AN EXPERIMENTAL PROTOCOL
FOR THE EVALUATION OF
GRAPHIC INPUT DEVICES IN MICROGRAVITY
by
JESS E. FORDYCE
Bachelor of Aerospace Engineering and Mechanics
University of Minnesota, Institute of Technology
(1983)
SUBMITTED IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR
THE DEGREE OF
MASTER OF SCIENCE IN
AERONAUTICS AND ASTRONAUTICS
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 1986
0 Massachusetts Institute of Technology 1986
Signature of Author
Certified by
d
Department oferonautics and Astronautics
May 9, 1986
Professor Steven R. Bussolari
Thesis Supervisor, Department of Aeronautics and Astronautics
Accepted by
Professor Harold Y. Wachman
Chairman, Department Graduate Committee
2
AN EXPERIMENTAL PROTOCOL
FOR THE EVALUATION OF
GRAPHIC INPUT DEVICES IN MICROGRAVITY
by
Jess E. Fordyce
Submitted to the Department of Aeronautics and Astronautics
on May 9, 1986 in partial fulfillment of the
requirements for the Degree of Master of Science in
Aeronautics and Astronautics.
ABSTRACT
An
experimental
protocol
was
developed to evaluate human
performance using three graphic input devices in the microgravity
environment of low earth orbit. The experiment is controlled by a Grid
Compass microcomputer which accepts cursor positioning commands from a
joystick, the computer keyboard, and a trackball.
The goal of the
experiment is to evaluate the performance and workload associated with
using the devices in microgravity.
Ground tests were conducted to verify the experimental operation
and to establish a baseline data base. Results of the ground tests
showed significant differences in both performance and workload between
the devices. Using the trackball resulted in the best performance
times,
followed by the joystick and keyboard, respectively.
In
addition, the trackball imposed the least subjective workload, while the
joystick had higher workload than the keyboard.
Thesis Supervisor: Dr. Steven R. Bussolari
Title: Assistant Professor of Aeronautics and Astronautics
ACKNOWLEDGEMENTS
The person whose help, patience and understanding contributed most
to this thesis was demonstrated by my wife, Barbara. In addition to
playing an active role by participating in the ground experiments as a
subject, she consistently bent a sympathetic ear when needed.
Perhaps
most importantly, she never complained while I constantly worked late
and neglected to come home.
The very existence of the project is due to the efforts of Dr.
Steven
Bussolari
and
Dr.
Byron
Lichtenberg, the co-principal
investigators who initially conceived of the project. Acting as thesis
supervisor, Dr. Bussolari guided the research leading to this document.
Statistical direction was graciously and diligently provided by Dr.
Alan Natapoff. I was constantly surprised by his ease of access and
willingness to instill an understanding of the concepts he knows so
well.
Without the support of friends, even
an otherwise pleasant
situation can be miserable. Their support becomes even more important
when
the
situation is inherently miserable, as demonstrated by
programming on the Grid! Had no one been around when I needed to launch
into an all too frequent tirade, I am firmly convinced that I would have
exploded, or at the very least, spontaneously combusted. Ed Kneller and
Steve Adkins had the misfortune (born of proximity) to bear the brunt of
my heated streams of vile profanity, so I should both apologize and
thank them for their unflagging good natures. My somewhat messy, yet
equally good natured office mate, Kathy Misovec was another good person
to talk to.
When it becomes necessary for a high strung, fast talking
Minnesotan to really blow off steam, nothing does the trick like another
high strung Minnesotan who is even more frantic. Keith Keller embodies
just such a person, and in our "conversations" together we seethed
enough energy to annihilate all enemies, whether foreign, domestic, or
circumstantial. Bill Kromydas and David Jarrett provided good sources
of comradery for conversation, beer guzzling, etc.
In addition to the people named above, I thank others who served as
subjects in the ground based experiments.
Adam Brody volunteered to
participate, based on his interest in human factors research, while Ted
McDade and Grant Schaffner served as both colleagues and participants in
the experiment. The contributions of the six U. C. Davis students at
the Ames Research Center is also greatly appreciated.
3
TABLE OF CONTENTS
ABSTRACT ............................................................
ACKNOWLEDGEMENTS....................................................
LIST OF FIGURES.....................................................
2
3
6
CHAPTER 1: INTRODUCTION
1.1 MOTIVATION FOR RESEARCH......................................
1.2 GOALS OF THE EXPERIMENT......................................
8
9
CHAPTER 2: BACKGROUND - REVIEW OF LITERATURE
2.1 OVERVIEW ..................................................... 11
2.2 MOVEMENT TIME - FITTS' LAW................................... 12
2.3 REACTION TIME................................................ 15
2.4 THE COMBINATION LAW.......................................... 16
2.5 THE RELATIONSHIP BETWEEN REACTION TIME AND MOVEMENT TIME ..... .17
2.6 SUBJECTIVE WORKLOAD.......................................... 19
2.7 A REVIEW OF PERFORMANCE USING GRAPHIC INPUT DEVICES .......... 21
CHAPTER 3: DESCRIPTION OF EXPERIMENT AND PROTOCOL
3.1 INTRODUCTION.................................................
3.2 EXPERIMENTAL DESIGN..........................................
3.2.1 THE TRIAL BLOCK..........................................
3.2.1.1 Memory Set Generation................................
3.2.1.2 Target Order Determination...........................
3.2.2 THE
3.2.2.1
3.2.2.2
3.2.2.3
3.2.3
TRIALS...............................................
Reaction Time Measurement............................
Movement Time Measurement............................
Accuracy of Time Measurements........................
23
23
29
29
30
32
32
32
33
SUBJECTIVE WORKLOAD RATINGS.............................. 34
3.3
PROGRAMMING THE FITTSBERG TASK ON THE GRID
COMPASS MICROCOMPUTER........................................
3.3.1: CHARACTERISTICS OF THE GRAPHIC INPUT DEVICES
3.3.1.1 Qualitative Description of the
Graphic Input Devices................................
3.3.1.2 Joystick and Trackball Movement Gains................
3.3.1.3 Keyboard Movement Gain...............................
38
39
45
46
3.3.2
THE INTEGRATED SYSTEM.................................... 41
3.3.3
A BRIEF OVERVIEW OF SOFTWARE............................. 44
3.3.3.1 File 10 ............................................... 45
3.3.3.2 Screen 10 and Graphics................................ 47
3.3.3.3 Results of Software Design............................ 47
5
TABLE OF CONTENTS
CHAPTER 4: RESULTS OF GROUND EXPERIMENTS
4.1 INTRODUCTION.................................................
4.1.1 DATA ANALYSIS..............................................
4.1.2 PREVIEW OF RESULTS.........................................
4.1.3 SUBJECTS...................................................
49
49
50
52
4.2 MOVEMENT TIME RESULTS & DISCUSSION
4.2.1 THE EFFECT OF GRAPHIC INPUT DEVICE
ON MOVEMENT TIME............................................ 53
4.2.2 THE EFFECT OF MOVEMENT DIRECTION ON MOVEMENT TIME ......... .63
4.2.3 THE EFFECT OF TARGET ID ON MOVEMENT TIME................... 69
4.3 REACTION TIME RESULTS & DISCUSSION
4.3.1 THE EFFECT OF GRAPHIC INPUT DEVICE
ON REACTION TIME..........................................
4.3.2 THE EFFECT OF MOVEMENT DIRECTION ON REACTION TIME .........
4.3.3 THE EFFECT OF TARGET ID ON REACTION TIME..................
4.3.4 THE EFFECT OF MEMORY SET SIZE ON REACTION TIME............
78
81
81
82
4.4 ARE REACTION TIME AND MOVEMENT TIME INDEPENDENT? .............
4.5 SUBJECTIVE WORKLOAD RATING RESULTS ...........................
4.5.1 THE EFFECT OF GRAPHIC INPUT DEVICE ON WORKLOAD ...........
4.5.2 THE EFFECT OF MOVEMENT DIRECTION ON WORKLOAD .............
4.5.3 THE EFFECT OF TARGET ID ON WORKLOAD ......................
4.5.4 THE EFFECT OF MEMORY SET SIZE ON WORKLOAD ................
84
87
88
90
90
90
4.6
4.7
THE EFFECT OF LEARNING ON PERFORMANCE ........................ 91
SUMMARY OF RESULTS........................................... 92
CHAPTER 5: CONCLUSION AND RECOMMENDATIONS
5.1 CONCLUSIONS..................................................
5.1.1 Summary of Fibdings.......................................
5.1.2 Applying the Results.......................................
5.2 RECOMMENDATIONS..............................................
APPENDIX 1 SUMMARY OF BLOCK DATA.......................
APPENDIX 2 DEVICE COMPARISON - "T" TEST................
APPENDIX 3 MOVEMENT TIME INTERACTION TABLES ............
APPENDIX 4 REACTION TIME INTERACTION TABLES ............
APPENDIX 5 WORKLOAD INTERACTION TABLES .................
APPENDIX 6 DATA DEPENDENCE UPON ID.....................
APPENDIX 7 DATA DEPENDENCE ON MEMORY SET SIZE ..........
APPENDIX 8 OVERALL DEVICE SUMMARY ......................
APPENDIX 9 WORKLOAD RATING SUMMARY .....................
APPENDIX 10 LEARNING EFFECT DATA........................
APPENDIX 11 THE EXPERIMENTAL PROTOCOL...................
APPENDIX 12 INSTRUCTIONS FOR WORKLOAD RATING SCALES .....
APPENDIX 13 TECHNICAL SPECIFICATIONS....................
REFERENCES..............................................
............
............
...........
............
............
............
...........
............
............
............
............
............
............
............
93
93
94
96
98
116
. 134
141
148
152
. 154
156
174
178
193
208
211
213
6
LIST OF FIGURES
FIGURE 2.1
2.2
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14
4
1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
4.13
4.14
4.15
4.16
4.17
4.18
4.19
4.20
4.21
4.22
4.23
4.24
4.25
4.26
A Linear Fitts Task................
Generic Bipolar Workload Rating....
12
19
Cardinal, Easy ID Targets...........
Cardinal, Hard ID Targets...........
Diagonal, Easy ID Targets...........
Diagonal, Hard ID Targets...........
A Typical Trial Block Memory Set....
A Typical Trial Block Memory Task...
Mental Workload Rating Scale........
Physical Workload Rating Scale......
Temporal Workload Rating Scale......
Performance Rating Scale............
Effort Rating Scale.................
Frustration Rating Scale............
The Integrated System Configuration.
Target Geometry.....................
27
27
28
28
vrA'avl 1
MoMvemeiii-
ema
m
C .........
CE4 Movement Time Average.............
CE4 Movement Time Average.............
CH4 Movement Time Average.............
CH4 Movement Time Average.............
DE4 Movement Time Average.............
DE4 Movement Time Average.............
DHl Movement Time Average.............
DH4 Movement Time Average.............
EASY ID, MS-1 Movement Time Average...
EASY ID, MS-4 Movement Time Average...
HARD ID, MS-1 Movement Time Average...
HARD ID, MS-4 Movement Time Average ...
Target ID Definition..................
Cardinal, MS-1 Movement Time Average..
Cardinal, MS-4 Movement Time Average..
Diagonal, MS-1 Movement Time Average..
Diagonal, MS-4 Movement Time Average..
Movement Time VS. Index of Difficulty [Cl]
Movement Time VS. Index of Difficulty [C4]
Movement Time VS. Index of Difficulty [Dl]
Movement Time VS. Index of Difficulty [D4]
MIT Overall Reaction Time Average.....
Reaction Time VS. Response Entropy....
Overall Subjective Workload...........
Device Ranking Scores.................
31
31
35
35
36
36
37
37
43
48
54
55
56
57
58
59
60
61
62
65
66
67
68
88
70
71
72
73
74
75
76
77
80
83
89
92
7
LIST OF FIGURES
Figure A13.1 Initial Start-Up Screen...................
A13.2 Selecting the Appropriate Name............
A13.3 Joystick Device Prompt....................
A13.4 Keyboard Device Prompt....................
A13.5 Trackball Device Prompt ...................
A13.6 Begin Testing Prompt ......................
A13.7 Memory Set Screen......................... .......
A13.8 Trial Screen-.............................. .. I . .
A13.9 Incorrect Target Acquired................. ......o
..... .
A13.10 Correct Target Acquired..................
.. .. .
A13.ll Initial Setting for Mental Workload......
A13.12 Final Cursor Position for Mental Workload
A13.13 Data for Blocks 1 - 4....................
A13.14 Data for Blocks 5 - 8....................
A13.15 The Exit Message .........................
A13.16 The Abort Menu Form ......................
195
195
197
197
198
198
200
200
201
201
203
203
205
205
206
206
8
CHAPTER 1: INTRODUCTION
Chapter 1.1
In
Motivation For Research:
recent
times,
computers
have
been
processes in a variety of environments.
used to automate complex
The proposed space station will
most certainly employ state-of-the-art computer systems to
control
vital
functions in the orbital environment.
human operator will
through
space,
to include
that
of
a
manager
Interactions
automated systems may be facilitated
through
The role
of
and
the
be expanded from that of a pilot flying the vehicle
accomplish mission goals.
user
monitor
the tasks in a
decisions
to
between the human operator and
by a computer that could guide the
logical
checklists and predetermined
making
order.
A
system
of
software
procedures could be used to enhance safety
and efficiency.
Anyone who has seen
Skylab
control stations of the mid 1970's has
no doubt been overwhelmed by the myriad array of switches and dials used
to control America's first space station.
station
In
order
to create a space
with an initial operational capability of 90 days
and
limited
ground support
requirements, it seems likely that a computer controller
system will be
implemented.
control
station
could
be
For
operator
may
sake
of
interact
through
a
menu
device
such
as
which
environmental
can
only
stressors
be
reached
will
have
using
the
level of efficiency,
if designers know
on
either
a joystick or trackball.
The operational station will have to attain a high
level
operator
selection protocol.
with the menu by using
keyboard, or a graphic input
a
convenience, a single
designed from which an on-station
could command all vital functions
The
the
human
what
effect
performance
in
9
weightlessness.
The extent
to
be impaired due to
which a person's motor control and productivity may
space
adaptation
syndrome upon initial exposure to
the space station environment is difficult to
predetermine.
Yet,
the
initial days of exposure to weightlessness may well be the most critical
time of the mission, since a great deal of work may be necessary to make
the station fully operational as crews are rotated, and so
important,
the
therefore,
operator
to
on.
It
is
to select an appropriate input device that helps
complete
the
necessary
tasks
quickly,
and
with
reasonably little workload.
Chapter 1.2
Goals Of The Experiment:
The primary goal of the experiment is to determine the interactions
of time performance and
perceptual workload associated with using three
graphic input devices to accomplish tasks with varying difficulty.
a practical standpoint, it is
also
useful to decide which input device
is "best" to use in weightlessness.
fastest
to
and have that
satisfy
The
desired
device
would
be the
use, impose the least workload, require the least training,
training
adaptation transients.
to
From
all
of
transfer
to the space environment with minimum
Although it may not
these criteria, it is
candidate compares to the others
for
be
possible for one device
useful
similar
to
know
how
each
task conditions, and how
various factors may interact to affect performance, holding
the
device
constant.
To
human
this
end,
an
experimental
paradigm was developed to measure
performance using three devices to
consisting of
a
memory
task,
and
complete
a
two
part
trial
a subsequent target acquisition in
10
response to the memory probe.
In preparation for executing the paradigm
in space, ground tests were run using fifteen subjects who established a
baseline database for the evaluation of computer
joystick, and a trackball.
keyboard arrow keys, a
Quantities measured were the
to a memory probe, time to acquire the appropriate
times to react
target, and a series
of subjective workload ratings, all of which will be described at length
later in this document.
The paradigm developed and the results
of
the
ground data are the topic of this thesis.
Chapter
to the use
reaction
2
is a review of literature that provides an introduction
of graphic input devices used in past experiments to measure
time,
movement time,
describes the experiment,
findings.
and
and
subjective
Chapter
4
is
workload.
a
Chapter
presentation
of
3
the
A concluding summary with recommendations for future efforts
is contained
in Chapter 5, and the many appendices provide the detailed
information referred to in the text.
11
CHAPTER 2: BACKGROUND - REVIEW OF LITERATURE
Overview:
Chapter 2.1
experimental paradigm developed to evaluate the use of graphic
The
input devices
elsewhere.
similar to other experiments that have been performed
is
The so called
"Fittsberg"
paradigm
task
dual
a Sternberg memory
Gopher, Hart, Lee, Dunbar, 1983) is a combination of
search
(Hartzell,
(Sternberg, 1975), coupled with a modified version of a
classic
Fitts target acquisition. (Fitts, Peterson, 1964). For our purposes, the
subject to identify and select one letter
the
requires
Sternberg task
(corresponding to a given memory
task is to move a cursor from
combining
two
the
possible to vary the
size,
an
set)
from a group of letters.
initial
of
have
on
it
workload
taken after each task condition were
evaluate
to
and performance.
developed
at
the
of
Workload ratings
NASA
Ames Research
this chapter will address the background of these
through a review of literature.
is
effect
Center and are the current standard for the evaluation of workload.
remainder
By
the memory set letters, as well as the
direction, and distance to the targets
these parameters
into a target.
location
tasks into the Fittsberg dual task paradigm,
number
Fitts'
The
topics
Movement Time - Fitts' Law:
Chapter 2.2
Fitts' law (Fitts,
Peterson,
1964)
is
intended to model the amount of information
by performing a movement.
an empirical relationship
that
a person can transmit
Essentially, the law states that the movement
time (MT) required for a person to make a
movement
of
a
distance (A)
into a target of size (W) is modeled by:
[ Eq. 2.1 1
MT - aID + b
where ID is the index of difficulty, classically expressed as:
ID - log 2 (2A/W).
Equation
1
is Fitts' law, and Equation 2 is one definition of ID.
Many other expressions
properties
movement
of
[ Eq. 2.2 ]
for
various
amplitude,
have
been
experiments.
The
and
W
ID
is
the
range
used
to suit the geometric
quantity
of
A
refers to
permissible
error
the
as
demonstrated in Figure 2.1.
W
target
initial
<- cursor
position
->
A
Figure 2.1
The
meaning
simply states the
of Fitts' law is fairly easy
the
time
required
to
to
acquire
depends on two things: the distance to the target,
One
may
controller
reasonably
device
expect
will
that
the
movement
take a finite amount of
conceptualize.
It
(or enter) a target
and the target size.
of a hand
time
to
or
cursor
travel
the
13
necessary distance to a target, depending on the maximum acceleration or
velocity that can be
obtained.
The
farther
away
the target is, the
Since the additional condition is imposed
longer it takes to get there.
that the target must be acquired (captured, not just passed through) the
person must use some sort of terminal control to
the target bounds.
satisfy
the
Thus, a smaller target may require a
difficulty
increases,
does the amount of information
so
Consider a man
that can be communicated.
pointing
the man has poor positioning accuracy, he may be
desired
country,
such
as
United States.
the
at a world map.
able
If
to land within a
With good accuracy, he
Given excellent aim, he could locate a
could point to a specific state.
city within the state, and so on.
From this standpoint, each subsequent
allows the man to convey
refinement
longer time to
Further, it can be reasoned that as
accuracy requirement.
the index of
stop the cursor within
more
information
with
a
single
movement.
Fitts' law
is closely related to information theory;
indeed, Fitts
used an analogy between human motor control and a communications channel
with a capacity
which can be predicted by Shannon's Theorem.
According
to Shannon's Theorem 17 (Kvalseth,1985), the information capacity (C) of
a
a communication channel with
finite
bandwidth (B) broadcast with an
average power (S) and a disturbing, independent
white Gaussian noise of
power (N) is given by:
C
-
[ Eq. 2.3 1
Blog2 [(S + N)/N] bits/sec.
By comparing equations 2 and 3 and considering
the
analogy,
it is,
apparent that the signal strength corresponds to the movement amplitude,
and
the
noise corresponds to the permissible movement error.
the two equations
don't
have
exactly
Although
the same form, they can be made
14
more alike by redefining A and W.
Although Fitts' law is primarily empirical, an attempt has been made
to derive a control model for rapid
hand
movements.
Connelly
(1984)
analyzed a first order model assuming that hand velocity can be directly
controlled, and a second order model assuming that hand acceleration can
be
directly controlled.
Connelly used the form of Fitts' law published
in 1954:
MT - Klog 2 (2A/W)
[ Eq. 2.4
A>W/2
which is identical to equation 1 excluding the added constant.
Assuming
an exponential solution for his linear control models, Connelly obtained
Fitts' law
as
the
solution for both the first and second order cases.
In summary, Connelly is not proposing that he has created the difinitive
model.
Instead,
he
maintains that
modeling
human
movements
as
a
physical system with some sort of rate control is more intuitive than an
analogy relating to communication channel capacity.
Many
studies have demonstrated
movement times
ID.
and
times
good
agreement
between
observed
predicted by Fitts' law over a wide range of
Because of this, Fitts' law is the generally accepted motor control
model
used
for predicting movement
reports the result
found
time.
However,
Connelly
(1984)
by Welford (1968) that the agreement was good
except for very small movement times, and for large movement times where
the
data
tended to "curve
gently
upwards".
As
a
result,
Welford
proposed a number of alternative expressions of Fitts' law, including:
MT - Klog 2 (A/W + 1/2)
A>W/2
[ Eq. 2.5 ]
in order to provide a better fit to the data.
There are many other versions of Fitts' law that have been proposed
as being "corrections"
or empirical adjustments to provide a better fit
15
to various data.
power
law
of
relationship
Departing
motor
control
presented
"empirical data from a
Data
came
from
two
from Fitts' law, Kvalseth (1982) proposed a
by
which
Fitts, and
substantial
that
-
log
states
supports
number
of
log
superior
his
claim
to
by
experimental
the
using
studies."
data for human locomotion.
plots
approximated straight lines.
of
time
versus
movement
It is unreasonable to
is a "correct" equation which exactly models human
point is to find a model that can be used
give
is
diverse areas: traditional industrial engineering
work measurement, and world record
found
he
as
a reasonable estimate of performance.
a
Kvalseth
amplitude
assume that there
motor
design
control.
The
tool which can
In this context, Fitts' law
is generally applicable.
Chapter 2.3
Reaction Time:
One important
element
of
the
reaction time required for the user
set item and respond to it.
portion
of
Fittsberg task is to determine the
to
identify the appropriate memory
Zaleski (1985)
experiment,
memory
of
Hick (1952) and Hyman (1953).
information content of a discrete target
a
The
acquisition can be measured by
the response entropy (H), which is related
correct
the
where the reaction time (RT) required
follows the relationship proposed by
probable,
that
the task is similar to that performed by the subject
choice reaction time
choices.
observed
to
the
number
of possible
If there are n stimuli (in our case letters) which are equally
then
the
probability
response is 1/n.
In
this
that each stimulus corresponds to the
form,
the
reaction
expressed as:
RT - c + dlog 2 (n)
[ Eq. 2.6 ]
time
can
be
16
Response entropy
(H)
is
defined
to be the number of possible bits of
information, calculated by:
[ Eq. 2.7 ]
H - log 2 (n)
With this definition, reaction time is expressed as:
RT - c + dH
[ Eq. 2.8 ]
Upon inspection, the expressions simply state that reaction time is
equal to a constant (overhead time) plus an additional time proportional
to the log 2 of the number
may
of
possible
choices.
The overhead constant
account for time not directly involved in mentally
information, such
as
focusing
on
the
processing
the
screen,
scanning the letters,
initiating a movement to indicate that a decision
has been made, and so
on.
It may also include some mental time constant
the
brain to begin processing the information, or neurophysical delays.
Overhead time
(c)
can
take
on
a
range
overhead
value
for
a
to prompt
of values, depending on the
specifics of the task and the manner in which
average
necessary
it
is
executed,
but an
Sternberg memory search task is on the
order of 400 milliseconds. (Johnson, 1981).
In
conclusion, the Hick -
Hyman law is not so much a model of how a human brain actually processes
information as it is an empirical equation which
observed
data
this light it
can
be
made
to
fit
reasonably well by carefully choosing the constants.
can
provide
a
means
of
In
predicting reaction times for
similar tasks and is useful as a design tool.
Chapter 2.4
The Combination Law:
A combination law has been proposed which suggests that
paradigm such as the Fittsberg task, a relationship may exist
predict
the
total
time to capture a target.
for a dual
which can
Capture time (CT)
could
then be expressed as a function of response entropy (H) and target index
of difficulty (ID).
This
prediction
was
tested by Zaleski (1985) who
used an expression of the form:
[ Eq. 2.9 1
CT - a + bH + cID
Zaleski experimented with a version of the
Fittsberg dual task paradigm
and concluded that the combination law predicted capture time as well as
the Hick - Hyman and Fitts' laws predicted
time, respectively.
proposed
1972.
and
reaction
and movement
According to Zaleski, the combination law was first
investigated by Beggs, Graham, Monk, Shaw, and Howarth in
Their results, however, were inconclusive.
Chapter 2.5
The Relationship Between Reaction Time And Movement Time:
The point which
makes
the validity of the combination law unclear
is its implicit indication that
time
the two processes, measured by reaction
and movement time, are independent.
thought on
this
issue:
one
There
Zaleski
summarized
the
results
of
are
manner,
as
suggested
independence
executed
to
parallel processing.
two
reaction
factors
time
is
One
to
in
carried
should
it
of
propose
depend
Using the notation: rt - f(H) to indicate that
entropy,
be
way
entropy, and movement time should depend only on
response
schools
well as Taylor's analysis (1976) of the
of the
independent, then
two
of
parallel.
Sternberg's experiments (1969) in
which certain memory searching tasks appeared
serial
are
maintains that the processes are serially
executed, the other suggests the processes
which
time
out in a
same
data
interpreting
that
only
if
they
are
on the response
the target difficulty.
reaction time depends on
is instructive to note the following summary
findings from some recent
the
Fittsberg tasks in Table 2.1.
of
The functional
17
18
relationships were reported only if they were statistically significant.
TABLE 2.1
Date
Experimenters
Findings
1984
Hart, Sellers, Guthart
1984
Zaleski, Sanderson
rt - f(H)
mt - f(ID)
rt - f(H)
mt - f(H,ID)
1985
rt - f(H)
Hart, Wickens
mt - f(ID)
1985
Yeh,Wickens,Hart
rt - f(H)
mt - f(ID)
1985
rt - f(H)
Zaleski, Moray
mt - f(ID) *
* mt interacted significantly with H
There is no clear consensus concerning the independence of reaction
time and movement time, at
What
can
be
reasonably
least
for
concluded
this definition of independence.
from an examination
detailed results of the aforementioned studies
bears
a
strong
dependence
on
"serial" in this sense.
dependence on
are
more
reaction
time
Reaction
Movement time data
time
demonstrates
appears to
a
strong
target difficulty, and is influenced by response entropy,
although the effect
results
that
the
response entropy, but no statistically
significant dependence on target difficulty.
be
is
of
not
is
not
really
always
surprising
found
even
findings, since the second part of the task
a subsequent memory search.
to
with
be
significant.
These
regard to Sternberg's
involves motor control, not
The issue of independence will be discussed
further in the context of our own experiment in Chapter 4.
19
Chapter 2.6
Subjective Workload:
In order to decide what configuration will be best
orbital
workstation,
it
is
objective
measures of performance.
instance, if a person demonstrates superior performance
as opposed to another, yet experiences a greater
fatigue
convenient
if
effects
the
may
resulting
had the lowest workload, but this may
necessary
input
in
with one device
Ideally, it
device,
screen
would
easier,
especially
be
information
the best time performance also
not
be
the case.
It may become
to sacrifice some level of performance in order to
operator's job
For
sense of workload as a
become important.
combination of
format, and user protocol
for the
important to consider human perception of
subjective workload as well as
result,
suited
make
the
if the tasks must be executed for a
substantial period of time.
Unfortunately,
the assessment of subjective
workload
is
difficult task since there is frequently a large
variance
in the data,
leading
to
successful
inconclusive
in
to
Nonetheless, such ratings have been
demonstrating differences
several Fittsberg related
used
results.
tasks
in
the
between
past.
obtain such measures is accomplished
indicate their subjective
feelings
a fairly
about
some
task
conditions
in
In general, the method
by
asking
subjects
to
aspects of workload on
bipolar rating scales as shown in Figure 2.2.
Mental Demand
Very Low
Very High
Figure 2.2
20
Usually there are from six to twelve
such ratings, including items
such as mental demand, physical demand, fatigue, and
others.
The group
of subjective workload rating scales is different for each task, and has
been in a development process at the NASA Ames
Research
various
implemented.
combinations
individual
rating
of
the
have
been
by
multiplying
The details of the process are
each
where
After the
measurements are recorded, they are blended
composite workload value
factor.
ratings
Center,
into
a
component by a weighing
somewhat variable (depending on
specific experiment) and a more detailed explanation
for
our
own
experiment will be given in Chapter 4.
A review of related literature leads to the following conclusions:
1)
Workload
is
a
function
of
both
response
entropy
and
target
difficulty.
(Hart, Sellers, Guthart, 1984)
(Yeh, Wickens, Hart, 1985)
(Staveland, Hart, Yeh, 1985)
2) Workload for
workload
ratings
individual
are
task
components
don't
add
linearly.
taken separately for Sternberg and Fitts
If
tasks,
their sum will not equal the corresponding value for the Fittsberg task.
Although the Fittsberg workload
is significantly greater than either of
the two components, it is less than the sum.
(Hart, Sellers, Guthart, 1984)
3)
Workload is influenced more
difficulty.
(Staveland, Hart, Yeh, 1985)
by
response
entropy
than
by
target
21
Chapter 2.7
A Review Of Performance Using Graphic Input Devices:
Other experiments
have
been
conducted
performance of graphic input devices used
tasks.
to determine the relative
to perform cursor positioning
Results are generally consistent and are summarized below.
Mount, Rudisill, and Schulze (1984) conducted a
task
cursor positioning
designed to evaluate graphic input devices that
space.
Subjects
used
devices
to
may
be
used
in
select targets on a computer screen
diagram representing spacecraft system components.
Results of the study
included a ranking of the four devices according to their speed, and are
summarized below.
1) Trackball
2) Joystick
3) Lightpen
4) Stepkeys
was
A study of seven graphic input
devices
conducted
resulting
by
Albert
(1982),
used for target acquistion
in the
following
order
according to speed:
1) Touchscreen
2) Light Pen
3) Data Tablet with Puck
4) Trackball
5) Force Joystick
6) Position Joystick
7) Keyboard
These results are perhaps a bit misleading, since the touchscreen, light
pen, and data tablet only required the
target;
whereas
the
trackball,
user
joysticks,
to
point
at
the desired
and keyboard required the
22
additional task of pushing a button to indicate that the cursor had been
positioned in the desired location.
to determine at what point the
Depending on the protocol developed
target
has actually been selected, this
condition may be entirely appropriate, however.
The other quantity measured in Albert's experiment was
accuracy, which had very different results.
positioning
From most accurate to least
accurate, the devices ranked as follows:
1) Trackball
2) Data Tablet with Puck
3) Force Joystick
4) Position Joystick
5) Keyboard
6) Light Pen
7) Touchscreen
Obviously, there is a sigificant tradeoff between positioning speed
and
accuracy.
Mehr
trackball
(1969)
to
aircraft on a
concluded
times,
compared
a
small
isometric
position a cursor over a
radar
large
for
that
CRT
using
contrary
to
screen
used
the
findings
a
of
cursor
may require the user to make several
simply
"blip"
air
representing
traffic
control.
an
He
a joystick resulted in shorter target acquistion
explanation is that moving
could
(force) joystick and a
other
studies.
One
possible
a large distance with a trackball
rotations of the ball, whereas one
apply force to the joystick and
the
cursor
would
keep
moving without any hand repositioning.
The results of
our
own
study are in general agreement with these
findings, as will be shown in Chapter 4.
23
CHAPTER 3: DESCRIPTION OF EXPERIMENT AND PROTOCOL
Chapter 3.1
The
Introduction:
experiment
devices
in
space
developed
will
be
to
evaluate
described in
the
this
use of graphic input
Chapter.
Aside
from
describing the inner workings of the computer program, the rationale for
the experimental design is presented.
Fittsberg
task
separately
mechanics
from
the
and
At
the
technical
experimental
risk
aspects
protocol.
with technical aspects involved with
of redundancy, the
will
of
exactly
discussed
The former portion deals
implementing the Fittsberg task on
the Grid Compass microcomputer, and is discussed below.
demonstration
be
what the user
sees
and
The latter is a
does
to
run
the
learning,
the
experiment, and is contained in Appendix 11.
Chapter 3.2
In
Experimental Design:
order to minimize
experiment was
effects
of
order
and
counterbalanced according to the parameters we felt were
most important, and
quantities
the
subject
to many practical constraints.
are listed below along
with
their
identifying
The varied
brackets.
symbol
1) Device
Joystick
[J]
Keyboard
(K]
Trackball (T]
2) Direction - position
of target with respect to cursor
Cardinal (C]
( up, down, left, right )
Diagonal (D]
( 300 in four quadrants )
in
24
3) Target Index of Difficulty
Easy [E]
Hard [H]
4) Size of Memory Set
To
ms - 1
[1]
ms - 4
[4]
begin,
the
three
devices
with
the
eight
possible
block
conditions of the factorial design are shown in Table 3.1 below.
Joystick
Keyboard
Trackball
Cardinal
/
Diagonal
\
/
\
Easy
Hard
Easy
Hard
ms - I
CEi
CHi
DEl
DH1
ms - 4
CE4
CH4
DE4
DH4
TABLE 3.1: DESIGN PARAMETERS
One of
the
most
important
elements
of
the
experiment
is
determine which device can be used with the greatest proficiency.
impractical,
however,
It is
to alternate devices after each block to balance
especially since there is a limited time on orbit
out learning effects,
that can be devoted
to
to
the
experiment.
Therefore, to reduce the time
required to connect and disconnect devices, the
subject
uses
a single
device for all eight trial blocks before moving on to the next device.
25
Subject
organized
to
this
side
constraint,
the
remaining variables were
according to their relative importance that
we
assigned
as
follows:
1) memory set ........... ( ms - 1 & 4)
2) index of difficulty...
E & H )
3) direction.............( C & D )
4) device................
Based on
J , K
,
& T)
this criteria, the three input parameter files listed in Table
3.2 were created
to
control
the order of the experiment.
There are a
fairly large number of such files that could be created according to our
hierarchy, but these three satisfy our goals for counterbalancing.
a subject enters the experimental program
subject
and chooses his name from the
menu (described later) a corresponding data
file
which is used to keep track of all experimental sessions.
the data
input
file
is
parameter
experiment.
The
A,B,C,A,B,C...
using the
same
a
letter
should
be
used to
end
letter
is
rotated
control
through
twice in a row.
testing with a different letter; so
exactly
what
effect
direction have on the targets.
attached
At the end of
a
one
flow
of
the
pattern
of
subject
from
Individual subjects begin
third of the people start with
file A, one third start with file B, and so on.
demonstrate
the
at the end of each session to prevent the
file
is
(either A, B, or C) which indicates which
file
input
When
varying
the
Figures 3.1 through 3.4
size,
distance,
and
26
File A
File B
File C
JCE1
JDH4
JCH1
JDE4
JDE1
JCH4
JDH1
JCE4
KDE4
KCE1
KDH4
KCH1
KCE4
KDE1
KCH4
KDH1
TCE4
TDH1
TCH4
TDE1
TDE4
TCH1
TDH4
TCE1
KDH1
KCE4
KDE1
KCH4
KCH1
KDE4
KCE1
KDH4
TCH4
TDH1
TCE4
TDE1
TDH4
TCH1
TDE4
TCE1
JDH1
JCH4
JDE1
JCE4
JCHi
JDH4
JCE1
JDE4
TCE1
TDE4
TCH1
TDH4
TDE1
TCE4
TDH1
TCH4
JDH4
JCE1
JDE4
JCH1
JCH4
JDE1
JCE4
JDH1
KDH1
KCE4
KDE1
KCH4
KCH1
KDE4
KCE1
KDH4
Table 3.2: LIST OF PARAMETER FILES
27
s
U
+ I
F
0
FD
Figure 3.1: Cardinal, Easy ID Targets
3
R
0
C
H
0
a
Figure 3.2: Cardinal, Hard ID Targets
28
K
a
z
0
Figure 3.3: Diagonal, Easy ID Targets
K
R
L
B
Figure 3.4: Diagonal, Hard ID Targets
29
Chapter 3.2.1
A
The Trial Block:
trial
block
acquisitions followed
consists
by
of
eight
memory
is
memorized.
seconds, the memory set
five
first
letters with corresponding target
boxes
appear
the same direction as the memory probe letter.
times,
and
target
During each
presented with the memory set to be
disappears
on
and
four
screen.
The
the
subject is instructed to move the cursor into the target
eight
and
six subjective workload ratings.
trial block, the subject
After
probes
box located in
This process is repeated
then the bipolar subjective workload ratings appear.
Figures 3.5 and 3.6
depict
a
typical trial block with four memory set
items and an easy target index of difficulty.
Chapter 3.2.1.1
Memory
Memory Set Generation:
set
letters
are
"randomly"
generated
by
a
pseudo-random
Pascal function in the software which uses a
multiplied by
a
"seed"
standard
clock
value to obtain a number between 1 and 26.
adding 64 to the result,
the
number
between capital A and capital Z.
time
By
corresponds to an ASCII character
The software will not allow any two of
the memory set letters to be the same, and the letter "V" is not allowed
since it looks too much like a "U"
the
on
the computer screen.
"dummy" letters presented in each trial that do not
the correct
response
are
generated
Similarly,
correspond
with the same rules.
to
In order to
keep the subject from learning the dummy letters and confusing them with
the memory set, new dummy characters
trial.
Using a memory set of four, each
be used as the memory probe twice.
used,
are
it
will
be
the
generated for each individual
letter
If only one
of the memory set will
memory
set
probe item for all eight trials.
letter
is
The order in
30
which memory set letters are used
during
the
trial
block is randomly
generated, with the constraint that there can be no more than one "pair"
of
memory
set
letters in a given block.
A pair
is
defined
as
condition where the same memory set letter is used twice in a row.
was done because the
pattern
the
This
becomes too obvious if multiple pairs are
allowed.
Chapter 3.2.1.2
Target Order Determination:
Similarly,
each
response twice, and
of the four targets
there
correspond
to
the
correct
can be no more than one pair of targets.
In
this way, the detailed experimental
design is done through the software
without
files
the
question.
probe
need
for large input
The software checks
letters
are
verify
specify
that
every
were
not
told
item
patterns.
of these rules ahead of
any
Of the fifteen subjects tested, none caught on to the pairs rule,
although several learned that each target would be acquired twice.
asked
agreed
in
all targets and memory
equally probable, and eliminates obvious
Needless to say, subjects
time.
to
to
if
they
that
put
this knowledge to use, all
they didn't since
especially with a
memory
set
not try to find any patterns.
it
required
of four.
too
experimental
much
added
When
subjects
effort,
In addition, most subjects did
31
SMemor'
sets
wx 1 s
Figure 3.5: A Typical Trial Block Memory Set
F
P
Figure 3.6: A Typical Trial Block Memory Task
32
The Trials:
Chapter 3.2.2
The eight trials begin after the
Figure 3.5) is replaced by the trial
displayed memory set (as shown in
screen
(Figure
The trial
3.6).
screen contains the four targets, four letters (one of which is a member
of the memory set) and a cursor in the center.
Reaction time is defined
until
to be the elapsed time from presentation of the trial screen
subject
cursor
initiates
movement.
its
The target is
illuminated when the cursor is within
Target acquisition requirements are met
bounds.
remains in
Time elapsed from initial cursor
the target boundaries is defined to be
movement until the cursor enters
the movement time.
the
target
for
the
0.4
when
the
cursor
at which time the screen is
seconds,
This
erased and a new trial screen is presented.
process
is repeated
until the eight trials of the block have been completed.
Reaction Time Measurement:
Chapter 3.2.2.1
Reaction
time
is
measurement
straight
very
keyboard, movement is initiated when the subject presses
Both
the
deadband.
joystick and trackball have a serial output and
If the
y
device
moved
been
an
arrow key.
a
built
in
joystick or trackball are not moved more than a small
amount, the x and
has
For the
forward.
output
non-zero value is sent to
are
both zero.
by an amount which
the
As soon as the the serial
exceeds
the
deadband,
a
computer's serial port and the criterion
for initial movement is satisfied.
Chapter 3.2.2.2
Movement Time Measurement:
Once the cursor has been moved, it must enter the target and remain
there for 400 milliseconds to satisfy the
capture
criterion.
Movement
33
time
is
measured
from
the initial cursor movement until
crosses the target boundary, and
for
400
milliseconds.
If
then
the
is
reentered, the last time the cursor crosses
exited
When the cursor
move if the
is
user
not
pressing
a
is within the target
be
no
to
for
account
doubt in the
rolling the trackball, or
key,
be released once
pushing on the joystick, any of the devices can simply
the target lights up.
subsequently
Since the cursor will not
that the target is acquired.
mind
and
into the target is used for
bounds the target box is illuminated, so there can
subject's
cursor
remains within the target bounds
target
the computation of movement time.
the
There is no purpose for a "steadiness"
criterion
tremors as in other Fittsberg tasks.
After the
motor
target
cursor has been in the
400
for
milliseconds,
there
is a 500
millisecond delay, and then the next trial screen is presented.
Chapter 3.2.2.3
Reaction
a special
Accuracy of Time Measurements:
and movement time measurements were accomplished by using
subprogram written in PLM to count the number of "ticks" from
tick corresponds to 200 nanoseconds and
Each
the 8087 microprocessor.
of
the code developed is capable
interpreting
time accurate to the nearest millisecond.
Loop times
whether or not the cursor has been moved are
millisecond,
approximately
and
loop
used
milliseconds.
seven
screen refresh rate
times
of
about 15 milliseconds.
66
Hertz
for
the ticks as an elapsed
on
required to check
the order of about one
measuring
movement
time are
The real accuracy limitation is the
which corresponds to a delay time of,
There is no way
to
change
the refresh rate of
the screen, so we must settle for this accuracy limitation.
34
Chapter 3.2.3
The
Subjective Workload Ratings:
method
used
to evaluate the workload imposed by the
task conditions will be
described
here.
Figures 3.7 through 3.12 show
the bipolar rating scales used for the ground experiments.
and rating scale descriptions (slightly modified
from
Instructions
NASA Ames) given
to the subjects prior to testing are included in Appendix 12.
flight
experiment uses an additional scale to indicate
the
level of nausea in case sickness has an effect on the data.
indicate a
subjective
workload
rating,
input device to move the cursor (shown
positioned
as
various
the
a
The space
subjective
In order to
subject uses the current
horizontal line initially
in the middle of the scale) to some point
between
the
end
points that best demonstrates the subjective workload experienced during
the trial
block.
Although
no
gradations are visible to the subject,
ratings are recorded on a 0 - 100 scale with 0 on the bottom, and 100 on
the top.
The ratings obtained can
then be compared for different block
conditions to assess the effect of increasing the memory set, increasing
the index of difficulty of the target,
or
to
compare workload between
devices.
In addition to the six raw ratings, a composite
obtained
by
multiplying
each
component
of
value
can be
workload by its weighing
factor, which reflects the importance of one workload component relative
to the others for a given subject.
forming
all
possible
pairs
indicate the member of the
Weighing
factors are determined by
of workloads, and asking the
subject
to
pair he considers to be the more significant
contributor to his perception of workload.
In
this manner, the weight
of a rating can take on a value between zero and five.
35
MENTAL
DEMND
Very High
Ver
Low
Figure 3.7: Mental Workload Rating Scale
PHYSbI CAL
Vr
High
-
Low
DEMAM
Figure 3.8: Physical Workload Rating Scale
36
TEMPORAL DEMNU
-- r.
High
IVery Low
Figure 3.9: Temporal Workload Rating Scale
Perfect
Failure
Figure 3.10: Performance Rating Scale
37
EFFORT
- Ury High
V"
Low
Figure 3.11: Effort Rating Scale
Vary High
Ver
Low
Figure 3.12: Frustration Rating Scale
38
Chapter
3.3
Programming
the
Fittsberg
Task
on
the
Grid Compass
Microcomputer:
One of the most significant qualities of the experimental
is that it is designed to run on the Grid
Compass
1109
paradigm
microcomputer.
The Grid was chosen simply because it is already space rated, and it has
a
bubble memory with roughly the same capacity as a 5.25" double sided,
double density
peripheral
abundant,
floppy
disk,
therefore
memory storage device.
however.
In
eliminating
This does not mean
fact,
the
driving
goal
the
need for any
that
of
memory
the
is
software
development was to develop code small enough to fit in the bubble memory
along
with
the necessary
including the operating
data
files,
system.
device
drivers,
accomplished
communicate
by
directly
using
with
the
rudiments
the operating
capabilites for the efficient
handling
procedures, menu forms, and file input
statements.
Instead,
of
system
on,
Pascal
to
use
These goals
and
PLX
its
to
unique
of screen graphics, serial port
-
output.
tedious and cumbersome since it does not use
output
so
In addition, the code has to be robust
enough to handle on-orbit contingencies and be easy to use.
were
and
Frankly, the code is
standard
Pascal input and
more efficient (albeit, awkward) code is
linked only to the system's built in compact system calls instead of the
large Pascal and PIA
libraries,
fast
The
as
possible.
thereby
tradeoffs
and
keeping the code as small and
added
difficulty
paid
off
eventually, since the bubble memory can hold all files necessary for the
mission.
39
Chapter 3.3.1: Characteristics of the Graphic Input Devices:
Chapter 3.3.1.1
Qualitative Description of the Graphic Input Devices:
The input devices chosen were the Measurement Systems 8534 joystick
and
the
8531
trackball.
characteristics
helpful to
Without
of the devices (listed
that
sends
cursor velocity increases.
quickly return
Appendix
13)
technical
it
will
be
The joystick is a
As the stick is deflected further,
When the
stick
is
released,
springs
it to the neutral position, and the cursor stops moving.
The trackball is
on how much the
a position controller that sends out information based
ball has rolled since the last update 6.66 milliseconds
It has a one to one correspondence with the cursor position on
the screen, so if the
proportionate
devices
the
out information based on how far it is
deflected from its neutral position.
earlier.
in
into
describe them from the user's standpoint.
rate control device
the
delving
are
amount.
ball
is
Also
rolled
included
the arrow keys on the
directions are possible
using
up, down, left, right, and
in
"up", the cursor scrolls "up" a
in
Grid's
the
keyboard.
the arrow keys.
any
keys are simultaneously depressed.
of
list of graphic
Eight
input
movement
The cursor can be moved
four diagonal directions if two
Continuous (although "jerky") motion
can be accomplished by holding the keys down and
using
the
keyboard's
automatic repeat function.
The
devices
have
same software driver is
both
devices
contain
identical serial output characteristics, so the
used
for
reading both devices.
microprocessors
that
information into digital x and y values sent
cord.
convert
through
Conveniently,
analog
voltage
an RS232-C serial
A built in hardware deadband eliminates the need for any software
40
manipulations.
One unique aspect of the serial output for the devices
value for the x and y positions do not directly correspond
the device is deflected.
separately.
In
information).
its
to
how much
For the sake of clarity, consider the joystick
neutral
Consider
is that the
a
position,
it
deflection
reasonably expect that as the stick
is
in
sends
out only zeros (no
the x direction.
deflected
value sent to the serial port would grow steadily
One
may
more and more, the x
larger.
Instead, the
device sends out l's slowly for a small deflection, and at an increasing
rate
for larger deflections.
exceed the
maximum
2's, and so on.
output
When the deflection gets large enough
rate
The trackball
to
of l's, the device begins to send out
works in an analogous way.
is apparent to the user, however, and
performance
None of this
using the devices is
quite good as described in Chapter 4.
Chapte? 3.3.1.3
As
Joystick and Trackball Movement Gains:
previously
software driver.
The
mentioned, both devices can be read with
moves the cursor by a
application
the
same
program reads in x and y values, and
corresponding amount.
The only difference in the
way the devices are treated by the software is the gain by which x and y
input values are multiplied.
For the joystick, it was found that a gain
of 1 worked very well, enabling
users
to move the cursor slowly enough
to acquire every pixel on the screen, yet fast enough to move across the
screen at a rate beyond that which anyone
trackball
acquire any
could
presented a different challenge.
desired
pixel,
but
it
actually control.
It could be easily used
The
to
did not move fast enough to reach
distant targets with a single hand movement.
Instead,
the
ball would
41
have to be rolled using several hand movements.
when
software
recommendation
Corporation,
drivers
came
1984).
Most software
are
from
written
one
mouse
for
are
"mice".
The
manufacturer.
"We strongly urge you
engineers
Similar problems
reluctant
to
try
following
(Mouse
2X
occur
Systems
magnification.
to do so, but after trying it,
they find the feeling of control and speed far outweigh the inability to
choose
individual
compromise
solution
multiplied by 1.99
1,3,5,7...
pixels..."
Upon
was reached.
The
experimentation,
however,
x
values
and
y
input
a
were
and then truncated, resulting in effective inputs of
This allows the cursor to move quickly and smoothly, without
sacrificing the ability to acquire individual pixels.
Chapter 3.3.1.4
Keyboard Movement Gain:
Aside from the serial
the
cursor
on the screen.
eventually chosen as
pixel
resolution.
although it
Given
devices,
is
the
By trial
gain
with
With a gain of 1,
possible
to
center
the
and
keyboard is also used to move
error,
A
gain
can easily jump out of
a
small
gain
of
3X
was
the best combination of speed and
the
cursor
the
cursor on any desired pixel.
the size of the targets, however, this
absolutely necessary.
a
fine
moves
very
resolution
slowly,
is
not
of 5 results in a "jumpy" cursor, which
target,
making positioning harder.
A
compromise gain of 3 allows for acceptable speed and resolution.
Chapter 3.3.2
The
The Integrated System:
operational
configuration
represented schematically as shown in
Compass Computer, one of the graphic
of
the experimental system can be
Figure 3.13.
This shows the Grid
input devices, a power supply used
42
to power the serial input device, and a representation of the wall.
the space flight configuration, the computer and
graphic
For
input devices
will be mounted on an adjustable work surface which can be positioned to
suit
an
individual's
neutral
body
posture
adjustable work surface is a major topic in
addressed in further detail in this document.
in
weightlessness.
itself,
and
will
not
The
be
43
Wall
Bus
Power
Grid
Compass
Computer
Graphic
Input
Device
Figure 3.13
Power
Supply
44
Chapter 3.3.3
The
A Brief Overview Of Software:
code
developed
to run the Fittsberg task on the Grid Compass
Microcomputer consists of nine
separately
which performs certain kinds of tasks.
in
compiled
PLM.
The
not arbitrary, in fact, there was no choice.
implement features such
and the timer, very
each
choice
languages
as input - output, graphics, menu select items,
specific
were
of
In order to
calls
have
to
be
made directly to the
operating system (which is written primarily in Pascal and PLM).
these
of
Seven of the modules are written
Pascal, and the remaining two are written in
languages was
modules,
Since
initially foreign to the programmer, some of the
early sections of code
are far from elegant.
of
thoroughly
the code have been
checked
Nonetheless, all sections
out
and
verified
through
extensive debugging and testing by fifteen subjects who used the program
without help
or
guidance.
A couple of problems were found during the
first few data collection sessions, but they were quickly resolved.
By virtue of being broken
into
reasonably
sized
and
separately
compiled modules, all sections of the code can be linked
exclusively to
the
instead of the
compact
calls
system
of
the
operating
system,
libraries normally associated with Pascal and PLM.
fast
and
directly,
small,
and
one
everything.
the code
since all features use routines which address
each
module
accessible by a 16 bit
change
This keeps
feature
can
address.
of
the
be
In
resident
addition,
code, it is not
The Grid takes approximately 30
in
contiguous
memory
memory
when it is necessary to
necessary
minutes
to
recompile
to compile all of
the source code, so it is extremely useful to be able to isolate certain
related tasks into smaller units.
that
comprise
Table 3.2 lists
the
separate
the code, and a brief description of the sort
of
files
tasks
45
performed.
The suffix
(.Pas)
designates Pascal source code, and (.PLM)
designates PLM.
Chapter 3.3.3.1 File 10
One of the major
the trials.
was
done
of
the program is to record data from
Lacking standard Pascal input through operating system calls.
system will only
example,
functions
read
or
write
output
procedures, all 10
Unfortunately, the operating
characters
from
or
this means that it is not possible to simply
"animal" to
a file.
to a file.
write
the
For
word
Instead, it is necessary to determine what and how
many characters need to be written to the file, create a string pointer,
(direct memory access) place each character in the string pointer, write*
the components of
the
string
pointer to the file, and free the memory
associated with the string pointer.
In
order
to write an end of line,
the ASCII characters carriage-return and linefeed must also
to the file.
Similarly, numbers can't be easily recorded in
A number must be converted to a string, placed
then
written
to
the
file
as
a
collection of characters.
be used by Pascal 10 libraries could mean the difference
program
and
other
written
the
would
Although
otherwise
between having
essential files fit in the bubble memory,
failure to meet mission goals.
file.
in a string pointer, and
tremendously tedious, the savings in memory space that
the
be
and
46
TABLE 3.2
File Name
Purpose
Main. Pas
This is the main driver for the experiment.
It
controls the flow of the session from a high
level, calling on many other routines to execute
necessary details.
Params.Pas
This module contains routines which set up screen
coordinates, control the position and location of
targets, and draw the screen graphics.
File.Pas
This module performs the file 10 to record data.
File.PLM
This PLM module is used to set up menu select
items in the same format the operating system
uses.
It is used to determine who the subject
is, and handle to abort contingencies.
Abort.Pas
This is the module which executes abort
procedures after the subject chooses the option
from File.PLM.
Bipolar.Pas
This module handles the subjective workload
rating scales.
ScreenInfo.Pas
This module displays a summary of experimental
data on
the screen at the end of the session so
that it can be photographed to provide a backup
in the event that data is somehow lost or damaged.
SerialRoutines.Pas
This module contains the routines used to read
and interpret the serial input.
TimerRoutines.PLM
This module contains the high resolution timer.
47
Chapter 3.3.3.2 Screen 10 and Graphics
Writing textual information to the
as writing to a file.
the
exception
that
The same
screen
string
is every bit as tedious
pointer technique is used, with
the exact position (in either pixel
coordinates) must be specified.
Screen
directly through the operating system.
graphics
or
are also implemented
The operating
system
choice of manipulating the screen's bit map of individual
of
pixels, or rectangles of pixels.
character
gives the
pixels, lines
Consider the targets used
in
the
The cardinal "easy" targets are located 60 pixels from
Fittsberg task.
the cursor origin,
and
are
20
X
20
pixels square.
targets are located 100 pixels from the origin, and
Cardinal "hard"
are 10 X 10 pixels.
"Diagonal" targets have the same geometric distance and size properties,
but
are located 300 from the horizontal.
scaling quantities
This is done
by the sine or cosine of 300
.
by
multiplying
See Figure 3.14 for a
qualitative explanation.
Chapter 3.3.3.3
Results of Software Design
The end result
of
the
software
is a computerized Fittsberg task
with subjective workload ratings that fits
in
the Grid's bubble memory,
and can be easily used by a subject without the need for an experimenter
to be present.
system
and
There is enough room in
other
The
parameter
Spacelab
files,
perform
for
the operating
and
up
to
27
complete data
mission will nominally have three crewmembers
executing the experimental paradigm
nine sessions.
bubble
essential system software, the executable experiment
code, the three input
sessions.
the
three
times
It may be possible, however, for
each,
for a total of
one or more crewmen to
an extra session, or for an additional, unexpected
participate.
crewman
Therefore, it is desirable to have a little extra room.
to
48
Hard target
10 pixels
Easy Target
20 pixels
._...../
100 pixels
60 pixels
Figure 3.14
49
CHAPTER 4: RESULTS OF GROUND EXPERIMENTS
Chapter 4.1
In
Introduction:
preparation
conducted
to
for
the
thoroughly
proceding
experiment,
ground
tests
were
check out the experimental procedure, and
establish a baseline data
Before
flight
set
further,
for
comparison
with
the
to
flight crew.
it may be useful to review once
again
the
experimental variables in question, as shown in Figure 4.1.
Joystick
Keyboard
Trackball
Cardinal
/
Diagonal
\
/
\
Easy
Hard
Easy
Hard
ms - 1
CEl
CHi
DEl
DHl
ms - 4
CE4
CH4
DE4
DH4
TABLE 4.1: DESIGN PARAMETERS
Chapter 4.1.1
Data Analysis:
The primary goal of
of
using
the data analysis was to determine the effects
the three graphic
conditions, both in
terms
input
Therefore,
it
is
in
each
of
eight
block
of performance and the feeling of subjective
workload imposed by the task.
to a single block condition,
devices
Data was recorded in blocks corresponding
thereby
isolating
sources
of
variance.
possible to compare two conditions directly using
simple two tailed "T"
test.
(Statistics for Engineers, 1982).
fashion, the changes in performance
and
workload
can
be
a
In this
observed as
50
functions
of
constant.
each
variable
independently,
The T test determines whether
holding all other factors
or
not
differences
between
means are significant at given confidence levels; however, it is another
matter to
determine if a statistically significant difference is really
of any importance.
not only on the
Importance is somewhat subjective, since it depends
magnitude of the difference between the means, but also
on the context in which the information is to be used.
Data reduction and subsequent analysis was done using the same Grid
compass
microcomputer
used
for the
statistical techniques employed were
written.
ground
fairly
The vast majority of the program
experiments.
simple,
is
Since
the
original code was
devoted
to manipulating
the data in various ways to display differences between block conditions
in an easy to use format.
Some of the tables
created
by the reduction
program appear in the Appendices, and will be refered to later.
Chapter 4.1.2
The
Preview of Results:
results
differences
between
whole, and for
statistical
in
this
chapter
differences
No
real
between
attempt
the
groups
as
a
was made to assess the
individuals,
since
individual
no direct bearing on the results we seek.
The primary
determine
which graphic input device is best suited for the
task overall, as well as
differences between specific block conditions.
Differences between individuals can be important,
subject
are intended to reflect
the block conditions, both for
individuals.
differences have
goal is to
presented
variability
however.
is very large, it may eliminate
between block conditions,
even
the
If between
significance
though significant differences may have
existed for each subject individually.
Fortunately, the consistency of
51
the data we obtained makes this
individual
subject
leads
issue
all
subjects.
maintained or even
Therefore,
The data for each
to statistically significant conclusions (at
least for performance measurements),
for
a mute point.
When subject
improved,
and
and these conclusions are the same
data
is
combined,
significance
the same conclusions can be reached.
there is overwhelming evidence to support the conclusions.
brief summary of findings can be
is
presented
quite
simply
A
in table 4.2
below:
fastest
Reaction Time
Trackball
Joystick
slowest
Keyboard
fastest
Trackball
Movement Time
Joystick
slowest
Keyboard
least
Trackball
Subjective Workload
Keyboard
most
Joystick
TABLE 4.2 Overall Rankings
Clearly,
the
trackball
is
performed in this experiment.
well as the lowest workload.
than
the
the
"best"
It has the
device for the type
fastest
of
task
performance times as
The joystick has better
time
performance
keyboard, but at the expense of a higher perceived
workload.
52
Detailed information about
these
overall
conclusions
and many others
will be presented in this Chapter.
Chapter 4.1.2
Subjects:
Data was generated by two groups of
Research Center, the other at MIT.
undergraduate
at the
subjects, one at the NASA Ames
The Ames group consisted of six paid
students participating in a set of experiments
bedrest
facility.
conducted
In addition to completing eight sessions in
our own experiment, they also served as subjects in several other tasks.
The MIT group consisted of one undergraduate and eight graduate students
(all
unpaid
subject's
volunteers)
who
completed
five
data was analyzed separately, and
combined data
were
the two groups was
separately processed.
that
sessions
each
of
the
data
gave
When
overall
subjects,
what
will really be presented are the results
data
is
discussed
is complete.
group are essentially identical, as
appendices.
groups
the Ames subjects did not complete most of the
results.
data
two
Each
The major difference between
subjective workload ratings, so their workload
group, since their
each.
can
for
the
inconclusive
combination
for
the
of
MIT
Performance results for the Ames
be verified in the appropriate
Chapter 4.2
Chapter 4.2.1
The
The Effect of Graphic Input Device on Movement Time:
results for overall movement time as
input device
(all
block
checked in any of the analyses.
times,
followed
the
significant
Appendix 1
lists
and Appendix 2
level,
by
of
graphic
each
important)
is the highest level
trackball has the fastest
the joystick,
and finally the
keyboard.
individual block condition, with
Figures 4.1 through
(and
function
which
Clearly, the
Similar results are obtained for
p<0.001 for all cases.
a
conditions and subjects blended together) are
significant at the 99.9% confidence
movement
53
Movement Time Results & Discussion:
4.9 graphically demonstrate
differences
between
the
devices.
the performance summary data for each block condition,
provides a table of the z statistics
between devices for
each block condition.
Results for individuals were
were not always obtained for
similar, but high significance levels
each individual block condition.
times averaged across all block conditions
were
significant.
for
For the fifteen subjects,
each
device,
Movement
however,
the following results were
obtained.
joystick faster than keyboard
15 subjects
14
keyboard slower than trackball
15 subjects
15
trackball faster than joystick
results
and
suggest,
3
provide
the
relevent
p<0.001
15 subjects
15
Appendices 2
p<0.001
p<0.001
numerical
data.
the variance between subjects had less of
As these
an
effect
than increasing the sample size, therefore the significance improved.
54
MIT OVERALL MOVEMENT TIME AVERAGE
m
e
C
n
1600 ..------------------------------------------------------------------------------..-....---..
1 4 0 0 - --------------------------------a-i -- -- - -- ----. -. - .--. .
1400 - .---.---------.
- -. -
| 1 00 ----------------------------- M.....---------------.-..
800 1- ---
----------
600 - -- -.---
.
-...
---....---
d.....
400.
200
-- ----
-------.....
. --.........
---
-- - -- ----
......
----.
......
JOYSTICK
KEYBOARD
TRACKBALL
FIGURE 4.1
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball faster than Joystick:
p<0.001
p<0.001
U
..
55
MIT CE1 MOVEMENT TIME AVERAGE
1200
- -----------------------------------------------------------------------------------------
..
.........
-..--.--- -----..--.-..-........
-.-.---......--...-..-..-.--.--.....- - - -- m 1000
~~~~ ~ ~ ~
.....................
-.......
.................
...
...
...
-..
.
800
S
e
c
0
n
- -----------------------------..-------------- -------- -------....
...
...
...
...
.......
...
...
...
..............
------ ------ ------- --- ------ --- ---- --- ---- --- ---- --- - ------ -- --- -- -- --- -- -- --- -- -- -...-. -.---...-.----.-..
...-..---... -..
-
600
400
d
S
......
200
-.
e.
- - - - -...
..--
....
0
JOYS TICK
JOYSTICK
KEYBOARD
KEYBOARD
T
FIGURE 4. 2
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<O.001
Trackball faster than Joystick:
RACKBAL L
TRACKEALL
p<O.001
p<O.001
56
MIT CE4 MOVEMENT TIME AVERAGE
1200-
m 1000
800S
e
600-
C
0
n
d
----
-----
-- --
................
200
--- --- ----..........
-
s
JOYSTICK
KEYB OARD
TRACKBALL
FIGURE 4.3
p<0.001
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball faster than Joystick:
p<0.001
....
57
MIT CH1 MOVEMENT TIME AVERAGE
1800
1600
1400
m 1200
S
c
1000
800
0
600
d
s
400
n
200
.----------------------------------------------------------------------------------------.
.... -------- -------- ------------------------------------------------..
--.........
-- - -- -- - -- .......
-.
-- - -- - -- --.--- - -- - -- -- - -- -- - -- --- --- --- --- --- --- --- --- ---.---- --- --- --- --- --- --- ---~~~~~~~~~ ----------------------------------------------------...
...
...
..
...............
................ ...
..
----- ------ -% - ------ ------ ----- ------ ----- -.--......-....
-----.
..... ........
--- --..
--..-...
.... . . . ...
....
-- -- -.--
0
JOYSTICK
KEYBOARD
TRACKBALL
FIGURE 4.4
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball faster than Joystick:
p<0.001
p<0.001
...
.
58
MIT CH4 MOVEMENT TIME AVERAGE
1800
1600
.
.................................
,..
.....................
m 1400
1200
e en
S 1000
0
800
C
0
n
d
S
...
.
.
.
600
400
..-...
.....:..
200
.'""
""""
""
0
JOYSTICK
KEYBOARD
TRACKBALL
FIGURE 4.5
p<0.001
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball
faster than Joystick:
p< 0 .0 0 1
50
MIT DE1 MOVEMENT TIME AVERAGE
1200
m 1000
..-------------------------............----------------.
....................................................
........ M,...........................
. -...................
800
S
e
600
C
0
400
d
S
200
n
............................
S................
. -------------....-.-.
S.t-.-
0
JOYSTICK
JOYSTICK
...
- -- - -- --.
KEYBOARD
TRACKBALL
TRACKBALL
KEYBOARD
FIGURE 4.6
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball faster than Joystick:
p<0.001
p<0.001
6o
MIT DE4 MOVEMENT TIME AVERAGE
1200
-------........................
....
..........
m 1000
800
S
S
C
600
0
400
d
s
200
n
. ......... ......................
........
-- ......
--
0
JOYSTICK
-
.
-.....-
KEYBOARD
KEYBOARDTRACK~BALL
FIGURE 4.7
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball
faster than Joystick:
p<0.001
p<0.001
....
U
MIT DH1 MOVEMENT TIME AVERAGE
2500 ----------------------------. ......................................
m 2000 ..-.----.--------.----.-----........
I
...
......
.-..
1500
.e
s
s0
5
- --
-
-..........
-
-
0JOYSTICK
JOYSTICK
KEYBOARD
KEYBOARDTAWm..
FIGURE 4.8
p<0.001
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball faster than Joystick:
p<0.001
TIRACl IALLT
62
MIT DH4 MOVEMENT TIME AVERAGE
-............................
2500
m 2000
........
.
1500'
m~
s
a
C 1000.
0
n
d
S
500.
0.
JOYSTICK
KEYBOARD
TRACKBALL
FIGURE 4.9
p<0.001
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball faster than Joystick:
p<0.001
63
Chapter 4.2.2
Movement
The Effect of Movement Direction on Movement Time:
direction had an effect on the time required to move into
a target, but to varying degrees.
the horizontal)
cardinal
targets
counterparts
(up,
Ames groups, the average
diagonal
targets
took
more
time
to
be
down, left, right).
result
resulted
In general, diagonal (300 relative to
was
in
that
acquired
than their
For both the MIT
for
all
and
block conditions,
longer movement times than did cardinal
targets in otherwise identical conditions.
These
results are shown in
Figures 4.10 through 4.13, with corresponding significance levels in the
captions.
Appendix 3 provides a table of
z
statistics
between
block
conditions, holding the device constant.
Although
on
the
average
diagonal targets had longer acquisition
times than cardinal targets, this
result
or all devices.
highlights
A summary of the
was not true for all subjects
subject data will be given here, but a thorough
for individual device and
treatment would require
literally hundreds of pages of graphs and tables, and therefore will not
be included in this document.
For the joystick, all block conditions for all subjects resulted in
longer
times
for
diagonal
were significant at the level
movements than for cardinal.
p<0.001.
Most results
This would not be surprizing to
anyone who has used the joystick, since it is spring loaded on two axes,
x and y.
The result is that it is easier to move the joystick along the
x or y axis than along a diagonal.
Keyboard
For
most
cardinal,
data
subjects,
but
many
for individual subjects is
diagonal
cases
perfectly executed, the user
targets
are
not
took
somewhat
longer
statistically
to
inconclusive.
acquire
significant.
than
If
should be able to acquire diagonal targets
64
faster than cardinal targets by holding down two keys at once and moving
at a 450 angle, and
then releasing one of the keys and moving laterally
into the diagonal target.
The keyboard
that if two keys are held down
into
repeat characteristics are such
simultaneously,
keystrokes
can
be fed
the keyboard buffer at a higher rate, thereby allowing the
cursor
to move
faster
than
if
only
one arrow key is pressed.
subject mastered this technique and has average
In fact, one
diagonal movement times
which are less than corresponding cardinal movement times
for all block
conditions, although the differences did not prove to be significant.
In
longer to
general, trackball data demonstrates that diagonal targets took
acquire than cardinal, but for individuals the result was not
very significant.
no
In
fact,
for six of the fifteen subjects, there was
observable difference between cardinal
other variables
with equal ease,
constant.
and
the
diagonal,
holding
all
The trackball can be rolled in any direction
magnitude of the roll angle in any direction
results in a proportional magnitude of
regardless of direction.
and
cursor
movement
on the screen,
65
MIT EASY ID, MS=1 MOVEMENT TIME AVERAGE
1200
M 1000
L
L 800
S
E
C
0
N
D
S
............
. . ..... .................................
....................
...............
......
.........
600
400
200
-
---
-
..-...
.
lmm
0
JOYSTICK
KEYBOARD
TRACKBALL
I CARDINAL M DIAGONAL
FIGURE 4.10
Joystick
- Cardinal faster than Diagonal: p<0.001
Keyboard
- Cardinal faster than Diagonal: p<0.05
Trackball - Cardinal faster than Diagonal: p<0.001
66
MIT EASY ID, MS=4 MOVEMENT TIME AVERAGE
1200 ---------------------------------------------
L
L
800 .
1 200 ---------S
E
600
...
C M 00
400 --
s
.
........
----. ------. ------.. ------------.
.A.
.AG.
........
200-JOYSTICK
KEYBOARD
TRACKBALL
D CARDINAL 13 DIAGONAL
FIGURE 4.11
Joystick
- Cardinal faster than Diagonal: p<0.001
Keyboard
- Cardinal faster than Diagonal: not sig.
Trackball - Cardinal faster than Diagonal: p<0.01
67
MIT HARD ID, MS=1 MOVEMENT TIME AVERAGE
2500
M
- - -
- -
- -
- -
-
- -
--.........----...............----
2000
L
L
1500
S
E
C 1000
0
N
D 500
S
-F-....
0
KEYBOARD
JOYSTICK
E
CARDINAL I
TRACKBALL
DIAGONAL
FIGURE 4.12
Joystick
- Cardinal faster than Diagonal: p<0.001
Keyboard
- Cardinal faster than Diagonal: p<0.001
Trackball - Cardinal faster than Diagonal: p<0.001
......
68
MIT HARD ID, MS=4 MOVEMENT TIME AVERAGE
2500
M
..---------------------- ----------------------..
.. .. . ................................
2000
L
---...
..
---..
-.
-- - - -- --- -- -- - ---- -..
-..................
I 1500
S
E
C 1000
0
N
D 500
S
----------
...
-........ ................. ..
.........
......
.............
0
JOYSTICK
KEYBOARD
TRACKBALL
C3 CARDINAL 0 DIAGONAL
FIGURE 4.13
Joystick
- Cardinal faster than Diagonal: p<0.001
Keyboard
- Cardinal faster than Diagonal: p<0 .0 0 1
Trackball - Cardinal faster than Diagonal: p<0.001
69
Chapter 4.2.3
It
The Effect of Target ID on Movement Time:
should
come as no surprize that movement time is influenced by
target index of
Fitts' law.
difficulty
(ID),
since
that
Data for both groups showed that
is the basic premise of
for all block conditions,
the difference in movement times as a function of
p<0.001.
Appendix
6
lists
the
z
conditions, where only ID is varied.
statistic
ID
was
between
The resulting graphs
group are shown in Figures 4.15 through 4.18.
similar block
for
specific task,
may seem appropriate to attempt to fit an equation to the
see how
well
the theory holds up.
it is a forgone
conclusion
that
the MIT
Since Fitt's law predicts
movement time as a function of index of difficulty for a
it
significant,
data
and
With only two data points, however,
a
perfect
linear
relationship will
result.
As
pointed
arbitrary.
out
in
Chapter 2, the definition of ID
For our geometric properties,
is
somewhat
however, the equation used to
define ID is:
ID where R
log 2 [(R
and R
+ R ) /
(Ro - R,)]
Eq. 4.1]
are defined in Figure 4.14 below.
initial
+ <- cursor
.position
target->
Ri
R
Figure 4.14
Figures 4.19 through 4.22 show the
MIT group.
resulting
data in this form for the
70
MIT CARDINAL, MS=1 MOVEMENT TIME AVERAGE
1800
--------------------------------------------------------------------------------------.....---.
.....-----.-------------------------..
---------------------------------------------.
M 1600
I 1400
L
L 1200
s 1000
E
C
800
o 600
N
D
S
400
200
----- -- ---.-
--
---.--o
--
-----
--- -. ...-
***-------.----.-..... . . ....-
--.............
-- --
0
JOYSTICK
-.
.
*
KEYBOARD
-- -
TRACKBALL
1EAsy 0 HARD
FIGURE 4.15
Joystick
-
Easy faster than Hard: p<0.001
Keyboard
-
Easy faster than Hard: p<0.001
Trackball
-
Easy faster than Hard: p<0.001
--...
. ....
71
MIT CARDINAL, MS=4 MOVEMENT TIME AVERAGE
1800
-------------------------------------------------------------------------------------------.
---------------------------------------------...---------------------------------
M 1600
.................................
1 1400 ..............................................
L
---------------------------------------------..--------------------------------.
L 1200
S
E
C
O
N
D
S
1000
800
600
. .
400
200
.
. . .-
--
. . . ...-....
.
-----------.--
0
KEYBOARO
JOYSTICK
El EASY
TRACKBALL
M HARD
FIGURE 4.16
Joystick
-
Easy faster than Hard: p<0.001
Keyboard
-
Easy faster than Hard: p<0.001
Trackball - Easy faster than Hard: p<0.001
72
MIT DIAGONAL, MS=1 MOVEMENT TIME AVERAGE
M
2500
----------- --------------------- ----------.-.. -.. ..-..............o.......0o............
2000
-----------------------------------------..-.
....
......... o.............o...........
L
1500
S
E
C 1000
0
N
D 500
S
S......
......
..................
..-------.......................
...
O.................
0
JOYSTICK
KEYBOARD
TRACKBALL
EASY IN HARD
I
FIGURE 4.17
Joystick
- Easy faster than Hard: p<0.001
Keyboard
- Easy faster than Hard: p<0.001
Trackball
- Easy faster than Hard: p<0.001
73
MIT DIAGONAL, MS=4 MOVEMENT TIME AVERAGE
M
2500
------------------------------ ----------------------------- -----------------------------
2000
-------------------------- ------------------.------ -------------------------
L
L
1500
S
E
C 1000
0
N
D 500
S
-- -- - - J-O------
-- - - - -
-----------------.-.--------------A--
-- - -- - - -- - - -- - - -- - -----
0
JOYSTICK
KEYBOARD
M EASY
TRACKBALL
0 HAMD
FIGURE 4.18
Joystick
- Easy faster than Hard: p<0.001
Keyboard
- Easy faster than Hard: p<0.001
Trackball - Easy faster than Hard: p<0.001
74
MOVEMENT TIME [MS] VS.
INDEX OF DIFFICULTY [BITS]
Cardinal targets,
1800 - ---------------------
- -------
ms -
1
---------------
-
-----
.-- .
K
m 1600 -
-------------------------
t
--- - -
1400
--
------------------
----------------
-------- ft--------------------
m
in
s
d
e
-- --------- m---
--------------- ---------
1200
-
1000
-
800
-
600
-
JJ
---------------
400
-
--------- ----- m------
K
--- -- - -
----------------------
t-------- -------.-
m-------
m-------- -----
.
C
-- -
d
s
--------------
-------------
--
--------.- .--T
------------------------
J
T
--------
------------------ -m----.------..
T
200
-
.----
- --
- - -- - -- ----------------.
.-------------------.-
02.5
3.0
3.5
4.0
2.81
4.5
4.39
ID
BITS
Figure 4.19
75
MOVEMENT TIME
rMSi
VS. INDEX OF DIFFICULTY rBITS1
Cardinal targets, ms - 4
1800 - --------------------------------------------------
------K
m 1600 - - -
--------------- -- -- - -----------
-------------
-
t
1400 m
1
1200 K
s
e
1000 ------
C
0
n
-- --------------------------
- - - - - - -----------
800 -
d
s
600 -
-- ------------ ------------------------------- -T
400 -
---
- -------
-
T
----------------
------------ - ----------------------------------------------
200 -
0 2.5
-
3.0
3.5
4.0
2.81
4.5
4.39
ID
BITS
Figure 4.20
76'
OF' DTF'FTC~T1T1PY
TIME
MOVEMENTMOV
ME
M rMSl
Tmsi vs.
v. INDEX
INDX
____ ___r- T -
r BTT
Diagonal targets, ms - 1
2400
m
K
. .....
K-.-
- -------------------------- -----------
2000
t
J
1600 ------ -- - -
S
-------------
----------------..----
-
- - - - - - - - - - - - - - - - - - - - . .
1200 -----------K
J
------- ------ - ......
800 ------ ----------------------------TT -400 - ---- --
- -
-- - - - - - - -
---------..---- --
n
d
S
0
3.0
2.5
3.5
4.0
2.81
4.5
4.39
ID
BITS
Figure 4.21
.....-
77
MOVEMENT TIME [MS1 VS. INDEX OF DIFFICULTY r BITS1
Diagonal targets, ms - 4
2400 -
m
K
-----------
2000 ------- ----------- m-------------------------------
t
J
1600 -------
------------- m--------
- -------
--------------- --
m
1200 ------- ----- m- -m-------K
800 ------400 ------ -
-
m------------------
--.- ---
J
T
T
- - - --
- - -- - ---------------
--
-----------
n
d
1
--
0
2.5
3.0
3.5
4.0
2.81
4.5
4.39
ID
BITS
Figure 4.22
78
Chapter 4.3
As
may
Reaction Time Results:
the size of
be expected, reaction time was significantly influenced by
the memory set (p<0.001).
caused by the input
device.
These
The other significant effect was
and
other phenomena are discussed
below.
Chapter 4.3.1
The Effect of Graphic Input Device on Reaction Time:
The graphic input device used had a significant
time.
Averaging
effect on reaction
across all block conditions for each group
gave
the
following results.
joystick faster than keyboard
(p<0.001)
(p<0.001)
Ames
MIT
keyboard slower than trackball
(p<0.001)
(p<0.001)
Ames
MIT
trackball faster than joystick
(p<0.001)
(p<0.01)
Ames
MIT
Appendix 1 lists
condition,
the
Appendix
means
and
standard
deviations
for
each block
2 lists the z statistics between devices for
block condition, and
Appendix
across all block conditions.
8
gives
the
overall
each
results averaged
Figure 4.23 shows the overall
result
for
reaction time as a function of graphic input device.
Based
on
experiment,
experience
these
results
that the joystick should be
is given to how the user
the
using
are
three
devices
tested
not really surprising.
in
this
It makes sense
faster than the keyboard when consideration
initiates
movement
with
each
device.
When
using the joystick, the user already has the stick in hand at the moment
the
decision
pushing
in
to
move
is
made.
the desired direction.
Initiating movement is as simple
With
the
keyboard,
however,
as
the
79
subject must decide which way to go,
key or keys.
There
is
an
and then push the appropriate arrow
inherent
lag
time
for the keyboard that
probably accounts for the difference, although there may be some
process
during
the
memory search that is different.
which can determine the
We have no
data
exact reasons why this result is true, but this
explanation is most likely the
significant difference.
mental
greatest,
if
not
only,
cause
of the
A similar argument holds for the trackball.
Differences between the joystick and trackball are less significant
(at
least
at
speculative
initiating
the
block
explanation
joystick
level),
is
that
movements
yet
the
difference
subjects
because
fact, many subjects mentioned that
they
took
greater
were
more
subjects indicated that they felt more confident,
their
intended
targets,
but
care
One
while
it was harder to control.
began movements using the joystick for this reason.
overshot
exists.
then
In
cautious when they
With
the trackball
and that they
usually
quickly "fine tuned" the
cursor position to land within the target.
Reaction
time
differences
for
individual
essentially the same, except the differences
trackball
lose
significance,
block conditions
between
the
as Appendix 2 demonstrates.
individual subjects are essentially the same.
are
joystick and
Results for
80
MIT OVERALL REACTION TIME AVERAGE
900
-------------......................
-----------------------------------
800
----------.........................
---------------------------------
700
...--
-------
------
600
....-
--------
----- 0
500
...-..
- ---
-------.....
c0 400
....-
300
....-
n
..------
2/
..\\\ ---
d 200 -....
S
100 - -..-0
64
J
JOYSTICK
KEYBOARD
TRACKBALL
FIGURE 4.23
p< 0 .001
Joystick
faster than Keyboard:
Keyboard
slower than Trackball: p<0.001
Trackball faster than Joystick:
p<0.01
----
Chapter 4.3.2
Movement
time,
had
a
nonetheless.
81
The Effect of Movement Direction on Reaction Time:
direction,
although
significant
effect
seemingly
for
not
the
related to reaction
joystick
and
keyboard
Appendix 4 shows the z statistic between each of the block
conditions.
Again, the reasons behind this result are probably more
the
way
in which the device is physically used than
processes of deciding
It seems reasonable to
related to
with
the
mental
which target corresponds to the correct response.
assume
that
for
the keyboard, longer diagonal
reaction times are the result of the fact that the subject has to locate
and
depress two keys instead of just one.
likely (based
on subject comments)
while initiating diagonal
tendency
results
to
in
move
movements
slightly longer
cardinal
and
and
because
reaction
no
diagonal
the
joystick,
it
is
that subjects are being more careful
along either the x or y
"preferred" directions,
between
For
of
the
axis.
times.
significant
stick's
This
The
greater
trackball
differences
reaction times.
Data
natural
has
care
no
are observed
for
individual
subjects are in substantial agreement with the group results.
Chapter 4.3.3
The
The Effect of Target ID on Reaction Time:
data essentially supports the
independent of
target
ID.
Appendix
similar blocks, varying only ID.
claim
6
that
reaction
time
is
gives the z statistic between
82
Chapter 4.3.4
In
The Effect of Memory Set Size on Reaction Time:
accordance with the Hick - Hyman
increased significantly
as
relationship,
time
a function of response entropy (p<0.001 for
all conditions).
Appendix 7
blocks,
only the size of the memory set.
varying
reaction
shows
the
z
statistic
between
Recalling
similar
equations
2.6 through 2.8,
RT - c + dH
In
this form,
H
is
equal
H - log
to
the
number
transmitted in response to the memory set.
1 bit,
and
a
2
of
device
bits
of
information
A memory set of one requires
memory set of 4 requires 2 bits.
only a function (significantly) of
(n)
and
Since reaction time is
memory set, Figure 4.24
demonstrates the effect of varying these quantities for the MIT group.
83
REACTION TIME [MS] VS. RESPONSE ENTROPY [BITS)
R
1200 ------------------------------------------------.............
T
u~oo
1100 ---------------------------------------------.-----...
..
K..
1000 ------------------------------------------- -..........-....-.
J
900 -------------------------------
----------.-----..-
m
......--
T
800 -------------------
1
------------------
.-.--
-.- ..........-
1
s
Ti
c
0
n
d
700 --------
6
~
-----------------
--
--------------.- ............-
K.
600 ---------------
--------------------------.- ..-----.........-
T
S 500 --------------m------------------m---------
1.0
1.25
1.5
H
BITS
Figure 4.24
------...
1.75
m
- ....-...
2.0
84
Chapter 4.4
This
Interaction Between Reaction Time and Movement Time
chapter is a follow-up to Chapter 2.5, where the issue of the
independence of
doubt.
reaction
time
The findings from our
and movement time was left very much in
own experient are similarly inconclusive,
but it does appear that movement
entropy
(H)
and
time
is
a
function of both response
target ID, contrary to other findings that
function only of ID.
MT
is
a
Appendices 7 and 8 show the relevent z scores for
RT and MT, varying either ID or H.
Consider first the Ames group
RT is a function of H, p<0.001
data.
Appendix 7 demonstrates that
for all cases.
Appendix 6 shows that in
some cases, RT also increases significantly with target ID, with four of
twelve
blocks significant at the p<0.05 level, and
Therefore, we
ID.
one
p<0.01.
block
might conclude that RT has a weak interaction with target
Movement times, however, show a strong dependence on target ID (all
blocks, p<0.001)
as
well
as
on
the size of the memory set.
increase with H for all block conditions and devices.
three of four blocks are significant (1
p<0.001,
All MT's
For the joystick,
2 p<0.01).
Similarly,
three of four keyboard blocks are significant (1 p<0.01, 2 p<0.05).
All
trackball blocks are significant (p<0.001).
This is the only point at which the
For
the
RT
p<0.001,
target
but
with
increased
target
ID
with
Ames
groups differ.
response entropy for all cases, with
produced no significant
ID
for
significantly affected by H for any
all
with (1 p<0.001,
cases
effect
(p<0.001),
on
RT.
MT
but
was
not
of the joystick or keyboard blocks.
Three of the four trackball blocks were
however,
and
MIT group, there was less interaction between the two portions
of the task.
increased
MIT
significantly
1 p<0.01, 1 p<0.05).
affected
by
The interaction is
H,
far
85
less significant than for the Ames group, however.
Considering the
different
two
groups
conclusions.
For
function of H (p<0.001), but
separately,
the
is
one
may
arrive
Ames group, it seems that
at two
RT
also influenced by target ID.
is
a
MT is a
function of target ID (p<0.001), but there is also a strong, significant
relationship to H.
On the other hand, RT for the MIT group depends only
on H (p<0.001), and MT is a
function
only
of
ID
(p<0.001)
for
the
joystick and keyboard, with evidence of interaction for the trackball.
The review of literature presented in Chapter 2 provides a clue for
what
may
be
reasonable
speculation
concerning
Zaleski's experiments of 1984 and 1985, two
subject
Fittsberg task with results similar to our own.
Zaleski's
two
this difference.
The
groups performed a
only difference in
experiments (based on what could be determined from
descriptions) was that
in
1984
For
were six undergraduate students,
and the MIT subjects
and
eight
substantially
the
same
his
he used undergraduate students, and in
1985 he used graduate students.
undergraduate
In
our
experiment, the Ames subjects
graduate students.
results
that
consisted
of one
Surprisingly, he reported
we
obtained
concerning
the
functional relationships of RT and MT on H and ID for the two groups.
Any explanation of the difference between
students
and
undergraduate
Throwing caution
students
have
to
the
developed
students
wind,
certain
it
is,
may
mental
of
the results for graduate
course,
be possible that the graduate
patterns
problems as a result of their educational background.
group
speculation.
to
solve
various
The MIT graduate
consisted of engineering students who had been solving
technical
problems in a step by step manner for years, and perhaps that experience
is responsible
for
the greater separation of the two tasks involved in
86
the Fittsberg paradigm.
From this standpoint, it is not unreasonable to
suggest that the manner
in
which mental processes (or problem solving)
are carried out may depend a person's specific educational background.
87
Chapter 4.5
Generally,
subjective
it
Subjective Workload Rating Results:
is
workload,
nonetheless.
The large
difficult
to
but
significant
some
variance
obtain
associated
decisive
results
with
results
were
for
obtained
subjective feelings
across block conditions, devices, and subjects tends to detract from the
significance
workload
of
(the
the
differences
weighted
To
treatments.
The
value obtained by multiplying each
rating by the subject's
Chapter.
between
weighing
factors)
will
be
overall
individual
discussed in this
demonstrate the disparity between how subjects weigh
the
importance of each of the ratings, consider Table 4.3 below.
Mental
Subject
Subject
Subject
Subject
Subject
Subject
Subject
Subject
Subject
Subject
Subject
Subject
Subject
Subject
Subject
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
Physical
Temporal
Performance Effort
Frustration
3
0
3
2
5
2
1
2
3
2
3
4
3
5
0
0
0
1
1
5
5
4
4
4
3
4
3
2
0
4
2
1
3
4
1
2
1
1
4
4
0
3
0
0
5
3
4
1
1
4
0
5
2
2
2
5
1
2
3
2
1
5
3
5
3
3
4
5
0
2
5
5
3
3
2
3
0
1
0
1
5
0
3
4
2
0
0
4
3
4
Table 4.3
Although the above
weights
are
weightings has been shown to reduce the
the
composite
disparate,
variance
workload result (Hart, et al.).
the end result of the
between
subjects for
For our experiment,
no
attempt was made to see if the weights had any effect on between subject
88
variance.
Differences between
were not significant,
individual
block
statistical
finally
group
values
conditions
significance),
yielded
workload means for individuals generally
summed
across all MIT subjects for
demonstrated
and
a trend (yet
summing all
significant differences,
increasing the sample
size.
Appendix
blocks
lists
lacked
each
device
for
primarily
9
still
as
a
result
of
the mean and standard
deviation of each workload component for the MIT group.
Chapter 4.5.1
The Effect of Graphic Input Device on Workload:
The subjective experience of workload varied as a
graphic
input
Appendix
device
used
to
execute
the
1 lists the means for each block
Appendix 2
lists
had
higher
condition
workload
differences
for the joystick
that for the joystick
is the diagonal block conditions
poor
handling
Differences
qualities
between
the
of
that
the
keyboard
as demonstrated by Figure 4.25, and
joystick
imposed
are
joystick
and
In reference to the
None of
than either the
the joystick -
trackball comparison.
-
-
trackball comparison, it
significant, reflecting the
for
diagonal
movements.
the differences become significant
verified
by
Appendix 8.
higher workload than the keyboard and
(p<0.001), and the trackball
It
and trackball were not significant.
Summing across all block conditions,
the
device,
achieve significance, and only three of the eight
conditions are significant
is instructive to note
and
ratings
keyboard or trackball for all block conditions.
keyboard
of the
target acquisition task.
the z statistic between devices.
MIT group, the joystick
function
imposed
less
although the difference is not significant.
Overall,
trackball
workload than the keyboard,
89
MIT OVERALL SUBJECTiVE WORKLOAD
40
...---------------------------------
.---
.--------------------.------- .-------------.---- .---------.
35
--.
--
---
-
--------------------------------------...............
...
30
--
M s
-.
-
-------.-0W%
f
...
f
.......
---...
..
. .
------...
25
~
20
15
10
~
-
~
IS
------.------...
.....
----.......
---........-----------....
-~~~~~
----.
5 mose
f
.........
---
-
Se---
~
--- 1iiiSkA mt-----------
\.~\\\\\\\
f
---------------.....
..
f
f
f
///ffafff
0
JOYSTICK
KEYBOARD
TRACKBALL
FIGURE 4.25
p<0.001
Joystick
more than Keyboard:
Keyboard
more than Trackball: not sig.
Trackball less than Joystick:
p<0.001
A------
-
90
Chapter 4.5.2
The Effect of Movement Direction on Workload:
Movement
direction
had
joytick and keyboard, but not
an
effect on subjective workload for the
for
the
trackball.
Appendix 5 contains
the z statistics between block conditions, holding the
device constant.
Joystick data shows that greater workload was imposed for
all
diagonal
block conditions, with two of the four differences significant (p<0.01).
Similarly, all diagonal blocks were greater for the keyboard, with three
of four blocks meeting the p<0.001 criteria.
Chapter 4.5.3
As
The Effect of Target ID on Workload:
evidenced by Apendix 7, workload is generally less for easy
targets than
for
hard,
but most differences are not significant.
only significant effect is for
the
keyboard
ID
The
- diagonal target blocks,
with p<0.05.
Chapter 4.5.4
The
The Effect of Memory Set Size on Workload:
size
of
the
memory
set was a
determination of workload for all block
significant
conditions
and
factor
in
devices.
the
The
mean results are presented in Appendix 1, and the z scores are listed in
Appendix
7.
Most of the differences are significant to at
p<0.05 level.
function of the
least
the
In summary, it appears that workload can be a significant
input
device, movement direction, and memory set size,
whereas target ID has little effect.
91
Chapter 4.6
The
The Effect of Learning on Performance:
results
obtained
in
our
significant, but it is also useful
effects
can
alter
the
experiment
Suppose,
performance
is shown to be better for one device
early trial
sessions.
device
may
block
this
to
potential
than
another
proficiently.
Then,
effect,
trends.
changes.
during
a
person
In order to
performance data averaged
each
subject
Individual
and
each
subjects
across
session
demonstrated
different learning effects, ranging from substantial improvement
appreciable
that
if only the subject
learn and initially faster.
conditions was plotted for
demonstrate learning
example,
be able to deliver better performance than with
the device that was easier
investigate
for
performance,
enough practice to use the device
using that
and
addition, consider the possibility that the
slower device could provide excellent
had
consistent
and prudent to determine if learning
conclusions.
In
were
to
very
to
Appendix 10 shows graphs and numerical data
no
for
all subjects and sessions.
In
order to
determine
performance conclusions,
if
the
each of the six subjects.
learning
Ames
The
data
first
effects
Some
reaction time
occurred.
devices
subjects
four
and/or
demonstrated
movement
the
major
was analyzed in two parts for
sessions
sessions were analyzed separately for each subject,
compared.
changed
and the last four
and the results were
significant
improvement
for
time, indicating that a learning effect
Nonetheless, for all subjects the relative performance of the
in
unaffected.
terms
of
both
reaction
time
and
movement
time
were
The levels of significance, however, sometimes dropped to a
lower value (from p<0.001 to p<0.01 or p<0.05).
92
Chapter 4.7
Summary of Results:
There were three quantities measured in
time, movement time, and subjective workload.
trackball
has
two
points
points for ranking last,
a
the
performance
to be made between the joystick and keyboard.
After subjects completed the
to rank the devices in
reaction
As demonstated above,
the most favorable qualities, and there are
versus workload tradeoffs
Assigning
experiment:
our
last experimental session, they were asked
1,2,3
order based on their overall opinions.
for ranking first,
and
summing
one for
over all
ranking
subjects,
second,
the following
scores resulted.
joystick
- 9
keyboard
- 9
trackball - 27
OVERALL DEVICE SCORE
30 ...................................................................................
25 ...........................
s 20 ......................................
...... .......
C
0
15
-------------------------------..............................................
R
E
5 ....-0
JOYSTICK
KEYBOARD
no
TRACKBALL
93
CHAPTER 5: CONCLUSION AND RECOMMENDATIONS
Chapter 5.1
Conclusions:
Significant
and subjective
conclusions concerning the interactions of performance
workload
from the ground data.
with
the various task conditions can be drawn
Detailed
results are presented in Chapter 4, and
a brief summary will be presented
sequence
here.
To
provide
a
framework and
for the results, consider the items that were varied as
shown
in Table 5.1 below.
Joystick
Keyboard
Trackball
Cardinal
/
\
Easy
Diagonal
/
\
Hard
Easy
Hard
ms - 1
CEl
CH1
DEl
DH1
ms - 4
CE4
CH4
DE4
DH4
TABLE 5.1: DESIGN PARAMETERS
Chapter 5.1.1
Summary of Findings:
Starting from the top, the choice of device resulted in significant
differences for
was
the
measured quantities.
the "best" device
movement times,
since
it
produced
Qualitatively, the trackball
the
fastest
as well as the least subjective workload.
had faster performance
reaction
and
The joystick,
numbers than the keyboard, but at the expense of
a higher perceived workload.
The effect of movement
summarize.
Generally,
direction
diagonal
is
somewhat
more
difficult to
targets resulted in longer acquistion
94
times than cardinal targets
joystick
and
considering
keyboard,
the
physical
and
and
effect
similar conditions.
was
significant
characteristics
indeed,
why
there
and
of those specific
significantly
longer
large.
for
were
conditions for both the joystick
expected,
devices.
Diagonal
the joystick and
no significant difference was noted for
Subjective workload ratings
For the
would be any difference for the
the difference was not very
target reaction times were
keyboard,
otherwise
this
There is no apparent reason
trackball,
in
the
trackball.
significantly higher for the diagonal
and
keyboard, but lacked significance
for the trackball.
As predicted by Fitts' law, "hard" targets
resulted
in
(small
and
far
longer movement times than their "easy" (large and close)
counterparts, and the results are significant for all devices.
times demonstrated no
Although
away)
average
significant
subjective
Reaction
effect as a function of target size.
workload ratings
increased
with
target
difficulty, the results were not statistically significant.
Response entropy (memory set
of
increasing
conditions.
the
reaction
-
time
1 or 4) had the predictable result
for
all
otherwise
similar
block
Generally, subjective workload also increased significantly
with memory set size.
Perhaps unexpectedly,
(to varying degrees) by the memory set.
movement time was affected
For
a
better
explanation of
this finding, see Chapter 4.4.
Chapter 5.1.2
The
Applying the Results:
ultimate
goal
paradigm as a space flight
to
design
of
the project is to execute the experimental
experiment, the results of which can be used
a better orbital workstation.
Even
if
the
space
flight
95
experiment demonstrates major
differences between the use of devices in
various block conditions, the results may only be valid for the specific
devices tested.
Nontheless,
if
substantial
ground data and flight data, there
lend
may
then
agreement is found between
be
a precedent which can
credence to the extrapolation of ground results to the
environment.
If
flight
data
weightless
is found to differ from ground data in a
consistent fashion, it may be appropriate to expect that similar devices
will demonstrate the same trends.
It will not be possible
to
have
complete information about every
conceivable effect that may be important in weightlessness to use in the
design.
as
a
Indeed,
a major function of the space
laboratory where the important aspects of human
studied.
In reality,
it
an
acceptable
factors
can
be
is undesirable and unrealistic to expect that
complete knowledge must be available
reach
station will be to serve
design.
or
even
implemented
Were that the case, few
of
in order to
the
great
have
been
discoveries
and
undertaken.
Nowhere is this more true than in the realm of spaceflight.
accomplishments
In the past twenty five years,
allegorically
progressed
has learned to crawl
then
of
history
however,
from
walk,
the
would
manned
infancy to childhood.
we
space
activities
with
space program has
Like a child who
have developed the ability to not
only reach space, but to accomplish something there.
conduct
ever
a greater degree of
Now is the time to
sophisication
by
consolidating and expanding the capability to work efficiently in space.
Be it ever so humble,
our experiment
may have results that can enhance
the efficiency of space station operations.
96
Chapter 5.2
Recommendations:
Perhaps the most obvious recommendation is to correlate ground data
with flight data to determine what
effects
the
weightless environment
has on performance and the perception of workload.
that
a
It
may
be possible
task easily completed in a normal gravity field could
be
very
difficult in weightlessness.
Of special interest is the possibilty that
the relative performance
devices may change.
due
to the type of
devices.
Using
of
motor
the
response
joystick,
for
necessary
to
No hand repositioning is required.
require the
user
to
interact
with
the
example, one simply causes a small
stick displacement and the cursor continues
released.
Such changes could be
move until the stick is
Trackball movements
may
to roll his hand over the ball several times to place
the cursor in the desired position, which may be difficult if the arm is
restrained in a support
sleeve.
moving the ball with fingertips
ground
observations
show
If the subject adopts a new stategy of
only,
that
holding the hand (or arm) still,
time
performance
will
suffer.
Manipulations of the arrow keys may also suffer from (or be enhanced by)
some effect of weightlessness.
One topic of interest is whether
degraded or enhanced in flight.
some
initial
degradation
weightlessness,
the first
was found to
(Space
be
Skylab reports
control
motor
suggest
that
performance
of
a
task
than
Physiology & Medicine, 1982).
quite
rapid,
however.
there is
upon initial exposure to
as evidenced by generally longer completion
inflight
performance
of
or not fine motor control will be
the
times
for
last preflight
Performance recovery
Other reports (Kubis, et al.
1977) suggest that fine motor control is more
impaired than gross motor
movements, with performance times returning to baseline levels after
an
97
This could be important if work schedules are very
average of ten days.
busy
during the early days of the mission, as is
necessary to
consider
the
possible
effects
likely.
It
may
be
of reduced work capacity
beforehand in order to enhance the safety and success of the mission.
APPENDIX 1
98
Appendix 1 provides a summary for the subject or group named in the
title. The summary consists of the mean and standard deviation for
reaction time and movement time for each device and block
condition.
APPENDIX 1
99
MITBigFile - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDHl
JDH4
rt mean
rt s.d.
497
6
885
25
495
6
888
23
532
8
1009
34
522
8
976
28
mt mean
mt s.d.
553
15
541
20
932
25
962
26
943
22
943
27
1824
43
1836
43
KCEl
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt s.d.
552
1012
553
941
759
1240
791
1242
9
28
8
23
13
34
14
31
mt mean
mt s.d.
1113
7
1134
11
1704
16
1729
21
1182
31
1183
30
2132
48
2093
51
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt s.d.
501
11
846
38
526
11
878
29
517
10
860
532
11
887
33
mt mean
mt s.d.
305
10
394
20
613
16
663
22
426
13
486
23
762
18
839
22
n - 360
29
100
APPENDIX 1
Subject A - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
540
869
530
990
558
956
521
1053
0
3
0
3
1
5
1
19
55
14
54
22
68
34
100
mt mean
mt var
mt s.d.
528
359
1091
973
864
851
1736
1885
1
35
1
28
6
74
6
78
2
49
2
47
6
74
11
106
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
670
1087
652
1019
929
1407
966
1326
1
24
5
69
1
25
4
60
1
29
5
69
1
38
4
63
mt mean
mt var
mt s.d.
1135
1103
1621
1598
1098
1033
1866
1923
2
0
0
0
5
2
10
13
43
21
18
7
68
48
99
114
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDHI
TDH4
rt mean
rt var
rt s.d.
467
1
35
654
842
6
77
538
1
33
1017
61
536
1
33
mt mean
mt var
mt s.d.
337
1
30
375
2
629
2
48
40
984
9
97
424
1
36
4
n - 40
10
206
513
1
33
959
6
77
587
8
90
835
2
42
802
43
2
46
APPENDIX 1
101
Subject B - SUBJECT SUMMARY
Times expressed in milliseconds.
JCEl
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
579
1
28
1118
11
103
554
0
16
1097
12
111
691
1
37
1462
17
131
642
1
28
1271
8
89
mt mean
mt var
mt s.d.
756
4
60
751
3
55
1172
8
89
1321
14
120
1367
5
68
1440
17
129
2711
41
204
2787
30
173
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
582
1
31
1203
9
95
566
0
19
1051
6
79
833
2
48
1408
10
100
716
2
41
1516
14
118
mt mean
mt var
mt s.d.
1133
0
20
1121
0
21
1771
3
54
1795
2
46
1496
18
134
1414
11
103
2359
28
166
2201
19
138
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
572
2
846
560
1167
551
1047
6
1
36
11
103
1272
12
77
22
148
636
1
43
1
32
34
109
mt mean
mt var
mt s.d.
358
1
36
531
685
2
814
6
495
2
530
69
49
77
42
76
861
7
82
994
5
n - 40
6
6
78
APPENDIX 1
102
Subject C - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
528
886
520
984
570
1058
561
1112
1
24
5
73
0
15
5
70
0
21
10
98
0
17
9
93
mt mean
mt var
mt s.d.
656
1
30
570
1
31
1272
6
74
1014
4
60
1141
2
41
1049
1
37
2193
13
115
2305
10
99
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
671
1
37
1368
15
123
670
1
28
1242
11
107
1012
3
54
1692
15
124
1074
1
31
1526
9
93
mt mean
mt var
mt s.d.
1099
0
11
1083
0
0
1747
2
49
1694
1
26
1123
3
58
1204
12
108
2375
20
142
2319
23
150
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
495
1
29
919
8
88
517
1
26
1084
10
99
558
0
20
892
7
83
522
0
19
870
6
78
390
1
26
463
5
69
837
1
37
779
2
569
2
585
2
1035
5
40
39
43
916
0
21
rt mean
rt
rt
var
s.d.
mt mean
mt var
mt s.d.
n - 40
69
APPENDIX 1
103
Subject D - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
526
862
502
902
514
819
526
790
0
14
3
55
0
15
5
72
0
15
4
60
0
13
3
53
mt mean
mt var
mt s.d.
547
558
893
845
1007
980
1819
1858
1
35
7
84
5
69
5
67
4
64
4
60
7
86
4
65
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
512
0
12
982
5
73
521
0
13
870
2
42
631
0
19
1010
3
51
766
3
59
1004
3
56
mt mean
mt var
mt s.d.
1097
0
15
1092
0
8
1683
4
60
1726
5
72
1190
19
137
1213
13
114
2326
28
167
2206
21
145
TCEl
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
403
746
3
503
739
499
785
6
483
889
6
78
mt mean
mt var
mt s.d.
383
1
36
1
26
57
352
1
37
1
28
574
4
66
1
29
79
1
31
465
532
2
3
3
6
2
781
2
41
56
56
78
41
49
583
n - 40
697
104
APPENDIX 1
Subject E - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
521
815
537
824
542
783
579
874
0
2
0
3
0
2
0
3
14
44
16
57
15
45
18
58
mt mean
mt var
mt s.d.
586
3
54
551
1
38
837
3
55
949
4
62
835
1
36
862
1
34
1640
8
88
1623
8
89
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
563
0
18
941
3
59
513
0
14
844
1
38
777
1
35
1261
9
95
840
1
33
1246
4
66
mt mean
mt var
mt s.d.
1109
0
15
1129
0
22
1629
0
17
1626
0
18
971
0
15
1000
1
32
1771
8
92
1882
13
113
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
604
0
20
1009
9
96
601
0
17
1022
826
2
44
656
1
25
907
3
72
566
0
17
mt mean
mt var
mt s.d.
219
0
15
226
0
561
2
459
7
85
775
3
45
373
0
20
685
2
15
460
1
23
49
56
5
n - 40
51
APPENDIX 1
105
Subject F - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCHi
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
393
0
8
669
2
40
399
0
9
663
2
50
429
0
16
738
2
49
403
0
15
829
3
55
mt mean
mt var
mt s.d.
379
1
30
397
3
56
636
2
49
696
3
50
688
3
54
642
3
54
1306
5
72
1154
5
72
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
435
840
494
844
653
1016
655
1035
0
4
1
4
0
5
0
5
12
64
28
60
18
70
16
74
mt mean
mt var
mt s.d.
1121
0
16
1109
0
15
1651
1
32
1639
1
32
1079
5
71
1037
2
50
1618
6
77
1550
2
48
TCEl
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
430
768
6
77
489
696
452
1
4
412
1
830
1
28
35
60
24
71
717
4
63
mt mean
mt var
mt s.d.
207
1
27
251
1
36
430
476
2
3
320
2
40
51
315
1
32
n - 40
5
42
1
23
541
1
23
603
3
51
106
APPENDIX 1
Subject G - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDHl
JDH4
rt mean
rt var
rt s.d.
472
0
13
874
7
85
484
0
19
974
5
68
535
0
17
993
6
75
524
1
23
903
4
61
mt mean
mt var
mt s.d.
454
566
813
1063
794
855
1621
1485
2
4
3
6
6
8
10
7
42
62
59
74
78
88
100
81
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
509
0
20
817
4
63
507
0
21
918
4
61
655
1
28
1026
6
77
740
1
34
1122
13
114
mt mean
mt var
mt s.d.
1092
1266
1715
1814
1023
1111
2014
2341
0
4
3
7
1
4
22
55
8
64
58
83
37
62
150
235
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
478
1
971
72
459
720
527
679
525
806
rt s.d.
25
269
1
32
4
67
1
28
4
61
1
29
15
121
mt mean
286
550
740
382
1
32
482
3
57
725
2
867
6
79
mt var
mt s.d.
1
12
610
2
27
108
47
7
86
n - 40
44
APPENDIX 1
107
Subject H - SUBJECT SUMMARY
Times expressed in milliseconds.
JCEl
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
422
862
418
745
432
1024
442
852
0
10
9
95
0
11
3
58
0
13
9
97
0
20
5
68
mt mean
mt var
mt s.d.
563
2
48
569
4
64
944
7
82
1058
4
60
989
5
69
1042
9
95
1965
17
129
1998
21
144
KCE1
KCE4
KCH1
KCH4
KDEl
KDE4
KDHl
KDH4
rt mean
rt var
rt s.d.
481
0
18
961
7
84
476
0
15
821
4
62
586
0
20
1206
28
168
658
0
21
1200
14
117
mt mean
mt var
mt s.d.
1113
0
20
1168
3
55
1825
4
65
1967
14
119
1533
16
128
1555
17
130
2828
24
156
2598
24
156
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
507
807
5
70
541
853
7
85
478
1
30
739
2
491
1
33
726
3
mt mean
mt var
mt s.d.
301
1
435
3
58
649
608
2
436
2
395
2
62
43
43
47
896
8
91
812
4
1
32
24
1
30
n - 40
49
55
4
66
108
APPENDIX 1
Subject I - SUBJECT SUMMARY
Times expressed in milliseconds.
JCEl
JCE4
JCHi
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
496
1013
508
816
520
1249
501
1103
0
14
8
87
0
17
3
52
1
23
32
178
0
19
15
121
mt mean
mt var
mt s.d.
509
2
546
5
726
2
737
3
806
3
764
4
1423
6
1426
4
42
72
46
54
57
65
78
63
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
545
905
573
861
759
1132
704
1206
1
4
0
2
1
5
0
5
22
67
22
49
35
69
22
71
mt mean
mt var
mt s.d.
1116
0
19
1134
1
32
1693
1
35
1700
2
45
1126
4
63
1081
'3
56
2029
16
128
1822
25
159
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
555
1
36
898
6
77
528
2
783
2
836
8
46
926
10
100
513
3
48
525
1
34
52
89
mt mean
mt var
mt s.d.
267
0
21
367
1
39
542
1
523
1
32
374
1
26
482
4
65
705
1
33
880
35
n - 40
6
76
APPENDIX 1
109
AmesBigFile - SUBJECT SUMMARY
Times expressed in milliseconds.
JCEl
JCE4
JCHl
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt s.d.
538
7
954
38
573
10
962
29
550
10
928
26
534
8
1022
32
mt mean
mt s.d.
549
17
649
34
913
22
991
34
894
20
1022
39
1577
30
1772
45
KCEl
KCE4
KCH1
KCH4
KDEl
KDE4
KDH1
KDH4
rt mean
rt s.d.
605
10
1055
29
605
10
1113
33
776
16
1249
31
794
16
1348
39
mt mean
mt s.d.
1122
7
1179
21
1706
13
1812
33
1241
27
1324
37
2013
39
2145
53
TCE1
TCE4
TCH1
TCH4
TDEl
TDE4
TDH1
TDH4
rt mean
473
771
506
788
455
748
487
796
rt s.d.
8
23
14
22
9
23
11
mt mean
mt s.d.
380
13
604
760
913
474
14
692
29
819
1026
17
34
35
19
34
n - 384
23
110
APPENDIX 1
Subject J - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
572
0
955
3
564
0
936
4
522
0
955
3
501
0
943
4
22
53
19
63
13
55
10
61
mt mean
521
396
931
1001
739
942
1422
1677
2
2
2
7
2
11
4
12
44
42
47
84
46
106
61
109
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
574
0
16
1101
3
58
585
0
18
1103
7
81
631
0
20
1060
4
60
672
1
28
1052
2
50
mt mean
mt var
mt s.d.
1118
0
14
1173
3
53
1684
1
30
1727
5
69
1650
5
71
1641
5
70
2538
8
89
2779
26
162
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
509
856
2
46
520
898
3
435
0
16
788
2
48
487
0
19
870
4
56
0
21
mt mean
mt var
mt s.d.
262
1
23
507
6
80
704
905
8
89
520
1
37
785
6
80
794
1
37
1000
6
80
nt
var
mt s.d.
0
18
1
29
n - 64
60
APPENDIX 1
111
Subject K - SUBJECT SUMMARY
Times expressed in milliseconds.
JCEl
JCE4
JCHI
JCH4
JDEl
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
458
0
566
1
468
0
720
2
430
1
731
2
402
0
746
3
17
25
12
42
26
42
19
55
mt mean
mt var
mt s.d.
624
598
993
1099
914
942
1634
3
5
2
7
2
3
5
14
52
73
46
83
48
53
73
118
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
511
0
13
838
3
59
532
0
17
1042
7
85
554
0
16
1025
4
61
593
0
19
1118
8
88
mt mean
mt var
mt s.d.
1117
0
1233
6
1788
2
2047
18
1253
3
1364
10
1963
8
2026
6
14
77
41
135
59
99
91
79
TCEl
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
378
640
439
620
319
541
376
631
0
2
4
2
0
2
0
1
rt s.d.
16
42
63
47
19
41
13
39
mt mean
413
1
28
798
9
94
806
1
33
940
8
88
559
2
42
716
4
59
840
1
34
1167
10
100
rt mean
rt var
mt var
mt s.d.
n - 64
1851
112
APPENDIX 1
Subject L - SUBJECT SUMMARY
Times expressed in milliseconds
JCEl
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
521
0
1107
12
633
2
1107
5
559
1
1013
5
552
0
1112
7
15
110
41
68
32
73
20
82
mt mean
mt var
nt s.d.
624
1000
1040
1142
964
1220
1584
2
15
5
14
2
14
5
25
40
122
74
119
46
120
71
158
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
662
1
23
1284
10
98
645
1
26
1116
7
84
872
2
46
1433
7
84
802
2
44
1672
14
118
mt mean
mt var
mt s.d.
1185
1
32
1172
3
54
1695
1
26
1774
2
40
1176
4
63
1260
5
71
2014
9
96
2191
24
156
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
403
0
21
533
1
29
365
0
21
528
2
342
0
20
521
1
320
1
22
610
1
37
mt mean
mt var
mt s.d.
509
1
36
976
15
122
595
1
25
952
7
83
1025
44
4
1429
11
60
105
n - 64
34
1116
2
49
1957
1291
9
93
APPENDIX 1
113
Subject M - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
567
1044
604
856
629
887
586
1047
0
18
0
2
0
2
0
4
14
136
18
41
18
44
19
60
mt mean
mt var
mt s.d.
668
1
31
754
5
72
1023
2
46
1126
5
68
1070
3
50
1059
3
51
1922
6
76
1975
10
100
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
733
1
29
1085
3
56
750
1
28
1396
11
106
1010
1
36
1304
3
55
1105
2
44
1500
8
91
mt mean
mt var
mt s.d.
1111
0
17
1183
2
46
1718
1
34
1793
3
58
1219
5
70
1308
8
92
2078
9
96
2143
15
124
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
564
0
16
989
3
56
639
0
15
957
1
36
595
0
16
977
3
619
0
16
967
mt mean
mt var
mt s.d.
396
1
35
440
699
3
51
809
432
1
27
539
5
73
759
1
996
3
54
3
59
n - 64
59
34
4
64
4
67
APPENDIX 1
114
Subject N - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
611
0
19
1220
13
113
645
0
21
1293
13
113
651
1
23
1224
7
83
684
0
22
1471
15
122
mt mean
mt var
mt s.d.
447
554
799
842
948
1032
1624
1662
2
39
7
81
3
54
5
72
2
45
11
106
5
74
5
69
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
674
1
24
1155
7
85
625
0
21
1077
3
53
937
1
38
1614
9
95
925
1
27
1679
11
103
mt mean
mt var
mt s.d.
1078
0
5
1093
0
12
1655
1
26
1653
1
23
1064
3
59
1250
9
94
1667
4
66
1680
5
74
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
511
0
20
948
6
80
558
972
535
1003
6
76
566
0
22
992
4
66
mt mean
mt var
mt s.d.
380
1
29
400
693
2
604
582
750
40
43
887
5
74
2
49
1
4
27
66
2
n - 64
1
23
414
1
24
3
57
1
33
APPENDIX 1
115
Subject 0 - SUBJECT SUMMARY
Times expressed in milliseconds.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt mean
rt var
rt s.d.
499
0
16
833
3
51
522
0
14
859
2
40
506
0
14
757
3
51
480
0
11
812
1
35
mt mean
408
595
691
739
728
938
1273
1508
1
5
1
2
2
11
4
6
35
68
37
43
39
106
61
80
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt mean
rt var
rt s.d.
479
0
14
870
2
46
493
0
18
945
3
53
649
1
24
1057
3
58
666
1
30
1068
5
70
mt mean
mt var
mt s.d.
1123
0
14
1220
2
49
1695
1
33
1877
9
96
1083
3
56
1120
10
101
1817
10
102
2048
16
126
TCE1
TCE4
TCHl
TCH4
TDE1
TDE4
TDH1
TDH4
rt mean
rt var
rt s.d.
474
662
2
512
0
755
2
505
660
553
41
17
48
704
1
36
mt mean
320
1
29
501
5
69
635
1
36
789
mt var
mt s.d.
mt var
mt s.d.
0
18
4
66
n - 64
0
17
325
1
29
1
33
578
4
66
1
36
656
1
30
812
4
63
APPENDIX 2
116
Appendix 2 gives the "T" test score for a comparison of similar
block conditions, varying the device used. The numbers in the
tables correspond to the value obtained by comparing the mean,
variance and sample size of the upper device- with that of the
lower. A positive number indicates that the mean of the upper
device is higher, while negative is lower.
APPENDIX 2
117
MITBigFile - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
rt
mt
sub
JCE1
JCE4
JCH1
-5.17
-3.37
-6.07
JCH4
JDE1
JDH1
JDH4
-4.77 -17.07
-6.41
-6.27
-6.02
-4.79
-3.88
-1.63 -14.63
-33.60 -26.15 -26.40 -23.32
JDE4
1.78
1.25
1.95
1.83
1.68
1.83
1.30
0.67
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
3.71
3.49
1.99
1.70
14.73
7.95
14.74
8.47
mt
67.73
31.99
49.08
35.39
22.45
18.78
26.61
22.50
sub
-0.05
0.19
-0.51
-0.46
0.73
0.14
1.31
1.61
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
0.31
-0.85
2.60
-0.26
-1.21
-3.11
0.74
-2.24
mt
-13.72
sub
-5.18 -10.88
-8.85 -20.29 -13.04 -22.86 -20.78
-1.61
-1.42
-1.54
-1.49
-2.36
-1.95
-2.68
-2.18
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt,mt n - 360
sub
n -
45
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
1.99
2.63
3.40
APPENDIX 2
Subject A - "T"
118
TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
rt
mt
sub
JCE1
JCE4
JCH1
-4.31
-2.47
-10.94 -21.34
JCH4
JDE1
JDE4
JDH1
JDH4
-4.32
-0.37 -10.16
-4.67
-8.74
-2.31
-6.91
-8.00
-2.80
-2.71
-1.05
-0.24
0.55
0.81
2.74
1.18
0.81
1.21
1.17
0.38
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
4.81
4.69
2.84
1.82
8.98
1.79
8.97
3.69
mt
15.17
14.02
22.52
6.35
8.76
4.38
9.58
9.11
sub
-0.71
-0.43
-2.98
-1.62
-0.27
-1.22
-0.33
0.09
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-1.85
-2.62
0.17
-1.58
-0.50
0.28
-0.16
-0.75
mt
-4.11
0.28
-5.47
0.09
-7.28
-2.61 -10.53
-9.38
0.10
-0.66
-0.22
-0.34
-0.48
-0.12
-0.69
-0.42
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
sub
rt,mt n - 40
sub
n -
5
p<0.05
p<0.01
p<0.001
1.99
2.63
3.42
2.31
3.36
5.04
APPENDIX 2
119
Subject B - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
-0.09
-0.61
-0.47
0.34
-2.34
0.33
-1.51
-1.66
mt
-5.95
-6.26
-5.76
-3.69
-0.86
0.16
1.34
2.65
sub
1.25
0.20
0.94
0.95
0.17
1.28
0.68
1.33
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
0.20
2.92
0.14
-0.69
4.72
2.51
1.50
1.52
mt
18.92
8.23
14.90
10.95
7.12
6.90
8.09
7.61
0.22
0.80
-0.38
-0.56
1.93
0.72
1.02
2.28
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-0.14
-2.12
0.18
0.38
-2.71
-2.49
-0.12
0.01
mt
-5.69
-2.49
-4.80
-3.55 -10.96
-6.07
-8.43
-9.44
sub
-1.45
-1.05
-0.61
-0.53
-2.20
-2.02
-1.68
-3.25
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
sub
p<0.05
p<0.01
p<0 .0 0 1
rt,mt n - 40
1.99
2.63
3.42
sub
2.31
3.36
5.04
n -
5
APPENDIX 2
Subject C - "T"
120
TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
rt
mt
sub
JCE1
JCE4
JCH1
JCH4
JDE1
-3.25
-3.37
-4.69
-2.02
-7.61
-5.32 -10.34
-13.63 -16.47
JDE4
JDH1
JDH4
-4.02 -14.27
-3.14
0.25
-1.37
-0.99
-0.08
-0.49
-1.70
0.61
0.59
0.35
-1.61
0.32
-1.65
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
3.73
2.97
4.02
1.09
7.89
5.38
14.92
5.40
mt
24.67
8.98
14.73
19.17
7.86
5.34
10.14
7.77
0.38
-0.46
1.01
-0.19
2.62
1.26
2.40
2.53
TCEl
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-0.85
0.28
-0.08
0.82
-0.43
-1.30
-1.50
-1.99
mt
-6.59
-1.41
-5.23
-3.25 -10.04
0.11
1.24
-0.95
-0.42
-1.07
2.45
-1.12
-1.48
JCE1
JCE4
JCHi
JCH4
JDE1
JDE4
JDH1
JDH4
sub
sub
-8.21 -10.97 -10.51
5
p<O.01
p< 0 . 0 0 1
rtmt n - 40
1.99
2.63
3.42
sub
2.31
3.36
5.04
p<0 .0
n -
5
APPENDIX 2
121
Subject D - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
0.77
-1.30
-0.98
0.38
-4.81
-2.43
-3.98
-2.76
mt
-14.57
-6.32
-8.64
-8.94
-1.20
-1.81
-2.71
-2.20
1.29
3.89
1.49
1.81
3.58
2.08
0.35
-0.90
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
3.86
2.54
0.58
1.68
3.79
2.40
4.27
1.20
mt
18.16
19.32
15.14
12.54
4.89
4.94
9.50
9.32
sub
-1.03
-0.45
-0.70
-1.51
-1.77
-1.46
2.08
2.00
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-4.18
-1.48
0.05
-1.67
-0.45
-0.34
-1.29
1.05
mt
-3.24
-2.23
-3.98
-3.00
-6.38
-4.57 -11.78 -13.30
sub
-0.14
-2.45
-1.06
-0.72
-1.66
-1.01
JCE1
JCE4
JCH1
JCH4
JDE1
sub
rt,mt n - 40
sub
n -
5
-2.63
-0.80
JDH1
JDH4
,JDE4
p<0.05
p<0.01
p<0.001
1.99
2.63
3.42
2.31
3.36
5.04
APPENDIX 2
122
Subject E - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCEl
JCE4
JCHi
JCH4
JDEl
JDE4
JDH1
JDH4
rt
-1.85
-1.72
1.14
-0.29
-6.15
-4.53
-6.99
-4.23
mt
-9.33 -13.20 -13.80 -10.50
-3.49
-2.96
-1.03
-1.81
sub
2.17
-0.68
1.61
1.78
-0.01
-0.16
0.84
0.30
KCEl
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
-1.48
-0.60
-3.97
-2.18
5.40
4.14
4.48
4.07
mt
41.73
33.60
22.44
39.66
24.34
5.93
10.44
8.80
sub
-0.90
0.22
-5.35
-4.03
-0.18
-0.11
0.26
1.33
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
3.35
1.83
2.74
2.15
1.05
0.68
2.49
0.42
mt
-6.52
-7.97
-3.90
-7.39 -11.17
-4.39
-9.45
-8.10
sub
-1.23
0.25
2.60
0.86
0.38
0.29
-0.96
-1.37
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt,mt n - 40
sub
n -
5
p<0.05
p<0.01
p<0.001
1.99
2.63
3.42
2.31
3.36
5.04
APPENDIX 2
123
Subject F - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
rt
mt
sub
JCE1
JCE4
JCH1
JCH4
JDE1
-2.84
-2.27
-3.23
-2.31
-21.47 -12.37 -17.36 -15.94
JDH1
JDH4
-9.14
-3.24 -11.59
-2.23
-4.39
-5.38
-2.98
-4.58
JDE4
1.29
-0.91
1.17
2.25
2.29
2.22
2.04
2.84
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
0.16
0.72
0.10
1.74
7.89
1.85
7.15
3.27
mt
28.88
21.87
23.83
19.36
9.88
11.00
13.46
13.49
sub
-0.60
-0.21
-0.25
-0.39
-0.98
-1.74
-0.74
-1.03
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
1.28
1.14
2.50
0.43
-0.58
1.06
1.76
-1.33
mt
-4.22
-2.20
-3.25
-3.07
-5.94
-4.69 -10.17
-6.23
sub
-0.08
0.66
-0.75
-1.41
-1.03
-0.85
-1.05
-0.83
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
p<0.05
p<0.01
p<0.001
rt,mt n - 40
1.99
2.63
3.42
sub
2.31
3.36
5.04
n -
5
APPENDIX 2
Subject G - "T"
124
TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
1.59
0.53
-0.81
0.61
-3.68
-0.31
-5.20
-1.69
mt
- 5.12
-7.81
-10.90
-6.73
-2.66
-2.38
-2.18
-3.45
sub
0.91
2.10
2.43
2.97
5.26
6.61
4.83
3.60
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
0.97
-0.55
1.26
2.19
3.25
3.54
4.78
1.90
mt
28.52
5.70
14.80
8.95
13.02
7.46
8.26
5.94
2.23
0.24
1.23
0.31
1.23
4.19
1.46
1.50
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
0.23
0.34
-0.68
-2.65
-0.25
-3.24
0.03
-0.72
mt
3.38
-0.13
-2.70
-2.82
-4.87
-3.55
-8.20
-5.46
sub
2.35
-2.36
-3.30
-3.15
-6.01
-9.99
-5.56
-4.89
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
sub
p<O.05
p<o. 01
p<o.001
rt,mt n - 40
1.99
2.63
3.42
sub
2.31
3.36
5.04
n -
5
APPENDIX 2
125
Subiect H - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
-2.97
-0.78
-3.21
-0.89
-6.39
-0.94
-7.31
-2.57
mt
-10.51
-7.12
-8.44
-6.82
-3.73
-3.18
-4.26
-2.82
1.13
-0.55
1.22
-0.39
1.09
0.21
-0.49
-1.33
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
-0.69
1.41
-1.95
-0.31
2.97
2.66
4.28
3.67
mt
26.32
9.16
13.10
10.72
8.11
8.40
10.70
10.56
0.21
0.57
-0.30
-0.14
0.48
0.59
0.95
2.97
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
2.55
-0.47
3.90
1.05
1.38
-2.62
1.28
-1.44
mt
- 4.87
-1.55
-2.86
-6.11
-6.81
-6.08
-6.76
-7.47
sub
- 1.09
-0.10
-0.80
0.54
-2.01
-0.59
-0.48
-1.20
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
sub
sub
p<0.05
p<0.01
p<0.001
rt,mt n - 40
1.99
2.63
3.42
sub
2.31
3.36
5.04
n -
5
APPENDIX 2
Subject I - "T"
126
TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
1.87
0.99
-2.39
-0.65
-5.71
0.61
-6.92
-0.74
mt
- 3.27
-7.45 -16.56 -13.68
-3.76
-3.68
-4.04
-2.31
1.16
3.80
0.36
0.35
-0.84
3.14
0.62
2.80
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
0.22
0.06
0.86
1.17
4.84
1.69
3.38
3.24
mt
30.15
15.33
23.21
21.32
10.96
6.97
10.04
5.33
4.94
0.38
0.92
2.33
2.75
0.46
1.90
1.15
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
1.52
-0.99
0.39
-0.47
0.13
-1.58
0.21
-1.77
mt
5.17
-2.19
-3.17
-3.38
-6.91
-3.05
-8.44
-5.51
sub
2.60
-2.62
-1.54
-2.46
-1.57
-4.95
-2.06
-3.29
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
sub
sub
rt,mt n - 40
sub
n -
5
p<O. 0 5
p<O.01
p<o.001
1.99
2.63
3.42
2.31
3.36
5.04
APPENDIX 2
127
AmesBigFile - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
rt
mt
JCEl
JCE4
JCH1
-5.51
-2.09
-2.35
JCH4
JDE1
-3.46 -12.27
-30.57 -13.27 -30.82 -17.38 -10.29
JDE4
JDH1
JDH4
-8.01 -14.21
-6.44
-5.66
-8.81
-5.33
KCEl
KCE4
KCH1
KCH4
KDEl
KDE4
KDH1
KDH4
rt
10.55
7.67
5.95
8.17
17.58
13.08
15.81
12.29
mt
49.80
14.10
41.39
19.00
25.00
13.40
27.95
17.75
TCEl
TCE4
TCH1
TCH4
TDEl
TDE4
TDH1
TDH4
rt
-5.93
-4.12
-4.02
-4.80
-6.94
-5.22
-3.49
-5.72
mt
-7.83
-0.94
-5.28
-1.64 -17.62
JCE1
JCE4
JCH1
rt,mt n - 384
JCH4
JDE1
-6.80 -22.09 -13.21
JDE4
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
JDH1
JDH4
APPENDIX 2
128
Subject J - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
-0.07
-1.87
-0.82
-1.63
-4.58
-1.29
-5.78
-1.40
mt
-12.97
-5.50 -10.33
-5.64
-11.47 -13.53
-6.65 -10.74
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
2.81
3.32
2.56
2.07
6.68
3.55
5.53
2.36
mt
31.34
6.92
23.48
7.31
14.07
8.05
18.07
9.85
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
- 2.33
-1.42
-1.67
-0.45
-3.50
-2.30
-0.65
-0.86
mt
- 5.22
1.24
-4.09
-0.78
-3.72
-1.18
-8.79
-5.01
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt,mt n - 64
p<0.05
p<0.01
p<o.001
1.98
2.61
3.36
APPENDIX 2
129
Subject K - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
-2.40
-4.22
-3.05
-3.41
-4.07
-3.99
-7.10
-3.60
mt
-9.14
-5.98 -12.85
-5.98
-4.47
-3.77
-2.83
-1.23
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
6.33
2.73
1.43
4.35
9.49
6.61
9.20
5.08
mt
22.47
3.59
18.49
6.85
9.62
5.63
11.60
6.74
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-3.38
1.51
-0.46
-1.57
-3.44
-3.26
-1.14
-1.71
mt
-3.57
1.68
-3.29
-1.31
-5.58
-2.85
-9.85
-4.41
JCE1
JCE4
JCHl
JCH4
JDE1
JDE4
JDH1
JDH4
rt,mt n - 64
p<0.05
p<0.01
p<0.001
1.98
2.61
3.36
APPENDIX 2
130
Sublect L - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
-5.09
-1.20
-0.25
-0.08
-5.55
-3.76
-5.17
-3.89
mt
-10.92
-1.30
-8.36
-5.03
-2.72
-0.29
-3.59
-1.05
KCEl
KCE4
KCH1
KCH4
KDEl
KDE4
KDH1
KDH4
rt
8.25
7.33
8.26
6.20
10.54
10.07
9.72
8.62
mt
14.08
1.48
10.30
3.07
8.56
2.83
8.33
4.96
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-4.60
-5.04
-5.85
-7.14
-5.69
-6.10
-7.73
-5.56
mt
-2.12
-0.14
-0.16
1.81
-7.07
-1.84
-5.41
-3.63
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt,mt n - 64
p<0.05
p<0.01
p<0.001
1.98
2.61
3.36
APPENDIX 2
131
Subject M - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
5.12
-0.28
-4.42
-4.73
mt
2.64
-5.03 -12.05
KCE1
KCE4
rt
1
5.05
JDE4
JDH1
JDH4
-9.33
-5.94 -10.89
-4.16
-7.47
-1.74
-2.36
-1.27
-1.06
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
1.20
3.52
3.91
10.42
4.07
10.44
4.77
8.33
10.45
16.48
11.96
10.54
6.55
12.91
8.17
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1l
TDH4LL
rt
0.13
-0.37
1.51
1.84
-1.42
1.23
1.32
-0.90
mt
5.84
-3.50
-4.68
-5.84 -14.05
-8.12
JCE1
JCE4
JCH1
rt
-
Jt
rt,mt n - 64
-3.51 -11.18
JCH4
JDE1
JDE4
p<0.05
p<0.01
p<O.001
1.98
2.61
3.36
JDH1
JDH4
APPENDIX 2
132
Subject N - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCEl
JCE4
JCH1
JCH4
JDEl
JDE4
JDH1
JDH4
rt
-2.06
0.47
0.68
1.73
-6.50
-3.09
-6.85
-1.30
mt
-16.11
-6.59 -14.33 -10.70
-1.58
-1.54
-0.43
-0.18
KCEl
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
5.24
1.77
1.93
1.23
9.18
5.00
10.19
5.60
mt
24.00
13.72
20.39
21.36
10.22
6.09
12.37
7.59
TCEl
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-3.70
-1.97
-2.55
-2.45
-3.58
-1.96
-3.76
-3.45
mt
-1.39
-1.62
-1.59
-2.83 -10.54
-3.75 -10.82
-7.67
JCE1
JCE4
JCH1
rt,mt n - 64
JCH4
JDEl
JDE4
p<0.05
p<0.01
p<0.001
1.98
2.61
3.36
JDH1
JDH4
APPENDIX 2
133
Subject 0 - "T" TEST RESULTS
The numbers in this table represent the z statistic between block
conditions, upper device minus lower device value.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
0.93
-0.53
1.29
-1.28
-5.22
-3.88
-5.77
-3.26
mt
-19.02
-7.50 -20.14 -10.83
-5.21
-1.25
-4.59
-3.63
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
0.23
3.38
-0.74
2.65
4.93
5.96
2.43
4.61
mt
24.81
8.51
21.43
9.38
12.03
4.50
10.92
8.79
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-1.03
-2.61
-0.47
-1.69
-0.05
-1.59
1.93
-2.17
mt
-1.93
-0.97
-1.07
0.64
-8.25
-2.89
-9.09
-6.85
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt,mt n - 64
p<0.05
p<0.01
p<0.001
1.98
2.61
3.36
APPENDIX 3
134
Appendix 3 gives the "T" test score for a comparison of block
conditions, holding the device constant. The numbers in the tables
correspond to the value obtained by comparing the mean, variance
and sample size of the column heading block with that of the row
heading. A positive number indicates that the mean of the column
block is higher, while negative is lower.
APPENDIX 3
135
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Movement Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
JCE1
JCE4
JCH1
JCH4
JDEl
JDE4
JDH1
JDH4
JCEl
0.00
-0.50
13.05
13.71
14.60
12.63
28.07
28.41
JCE4
0.50
0.00
12.37
13.02
13.63
12.07
27.31
27.63
JCH1
-13.05 -12.37
0.00
0.85
0.36
0.31
18.11
18.39
JCH4
-13.71 -13.02
-0.85
0.00
-0.54
-0.51
17.33
17.61
JDEl
-14.60 -13.63
-0.36
0.54
0.00
-0.02
18.35
18.64
JDE4
-12.63 -12.07
-0.31
0.51
0.02
0.00
17.48
17.75
JDH1
-28.07 -27.31 -18.11 -17.33 -18.35 -17.48
0.00
0.20
JDH4
-28.41 -27.63 -18.39 -17.61 -18.64 -17.75
-0.20
0.00
p<O. 0 5
mt n - 360
1.96
p<O. 0
1
2.58
p<O.001
3.29
136
APPENDIX 3
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Movement Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
KCE1
0.00
1.61
34.53
28.34
2.17
2.32
20.98
19.00
KCE4
-1.61
0.00
29.55
25.34
1.46
1.55
20.21
18.32
KCH1
-34.53 -29.55
0.00
0.96 -14.98 -15.59
8.46
7.28
KCH4
-28.34 -25.34
-0.96
0.00 -14.65 -15.16
7.70
6.61
KDE1
-2.17
-1.46
14.98
14.65
0.00
0.02
16.58
15.22
KDE4
-2.32
-1.55
15.59
15.16
-0.02
0.00
16.82
15.42
KDH1
-20.98 -20.21
-8.46
-7.70 -16.58 -16.82
0.00
-0.54
KDH4
-19.00 -18.32
-7.28
-6.61 -15.22 -15.42
0.54
0.00
mt n - 360
p<O. 0 5
p<0 .0 1
p<0 .001
1.96
2.58
3.29
137
APPENDIX 3
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Movement Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
TCEl
0.00
3.96
16.60
14.89
7.45
7.35
21.99
22.01
TCE4
-3.96
0.00
8.53
9.00
1.32
3.02
13.48
14.81
TCH1
-16.60
-8.53
0.00
1.85
-9.18
-4.62
6.18
8.30
TCH4
-14.89
-9.00
-1.85
0.00
-9.30
-5.63
3.48
5.64
TDE1
-7.45
-1.32
9.18
9.30
0.00
2.30
14.99
16.08
TDE4
-7.35
-3.02
4.62
5.63
-2.30
0.00
9.52
11.17
TDH1
-21.99 -13.48
-6.18
-3.48 -14.99
-9.52
0.00
2.66
TDH4
-22.01 -14.81
-8.30
-5.64 -16.08
-11.17
-2.66
0.00
mt n - 360
p<0.05
p<0.01
p< 0 . 0 0 1
1.96
2.58
3.29
APPENDIX 3
138
AmesBigFile - "T" TEST INTERACTION RESULTS
AmesBigFile - Movement Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
JCE1
0.00
2.66
13.01
11.63
13.24
11.21
29.72
25.25
JCE4
-2.66
0.00
6.53
7.14
6.26
7.27
20.53
19.87
JCHi
-13.01
-6.53
0.00
1.94
-0.65
2.46
17.81
17.04
JCH4
-11.63
-7.14
-1.94
0.00
-2.49
0.60
12.91
13.78
JDE1
-13.24
-6.26
0.65
2.49
0.00
2.97
19.07
17.80
JDE4
-11.21
-7.27
-2.46
-0.60
-2.97
0.00
11.35
12.60
JDH1
-29.72 -20.53 -17.81 -12.91 -19.07 -11.35
0.00
3.59
JDH4
-25.25 -19.87 -17.04 -13.78 -17.80 -12.60
-3.59
0.00
mt n - 384
p<0.05
p<0.01
p<o.001
1.96
2.58
3.29
APPENDIX 3
139
AmesBigFile - "T" TEST INTERACTION RESULTS
AmesBigFile - Movement Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
KCEl
KCE4
KCH1
KCH4
KDEl
KDE4
KDH1
KDH4
KCE1
0.00
2.53
38.54
20.52
4.17
5.37
22.24
19.02
KCE4
-2.53
0.00
21.02
16.18
1.77
3.40
18.63
16.83
KCH1
-38.54 -21.02
0.00
2.99 -15.25
-9.78
7.39
8.00
KCH4
-20.52 -16.18
-2.99
0.00 -13.35
-9.90
3.93
5.32
KDE1
-4.17
-1.77
15.25
13.35
0.00
1.80
16.08
15.08
KDE4
-5.37
-3.40
9.78
9.90
-1.80
0.00
12.79
12.68
KDH1
-22.24 -18.63
-7.39
-3.93 -16.08 -12.79
0.00
1.99
KDH4
-19.02 -16.83
-8.00
-5.32 -15.08 -12.68
-1.99
0.00
mt n - 384
p<0.05
p<0.01
p<O.001
1.96
2.58
3.29
0
APPENDIX 3
140
AmesBigFile - "T" TEST INTERACTION RESULTS
AmesBigFile - Movement Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
TCEl
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
TCEl
0.00
6.02
16.77
14.60
5.02
9.70
20.90
17.88
TCE4
-6.02
0.00
3.96
6.34
-3.47
1.94
5.59
8.71
TCH1
-16.77
-3.96
0.00
3.92 -12.41
-1.96
2.37
6.89
TCH4
-14.60
-6.34
-3.92
0.00 -11.94
-4.89
-2.46
2.35
TDEl
-5.02
3.47
12.41
11.94
0.00
6.72
16.11
15.17
TDE4
-9.70
-1.94
1.96
4.89
-6.72
0.00
3.77
7.45
TDHl
-20.90
-5.59
-2.37
2.46 -16.11
-3.77
0.00
5.49
TDH4
-17.88
-8.71
-6.89
-2.35 -15.17
-7.45
-5.49
0.00
05
p<0 .0 1
p<o.001
1.96
2.58
3.29
p<o.
mt n - 384
APPENDIX 4
141
Appendix 4 gives the "T" test score for a comparison of reaction
time block conditions, holding the device constant. The numbers in
the tables correspond to the value obtained by comparing the mean,
variance and sample size of the column heading block with that of
the row heading. A positive number indicates that the mean of the.
column block is higher, while negative is lower.
APPENDIX 4
142
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Reaction Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
JCEl
0.00
14.96
-0.32
16.08
3.43
14.70
2.45
16.75
JCE4
-14.96
2.92 -13.76
2.42
JCH1
0.32
2.83
16.93
JCH4
-16.08
2.92 -14.77
2.42
JDE1
-3.43
JDE4
-14.70
JDH1
-2.45
JDH4
-16.75
0.00 -15.17
15.17
0.00
-0.08 -16.31
13.37
-3.87
16.31
3.87
0.00 -14.35
14.35
0.00
-2.92 -14.83
-2.92 -13.56
13.76
14.77
-2.83
-2.42 -16.93
rt
0.08 -13.37
n - 360
0.90
-2.42 -15.29
14.83
13.56
-0.90
15.29
0.00 -13.85
-0.75
13.85
0.00
15.65
0.75 -15.65
0.00
p<O.05
p<O.01
p<o.001
1.96
2.58
3.29
143
APPENDIX 4
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Reaction Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
KCEl
KCE4
KCH1
KCH4
KDEl
KDE4
KDH1
KDH4
KCE1
0.00
15.82
0.04
16.15
13.13
19.57
14.90
21.62
KCE4
-15.82
0.00 -15.92
-1.97
-8.19
5.19
-7.14
5.57
KCH1
-0.04
0.00
16.30
13.43
19.65
15.23
21.74
KCH4
-16.15
1.97 -16.30
0.00
-6.94
7.31
-5.71
7.90
KDE1
-13.13
8.19 -13.43
6.94
0.00
13.12
1.66
14.40
KDE4
-19.57
-5.19 -19.65
0.00 -12.23
0.06
KDH1
-14.90
7.14 -15.23
KDH4
-21.62
-5.57 -21.74
15.92
rt
n - 360
-7.31 -13.12
5.71
-1.66
-7.90 -14.40
12.23
0.00
13.42
-0.06 -13.42
0.00
p<O.05
p<0.01
p<0.001
1.96
2.58
3.29
APPENDIX 4
144
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Reaction Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
TCEl
0.00
8.66
1.64
12.14
1.10
10.22
2.01
12.61
TCE4
-8.66
0.00
-8.04
0.67
-8.32
0.27
-7.86
0.84
TCH1
-1.64
8.04
0.00
11.34
-0.62
9.51
0.40
11.79
TCH4
-12.14
-0.41 -11.09
0.20
TDE1
-1.10
8.32
0.62
11.76
0.00
9.86
1.03
12.24
TDE4
-10.22
-0.27
-9.51
0.41
-9.86
0.00
-9.31
0.60
TDH1
-2.01
7.86
-0.40
11.09
-1.03
9.31
0.00
11.55
TDH4
-12.61
-0.60 -11.55
0.00
-0.67 -11.34
-0.84 -11.79
rt
n - 360
0.00 -11.76
-0.20 -12.24
p<O.05
p<O.01
p<o.001
1.96
2.58
3.29
APPENDIX 4
145
AmesBigFile - "T" TEST INTERACTION RESULTS
AmesBigFile - Reaction Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
JCE1
0.00
10.67
2.82
14.35
0.93
14.55
-0.34
14.51
JCE4
-10.67
0.00
-9.66
-0.58 -10.72
1.34
JCH1
-2.82
9.66
0.00
JCH4
-14.35
JDE1
-0.93
JDE4
-14.55
JDH1
0.34
JDH4
-14.51
-0.16 -12.90
10.24
1.68
0.58 -12.92
10.72
2.99
-1.34 -13.25
rt
n - 384
0.16 -10.24
12.90
-1.68
0.00 -13.65
13.65
0.00
0.89 -13.74
14.37
1.19
-1.38 -13.92
12.92
-2.99
13.25
-0.89 -14.37
1.38
13.74
-1.19
13.92
0.00 -14.55
2.27
14.55
0.00
14.54
-2.27 -14.54
0.00
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
APPENDIX 4
146
AmesBigFile - "T" TEST INTERACTION RESULTS
AmesBigFile - Reaction Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
KCE
KCEl
0.00
KCE4
-14.57
KCH1
0.01
KCH4
-14.79
KDE1
-9.28
KDE4
-19.97
KDH1
-9.98
KDH4
-18.54
KCE4
KCH1
KCH4
KDE1
KDE4
KDHl
KDH4
14.57
-0.01
14.79
9.28
19.97
9.98
18.54
0.00 -14.56
1.30
-8.42
4.56
-7.79
6.01
0.00
14.77
9.24
19.95
9.94
18.52
-1.30 -14.77
0.00
-9.25
3.02
-8.68
4.62
9.25
0.00
13.73
0.81
13.66
0.00 -13.09
2.00
14.56
8.42
-9.24
-4.56 -19.95
7.79
-9.94
-6.01 -18.52
rt
n - 384
-3.02 -13.73
8.68
-0.81
-4.62 -13.66
13.09
0.00
13.15
-2.00 -13.15
0.00
p<O.05
p<0 .0 1
p<o.001
1.96
2.58
3.29
APPENDIX 4
147
AmesBigFile - "T" TEST INTERACTION RESULTS
AmesBigFile - Reaction Time
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
TCE1
0.00
12.45
2.06
13.33
-1.44
11.38
1.03
13.50
TCE4
-12.45
-0.72 -11.39
0.76
TCH1
-2.06
TCH4
-13.33
0.00 -10.09
0.00
10.09
-0.53 -10.86
1.44
12.92
3.04
TDE4
-11.38
0.72
-9.15
TDH1
-1.03
11.39
1.08
TDH4
-13.50
TDE1
-0.76 -11.04
rt
n - 384
0.53 -12.92
10.86
-3.04
0.00 -13.78
13.78
0.00
1.26 -11.88
12.22
-2.22
-0.23 -13.94
9.15
-1.08
11.04
-1.26 -12.22
0.23
11.88
2.22
13.94
0.00 -10.38
1.48
10.38
0.00
12.39
-1.48 -12.39
0.00
p<0 . 0 5
p<0 .0 1
p<0 .0 0 1
1.96
2.58
3.29
APPENDIX 5
148
Appendix 5 gives the "T" test score for a comparison of subjective
workload for block conditions, holding the device constant. The
numbers in the tables correspond to the value obtained by comparing
the mean, variance and sample size of the column heading block with
that of the row heading. A positive number indicates that the mean
of the column block is higher, while negative is lower.
149
APPENDIX 5
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Subjective Workload
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
JCE1
JCE4
JCHi
JCH4
JDE1
JDE4
JDH1
JDH4
JCE1
0.00
3.20
1.71
3.67
3.07
4.81
4.91
5.79
JCE4
-3.20
0.00
-1.55
0.62
-0.30
1.74
1.57
2.66
JCH1
-1.71
1.55
0.00
2.11
1.32
3.25
3.21
4.20
JCH4
-3.67
-0.62
-2.11
0.00
-0.93
1.09
0.89
1.98
JDE1
-3.07
0.30
-1.32
0.93
0.00
2.07
1.94
3.06
JDE4
-4.81
-1.74
-3.25
-1.09
-2.07
0.00
-0.26
0.86
JDH1
-4.91
-1.57
-3.21
-0.89
-1.94
0.26
0.00
1.18
JDH4
-5.79
-2.66
-4.20
-1.98
-3.06
-0.86
-1.18
0.00
sub
n -
45
p<0.05
p<0.01
p<o.001
1.99
2.66
3.46
APPENDIX 5
150
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Subjective Workload
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
KCE1
KCE4
KCH1
KCH4
KDEl
KDE4
KDHl
KDH4
KCEl
0.00
3.88
1.39
3.54
3.24
5.27
5.13
7.12
KCE4
-3.88
0.00
-2.34
-0.07
-0.67
1.30
1.54
3.47
KCH1
-1.39
2.34
0.00
2.13
1.65
3.64
3.65
5.62
KCH4
-3.54
0.07
-2.13
0.00
-0.57
1.32
1.56
3.44
KDEl
-3.24
0.67
-1.65
0.57
0.00
1.98
2.16
4.09
KDE4
-5.27
-1.30
-3.64
-1.32
-1.98
0.00
0.31
2.24
KDH1
-5.13
-1.54
-3.65
-1.56
-2.16
-0.31
0.00
1.86
KDH4
-7.12
-3.47
-5.62
-3.44
-4.09
-2.24
-1.86
0.00
sub
n -
45
p<0.05
p<0.01
p<0.001
1.99
2.66
3.46
APPENDIX 5
151
MITBigFile - "T" TEST INTERACTION RESULTS
MITBigFile - Subjective Workload
These numbers represent the z statistic obtained by subtracting the
row heading block from the column heading block.
TCEl
TCE4
TCH1
TCH4
TDEl
TDE4
TDH1
TDH4
TCE1
0.00
3.37
1.91
4.17
2.27
4.72
4.00
5.24
TCE4
-3.37
0.00
-1.67
0.60
-1.18
1.33
0.44
1.99
TCH1
-1.91
1.67
0.00
2.38
0.49
3.08
2.21
3.69
TCH4
-4.17
-0.60
-2.38
0.00
-1.85
0.76
-0.15
1.47
TDEl
-2.27
1.18
-0.49
1.85
0.00
2.51
1.63
3.12
TDE4
-4.72
-1.33
-3.08
-0.76
-2.51
0.00
-0.90
0.72
TDH1
-4.00
-0.44
-2.21
0.15
-1.63
0.90
0.00
1.59
TDH4
-5.24
-1.99
-3.69
-1.47
-3.12
-0.72
-1.59
0.00
p<O.05
sub
n -
45
1.99
p<0 .0
1
2.66
p<O.001
3.46
APPENDIX 6
152
MITBigFile - "T" TEST FOR INDEPENDENCE
The values in this table represent the z statistic between easy and
hard targets, calculated from easy minus hard values.
JCE1
JCH1
JCE4
JCH4
JDE1
JDH1
JDE4
JDH4
rt
0.32
-0.08
0.90
0.75
mt
-13.05
-13.02
-18.35
-17.75
sub
-1.71
-0.62
-1.94
-0.86
KCE1
KCH1
KCE4
KCH4
KDE1
KDH1
KDE4
KDH4
rt
-0.04
1.97
-1.66
-0.06
mt
-34.53
-25.34
-16.58
-15.42
sub
-1.39
0.07
-2.16
-2.24
TCE1
TCH1
TCE4
TCH4
TDE1
TDH1
TDE4
TDH4
rt
-1.64
-0.67
-1.03
-0.60
mt
-16.60
-9.00
-14.99
-11.17
-1.91
-0.60
-1.63
-0.72
sub
rt,mt n - 360
sub
n -
45
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
1.99
2.63
3.40
APPENDIX 6
153
AmesBigFile - "T" TEST FOR INDEPENDENCE
The values in this table represent the z statistic between easy and
hard targets, calculated from easy minus hard values.
JCEl
JCH1
JCE4
JCH4
JDE1
JDHl
JDE4
JDH4
rt
-2.82
-0.16
1.19
-2.27
mt
-13.01
-7.14
-19.07
-12.60
KCEl
KCH1
KCE4
KCH4
KDEl
KDH1
KDE4
KDH4
rt
0.01
-1.30
-0.81
-2.00
mt
-38.54
-16.18
-16.08
-12.68
TCE1
TCH1
TCE4
TCH4
TDE1
TDH1
TDE4
TDH4
rt
-2.06
-0.53
-2.22
-1.48
mt
-16.77
-6.34
-16.11
-7.45
rt,mt n - 384
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
APPENDIX 7
154
MITBigFile - "T" TEST FOR INDEPENDENCE
The values in this table represent the z statistic between
ms-1 and ms-4 conditions, calculated from ms-1 minus ms-4.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
-14.96
-16.31
-13.56
-15.65
mt
0.50
-0.85
0.02
-0.20
sub
-3.20
-2.11
-2.07
-1.18
KCE1
KCE4
KCH1
KCH4
KDEl
KDE4
KDH1
KDH4
rt
-15.82
-16.30
-13.12
-13.42
mt
-1.61
-0.96
-0.02
0.54
sub
-3.88
-2.13
-1.98
-1.86
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-8.66
-11.34
-9.86
-11.55
mt
-3.96
-1.85
-2.30
-2.66
sub
-3.37
-2.38
-2.51
-1.59
rt,mt n - 360
sub
n - 45
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
1.99
2.63
3.40
APPENDIX 7
155
AmesBigFile - "T" TEST FOR INDEPENDENCE
The values in this table represent the z statistic between
ms-i and ms-4 conditions, calculated from ms-1 minus ms-4.
JCE1
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
rt
-10.67
-12.90
-13.74
-14.54
mt
-2.66
-1.94
-2.97
-3.59
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
rt
-14.57
-14.77
-13.73
-13.15
mt
-2.53
-2.99
-1.80
-1.99
TCEl
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
rt
-12.45
-10.86
-11.88
-12.39
mt
-6.02
-3.92
-6.72
-5.49
rt,mt n - 384
p<O. 0 5
p<o.01
p<O.001
1.96
2.58
3.29
APPENDIX 8
156
Appendix 8 gives the mean and standard deviation for reaction time,
movement time, and subjective workload, as well as the "T" test score
for a comparison of devices. The numbers in the tables correspond to
the value obtained by comparing the mean, variance and sample size of
the upper device with that of the lower. A positive number indicates
that the mean of the upper device is higher, while negative is lower.
APPENDIX 8
157
MITBigFile - Device Summary
JOYSTICK
KEYBOARD
KEYBOARD
TRACKBALL
TRACKBAL
rt mean
rt s.d.
726
8
886
9
694
9
mt mean
mt s.d.
1067
1534
561
7
rate mean
rate s.d.
13
13
36
1
30
1
28
1
MITBigFile - "T" TEST RESULTS
JOYSTICK
rt
-13.06
mt
-24.71
sub
3.78
KEYBOARD
rt
14.92
mt
64.20
sub
1.16
TRACKBALL
rt
-2.60
mt
-32.99
sub
-5.57
JOYSTICK
p<0.05
p<0.01
p<0.001
rt,mt n - 2880
1.96
2.58
3.29
sub
1.99
2.63
3.40
n -
360
APPENDIX 8
158
Subject A - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
TRACKBALL
mean
s.d.
752
22
1007
691
33
mt mean
mt s.d.
1036
36
1422
30
622
10
1
7
1
9
rt
rt
rate mean
rate s.d.
23
24
1
Subject A - "T" TEST RESULTS
JOYSTICK
rt
-8.01
mt
-8.26
sub
2.48
KEYBOARD
rt
7.91
mt
21.08
sub
-2.44
TRACKBALL
rt
-1.54
mt
-9.53
sub
-0.85
JOYSTICK
p<0.05
p<0.01
rt,mt n - 360
1.96
2.58
3.29
sub
1.99
2.63
3.42
n -
40
p<0.001
APPENDIX 8
159
Subject B - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
TRACKBALL
rt mean
rt s.d.
927
984
32
831
33
mt mean
mt s.d.
1538
1661
43
659
26
30
22
2
33
60
rate mean
rate s.d.
40
4
3
Subject B - "T" TEST RESULTS
JOYSTICK
rt
-1.24
mt
-1.67
sub
1.79
KEYBOARD
rt
3.31
nit
20.16
sub
2.00
TRACKBALL
rt
-2.02
mt
-13.51
sub
-3.65
JOYSTICK
p<0.05
p<0.01
p<0.001
rt,mt n - 360
1.96
2.58
3.29
sub
1.99
2.63
3.42
n -
40
APPENDIX 8
160
Subject C - Device Summary
JOYSTICK
KEYBOARD
TRACKBALLL
TRACKBA
rt mean
rt s.d.
777
25
1157
35
732
25
mt mean
mt s.d.
1275
41
1580
697
20
3
1
4
rate mean
rate s.d.
42
2
0
1
Subject C - "T" TEST RESULTS
JOYSTICK
rt
-8.74
mt
-5.20
sub
-1.02
KEYBOARD
rt
9.76
mt
19.10
sub
2.14
TRACKBALL
rt
-1.26
mt
-12.57
sub
-1.12
JOYSTICK
p<0.05
p<0.01
p<0.001
rt,mt n - 360
1.96
2.58
3.29
sub
1.99
2.63
3.42
n -
40
APPENDIX 8
161
Subject D - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
rt mean
rt s.d.
680
787
20
631
21
mt mean
mt s.d.
1063
36
1567
45
546
19
54
3
38
4
43
18
rate mean
rate s.d.
3
Subject D - "T" TEST RESULTS
JOYSTICK
rt
-4.03
mt
-8.73
sub
3.08
KEYBOARD
rt
5.45
mt
20.72
sub
-0.85
TRACKBALL
rt
-1.78
mt
-12.74
sub
-2.58
JOYSTICK
p<0.05
p<0.01
p<0.001
rt,mt n - 360
1.96
2.58
3.29
sub
1.99
2.63
3.42
n -
40
APPENDIX 8
162
Subject E - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
rt mean
rt s.d.
684
16
873
23
774
20
mt mean
mt s.d.
985
1390
27
470
19
31
rate mean
rate s.d.
22
24
1
23
1
1
Subject E - "T" TEST RESULTS
JOYSTICK
rt
-6.79
mt
-9.87
sub
1.18
KEYBOARD
rt
3.23
mt
27.72
rt,mt n -
sub
-1.13
sub
n -
TRACKBALL
rt
3.49
mt
-14.34
sub
-0.28
JOYSTICK
p<0.05
p<0.01
p<0.001
360
1.96
2.58
3.29
40
1.99
2.63
3.42
163
APPENDIX 8
Subject F - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
TRACKBALL
rt mean
rt s.d.
566
746
16
21
599
20
mt mean
mt s.d.
737
26
1350
22
393
16
46
39
1
43
rate mean
rate s.d.
KEYBOARD
1
1
Subject F - "T" TEST RESULTS
JOYSTICK
rt
-6.94
mt
-17.84
sub
4.13
KEYBOARD
rt
5.05
mt
35.17
rt,mt n - 360
sub
-2.05
sub
TRACKBALL
rt
1.33
mt
-11.32
sub
-1.71
JOYSTICK
n -
40
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
1.99
2.63
3.42
APPENDIX 8
164
Subject G - Device Summary
JOYSTICK
KEYBOARD
TRACKBALLL
rt mean
rt s.d.
720
787
24
645
40
mt mean
mt s.d.
956
34
1547
46
580
25
45
3
30
2
26
2
22
rate mean
rate s.d.
TRACKBA
Subject G - "T" TEST RESULTS
JOYSTICK
rt
-2.03
mt
-10.27
sub
4.17
KEYBOARD
rt
3.03
mt
18.32
sub
1.33
TRACKBALL
rt
-1.63
mt
-8.89
sub
-5.36
JOYSTICK
p<0.05
p<0.01
p<0.001
rt,mt n - 360
1.96
2.58
3.29
sub
1.99
2.63
3.42
n -
40
165
APPENDIX 8
Subject H - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
rt mean
rt s.d.
650
24
799
33
643
mt mean
mt s.d.
1141
43
1823
51
566
23
44
43
40
2
2
2
rate mean
rate s.d.
TRACKBALL
20
Subject H - "T" TEST RESULTS
JOYSTICK
rt
-3.66
mt
-10.13
sub
0.26
KEYBOARD
rt
4.08
mt
22.31
sub
1.37
TRACKBALL
rt
-0.22
mu
-11.70
sub
-1.81
JOYSTICK
p<0.05
p<0.01
p<0.001
rt,mt n - 360
1.96
2.58
3.29
sub
1.99
2.63
3.42
n -
40
APPENDIX 8
166
Subject I - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
rt mean
rt s.d.
776
836
696
34
21
25
mt mean
mt s.d.
867
28
1463
35
517
19
59
54
3
2
44
3
rate mean
rate s.d.
TRACKBALL
Subject I - "T" TEST RESULTS
JOYSTICK
rt
-1.50
mt
-13.20
sub
1.68
KEYBOARD
rt
4.31
mt
23.79
sub
2.70
TRACKBALL
rt
-1.91
mt
-10.26
sub
-4.01
JOYSTICK
p<0.05
p<0.01
p<0.001
rt,mt n - 360
1.96
2.58
3.29
sub
1.99
2.63
3.42
n -
40
167
APPENDIX 8
AmesBigFile - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
rt mean
rt s.d.
758
9
943
10
628
7
mt mean
mt s.d.
1046
13
1568
13
708
10
AmesBigFile - "T" TEST RESULTS
JOYSTICK
rt
-13.67
mt
-27.93
KEYBOARD
rt
25.83
mt
52.16
TRACKBALL
rt
-11.47
mt
-20.48
JOYSTICK
rt,mt n - 3072
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
APPENDIX 8
168
Subject J - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
rt mean
rt s.d.
743
17
847
19
670
16
mt mean
mt s.d.
953
31
1789
38
685
24
Subject J - "T" TEST RESULTS
JOYSTICK
rt
-3.97
mt
-17.16
KEYBOARD
rt
6.99
mt
24.76
TRACKBALL
rt
-3.07
mt
-6.84
JOYSTICK
rt,mt n - 512
p<0.05
p<0.01
p<O.001
1.96
2.58
3.29
169
APPENDIX 8
Subject K - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
rt mean
rt s.d.
565
13
777
22
493
15
mt mean
mt s.d.
1082
31
1599
33
780
25
Subject K - "T" TEST RESULTS
JOYSTICK
rt
-8.36
mt
-11.35
KEYBOARD
rt
10.85
mt
19.73
TRACKBALL
rt
-3.69
mt
-7.54
JOYSTICK
rt,mt n - 512
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
APPENDIX 8
170
Subject L - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
TRACK A L
rt mean
rt s.d.
826
25
1061
30
453
11
mt mean
mt s.d.
1191
40
1558
32
987
30
Subject L - "T" TEST RESULTS
JOYSTICK
rt
-5.99
mt
-7.19
KEYBOARD
rt
18.90
mt
12.88
TRACKBALL
rt
-13.51
mt
-4.10
JOYSTICK
rt,mt n - 512
p<0.05
p<0.01
p<O.001
1.96
2.58
3.29
APPENDIX 8
Subject
171
M - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
TRACKBA L
rt mean
rt s.d.
778
22
1110
25
788
16
mt mean
mt s.d.
1200
31
1569
31
634
20
Subject M - "T" TEST RESULTS
JOYSTICK
rt
-10.06
mt
-8.44
KEYBOARD
rt
10.87
mt
24.98
TRACKBALL
rt
0.40
mt
-15.40
JOYSTICK
rt,mt n - 512
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
APPENDIX 8
172
Subject N - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
TRACK A L
rt mean
rt s.d.
975
31
1086
28
761
21
mt mean
mt s.d.
989
31
1393
23
589
18
Subject N - "T" TEST RESULTS
JOYSTICK
rt
-2.65
mt
-10.56
KEYBOARD
rt
9.30
mt
27.85
TRACKBALL
rt
-5.68
mt
-11.21
JOYSTICK
rt,mt n - 512
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
173
APPENDIX 8
Subject 0 - Device Summary
JOYSTICK
KEYBOARD
TRACKBALL
TRACKBALL
rt mean
rt s.d.
659
14
778
18
603
12
mt mean
mt s.d.
860
27
1498
33
577
20
Subject 0 - "T" TEST RESULTS
JOYSTICK
rt
-5.27
mt
-15.06
KEYBOARD
rt
7.99
mt
24.03
TRACKBALL
rt
-3.03
mt
-8.50
JOYSTICK
rt,mt n - 512
p<0.05
p<0.01
p<0.001
1.96
2.58
3.29
APPENDIX 9
174
Appendix 9 gives a summary of the individual workload ratings for
each device averaged across all sessions for the MIT group.
APPENDIX 9
175
MITBigFile - WORKLOAD RATING SUMMARY
JCEl
JCE4
JCH1
JCH4
JDE1
JDE4
JDH1
JDH4
Ment
mean
s.d.
14
3
43
4
18
3
43
4
19
3
46
4
23
4
43
4
Phys
mean
s.d.
25
4
33
4
33
4
37
4
40
4
44
5
50
5
50
5
Temp
mean
s.d.
27
36
30
39
35
41
42
46
4
4
4
5
4
5
5
5
Perf
mean
s.d.
78
72
74
3
3
63
3
58
3
68
3
67
3
3
57
4
mean
s.d.
30
4
37
4
36
4
40
4
40
4
43
4
49
5
50
4
Frus
mean
s.d.
18
29
27
31
32
41
44
46
3
4
4
4
4
5
5
5
22
2
35
3
28
3
38
4
33
3
44
4
42
3
48
4
Eff
OVER
mean
s.d.
APPENDIX 9
176
MITBigFile - WORKLOAD RATING SUMMARY
KCE1
KCE4
KCH1
KCH4
KDE1
KDE4
KDH1
KDH4
14
3
43
4
17
42
4
20
43
4
23
4
50
4
Phys
mean
s.d.
17
23
3
22
3
26
4
34
4
36
4
42
5
47
3
Temp
mean
s.d.
21
32
27
33
32
36
36
42
4
4
4
4
4
4
4
5
83
3
74
81
3
78
67
61
3
78
3
72
3
3
3
3
mean
s.d.
23
3
30
3
28
3
30
4
34
4
37
4
44
4
46
4
Frus
mean
s.d.
15
3
23
3
17
3
21
3
22
3
27
4
33
38
4
5
OVER
mean
s.d.
17
2
29
3
21
2
29
3
27
35
36
3
3
3
45
4
Ment
mean
s.d.
Perf
mean
s.d.
3
4
5
Eff
4
APPENDIX 9
177
MITBigFile - WORKLOAD RATING SUMMARY
TCE1
TCE4
TCH1
TCH4
TDE1
TDE4
TDH1
TDH4
Ment
mean
s.d.
14
3
41
4
17
3
42
4
17
3
45
4
19
3
44
4
Phys
mean
s.d.
18
3
24
3
26
3
29
4
26
4
33
4
35
4
37
4
Temp
mean
s.d.
22
28
24
32
27
32
32
36
4
4
4
4
4
4
4
5
Perf
mean
s.d.
82
73
77
73
74
68
68
67
2
3
3
3
3
3
3
3
23
31
31
36
31
34
38
39
3
4
4
4
4
3
4
4
Frus
mean
s.d.
12
2
21
3
19
3
24
22
3
26
26
3
4
3
30
4
OVER
mean
s.d.
17
2
29
3
22
2
31
24
3
34
3
30
3
Eff
mean
s.d.
3
37
3
Appendix
10
Subject A
Reaction Time
126
-------------------
I M0 ----------------ls
400
----
----------
SW I --- -------------
400
---
--
@--I
JOYTIC
KEBOR
mSESSION
EMSESS ION I
SESSION 4
2
11SESSION 5
TRAKBL
ESESSION 3
SubJect A
Movement Time
120 0
s80
400
a
JOYST I CK
SESSION 1
SESSION 4
TRACKBALL
SESSION 2
SSESSION
5
[2SESSION
3
Subject A
Reaction
Time
(ms]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
2
3
4
753
854
708
686
5
759
1
Movement
Time
(mS]
KEYBOARD TRACKBALL
1097
1108
975
981
875
721
639
557
845
692
JOYSTICK
1142
1076
1018
970
973
KEYBOARD TRACKBALL
1536
591
1409
698
1367
582
1408
666
572
1391
4
Ap peni-x
10
179
Raetinn Tim&
1296
I
C
d
s
'TICK
SESSIOH 1
MSESSION 4
KEYBOARD
TRACKBALL
SESSION 2
SESSION 5
MSESSION 3
SubjOct 9
- -.-
2 M
Movement Tipsm
- -
Ii--..
----------
r41698
See
I
...........
-
RACKBAL
JOYSTICK
SESSION 1
SESSION 4
ESSION 2
S ESSION 5
MSESSION 3
Subject B
Reaction
Time
(ms]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
1085
1026
831
845
846
Movement
Time
[ms]
KEYBOARD TRACKBALL
1006
976
1066
857
1071
816
900
783
879
725
JOYSTICK
1785
1798
1317
1311
1481
KEYBOARD TRACKBALL
1814
734
1623
663
1745
1546
646
610
640
1578
Appendix
Subject C
Reaction Time
1686
10
18 p
1400 -----------------------------------------------.-----------------
1200 -----------------
-------- ....- ..
1w e -- -------- - -0 600
---
---
.--
----
n
d400.
200
SESSION I
SESSION 4
SESSION 2
SESSION 5
QSESSION 3
Subject C
Movement Time
2866
.---------------------
N1600 -----------------
-----------------
----
11200
0
400-----
JYTICK
SESSION 1
SESSION 4
KEYBOARD
SESSION 2
SESSION 5
TRACKBALL
CSESSION 3
Subject C
Reaction
Time
[ms]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
873
761
689
763
800
Movement
Time
[ms]
KEYBOARD TRACKBALL
JOYSTICK
768
643
1373
1398
1204
1272
1127
1351
1056
1177
1200
1001
776
643
829
KEYBOARD TRACKBALL
1745
1658
1610
1401
1487
676
765
708
623
712
Appendix
Subject 0
Reaction Time
9wg
ane ----------------1
ESSI
.---------------------.---
--
SESSION 4
3
-----
SESSION
s4g
600
10
5
Subject 0
Movement
29aa
SESSION
SESSION 4
Time
SESSION 2
9 SESSION 5
(SESSION
3
Subject D
Reaction
Time
[ms]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
647
730
651
686
688
Movement
Time
[ms]
KEYBOARD TRACKBALL
846
838
763
532
741
603
747
692
617
711
JOYSTICK
1094
1135
1113
1012
962
KEYBOARD TRACKBALL
1654
772
1785
550
1566
445
1412
508
1417
454
181
SjectE
Reaction Time
1286
Appendix
10
182
1889 ------------------------------------------------- -- ------
ame ---------------ee-
-
S606.
c
a
see-----
d
s
JOYSTICK
KEYBOARD
0 SESSION 1
0 SESSION 4
SESSION 2
SESSION 5
ZSESSION 3
Subject E
Tipw
Mnt
1500
1200
TRACKBALL
-----------------
------------
-----------------
900-150
JOYSTICK
SESSION I
SESSION 4
KEYBOARD
SESSION 2
11SESSION 5
TRACKBALL
EZSESSION 3
Subject E
Reaction
Time
(ms]
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
JOYSTICK
706
690
662
687
677
Movement
Time
[ms]
KEYBOARD TRACKBALL
968
884
868
824
728
781
847
776
799
759
JOYSTICK
1204
986
921
963
852
KEYBOARD TRACKBALL
1419
1389
1349
1435
1357
597
460
467
415
409
Appendix
10
Reaction Time
900
BOB -----------------
--------------------.---.------------
N700 ---------------------
-
600 -500-----
-
-----
-
-
aC00---
SSESS ION I
SSESSION 4
SESSION 4
KEYBOA
TRACKBALL
2
ISESSION
SESSION 5
SESSION 5
M2SESSION 3
Subject F
Movement Time
I see
150
-------------
1200 -----------S goo
----
......-.....----.-.-
----
.......-......---
.-------
--.-
......-...........
260weS306-
JOYSTICK
0 SESSION 1
SESSION 4
KEYBOARD
SESSION 2
SESSION 5
TRACKBALL
2SESSION 3
Subject F
Reaction
Time
Movement
Time
[ms]
(ms]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
570
612
564
575
506
KEYBOARD TRACKBALL
862
851
726
626
666
757
601
539
622
479
JOYSTICK
891
812
657
723
604
KEYBOARD TRACKBALL
1487
450
1330
430
1316
347
1316
366
1302
372
183
Appendix 10
164
Reaction Time
1200
low0 -----------------------------------------------
am0
------------------------
------------
* 600
IC
0
n 400
d
200
JOYSTICK
M SESSION 1
C SESSION 4
KEYBOARD
MSESSION
2
CUSESSION 5
TRACKBALL
[ZSESSION 3
Subject G
Movement Time
2888
1600 --------------------
-----------------------.................
1120 -----------------
8 JOYSTICK
SESSION 1
SESSION 4
KEY9OARO
mSESSION
SESSION
2
TRACKDALL
[ZSESSION 3
5
Subject G
Reaction
Time
[ms]
Movement
Time
[ms]
SESSION 1
JOYSTICK
789
SESSION 2
722
843
608
1054
1639
500
SESSION 3
SESSION 4
764
671
741
704
762
633
872
811
1512
1550
730
486
SESSION 5
654
680
583
887
1479
489
KEYBOARD TRACKBALL
967
641
JOYSTICK
1158
KEYBOARD TRACKBALL
1555
696
Appendix 10
UbDJact H
R m.M4tn
, -
IAA9
swe -----------------
Tm
---- -------...--------
-
goo
....
---
0C4-d
s
288
JOSIK
SESSION 1
SESSION 4
KYOR8RCBL
[SESSION 3
SESSION 2
ISESSION 5
SubJect H
Movement Time
---
-
2500
FMc ------m230N ------ ----
-------------------------.--
-------.----------
15.-------------- --
SESSION 1
SESSION 4
USESSION
(ZSESSION
2
SESSION 5
3
Subject H
Reaction
Time
[ms]
SESSION
SESSION
SESSION
SESSION
I SESSION
1
2
3
4
5
JOYSTICK
729
542
660
704
613
Movement
Time
(ms]
KEYBOARD TRACKBALL
890
847
694
697
865
710
736
604
593
571
JOYSTICK
1450
1261
1139
887
967
KEYBOARD TRACKBALL
2063
650
1868
567
1777
563
1757
575
1651
478
18 5
Appendix 10
Subject I
6Reaction
0 ----
4
1
s
186
Time
----LI
-------------------
--
--
---
660
0
C 400
s 266
0
KEYBOARD
JOYSTICK
TRACKBALL
mSESSION
SESSION 1
SESSION 4
(22SESSION
2
SESSION 5
3
Subject I
Movement Time
2Mg
M1600 -----------------
---------------------------
166
11290 -----------------
-----------------
11200
4an
...
...
KEYBOARD
JOYSTICK
SESSION I
SESSION 4
SESSION 2
11SESSION S
.......
TRACKBAlL1
MSESSION 3
Subject I
Reaction
Time
[ms]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
685
873
758
736
827
Movement
Time
[ms]
KEYBOARD TRACKBALL
878
897
903
820
681
619
704
773
696
686
JOYSTICK
924
973
906
750
781
KEYBOARD TRACKBALL
1571
673
1468
597
467
414
435
1509
1445
1319
Appendix 10
Jui
React on
1286
137
Time
low0 -------------------------------.--------------.--
SM --
----.- .--......-
...-
600
C
40d
0 JOYSTICK
SION 4 SI1
SESSION 4
SESSION 7
KEYBOARD
TRACKSAL
SESSION 2
SESSION 5
SESSION 8
buojec;
2586
SESSION 3
SESSION 6
j
TiI
Movemet
----------------
----------------------...-
----------------
--------------.-
1856
500
Se
----JlOYSTICK
SESS ION 1
SESS ION 4
SESS ION 7
--------
KEYBOARD
TRACKBALL
K
SESSION 2
SESSION 5
SESSION 8
S ESSION 3
6
US ESSION
Subject J
Reaction
Time
[ms]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
6
7
8
780
744
782
764
716
684
745
732
Movement
Time
[ms]
KEYBOARD TRACKBALL
920
837
973
760
800
854
800
834
JOYSTICK
698
669
1307
1086
694
674
648
652
1115
909
785
667
661
901
801
725
KEYBOARD TRACKBALL
2111
2020
1921
1734
1633
1699
1571
1621
849
903
707
559
564
719
573
604
Appendix 10
188
Reaction Time
lowg
8W ..- --
.----------.......
.........--
-.
0J
-
------------R
600
ie2ee
---- ------
eC400----
JOYSTICK
M SESSION
I
SESSION 4
SESSION 7
KEYBOARD
SESSION 2
:SESSION 5
SESSION 8
TRACKBALL
SESSION 3
LSESSION 6
ec
2666
SESSION 1
SESSION 4
SESSION 7
mSESSION
SESSION
2
SESSION 3
SESSION 6
5
::SESSION 8
Subject K
Movement
Time
[ins]
Reaction
Time
[ins]
KEYBOARD TRACKBALL
JOYSTICK
919
822
821
612
551
538
1381
1102
996
1732
1588
1690
786
776
750
547
786
449
1037
1646
721
554
542
408
578
833
683
717
633
622
418
376
377
1122
1066
992
960
1596
1527
1535
1476
761
798
893
753
JOYSTICK
SESSION 1
SESSION 2
SESSION 3
620
601
671
SESSION 4
SESSION
SESSION
SESSION
SESSION
5
6
7
8
KEYBOARD TRACKBALL
AppendiX 10
SubJect L
Reactiom Time
1461
18
189
1266
ING0
0
-
SESO
SESSION 7
ESION 5
a SESSION 8
ESON
6
Moveme.tTime
466
1600 0 JOYSTICK
KEYBOARD
SESSION 4
HSESSION 7
SESSION 5
ISESSION 8
SESSION 1
SESSION 2
TRACKBALL
1SESSION 6
U SESSION
3
Subject L
Reaction
Time
(ms ]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
6
7
8
996
1003
795
783
757
613
989
670
Movement
Time
(ms]
KEYBOARD TRACKBALL
1307
1177
1165
1023
897
483
442
443
451
497
JOYSTICK
1438
1127
1022
1055
1021
957
470
1168
1029
1339
940
439
1354
397
KEYBOARD TRACKBALL
1640
976
1618
989
1553
1156
1483
835
1415
915
1523
1014
1654
1165
1580
843
Appendix 10
Reaction Time
14
1200
190
---------------------
--
-----------------
----------------
low0 -----------------s
0
JOYSTICK
M SESSION 1
SESSION 4
IBSESSION 7
KEYBOARD
TRACKBALL
SESSION 3
HSESSION 6
SESSION 2
SESSION 5
:SESSION 8
jec
M
2080
Tm
w
SESSION 1
SESSION 4
USESSION
SESSION 7
::SESSION
SESSION
2
5
8
SESSION 3
LJSESSION 6
Mike
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
6
7
8
794
830
844
719
785
707
780
760
Reaction
Movement
Time
(ins]
Time
[ins
KEYBOARD TRACKBALL
JOYSTICK
882
822
784
835
788
787
705
705
1174
1169
1163
1160
1274
1231
1285
1140
1306
1229
1106
1029
1332
890
964
1026
KEYBOARD TRACKBALL
1721
1523
1510
1577
1566
1763
1444
1449
682
594
559
559
723
658
632
663
Appendix
10
191
Subject N
Reaction Time
1400
1200
---------------------- --------......-.---...--
low
S
2am-
699
S
- --.-- -
-- -----
-
----
.
m etTm
-
1800
SESSION I
SESSION 4
SESSION 7
SESSION 2
SESSION 5
SESSION 8
T
SESSION 3
SESSION 6
uject
JOYSTICK
KEYBOARD
SESSION 7
PSESSION 8
SESSION 1
SESSION 4
SESSION 2
SESSION 5
TRACKBALL
SESSION 3
ISESSION 6
Subject N
Reaction
Time
[ms]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
6
7
8
963
1041
1068
1160
817
976
904
870
Movement
Time
(ms]
KEYBOARD TRACKBALL
1163
1104
1076
1181
1103
1101
901
758
858
773
690
732
971
714
659
986
JOYSTICK
997
KEYBOARD TRACKBALL
1454
571
1068
1407
650
1085
1424
558
960
873
1385
1371
618
536
940
1343
649
1115
870
1359
1397
569
559
Appendix 10
192
SubJect 0
Reaction Time
1206
low0 ----------------- ----------------------------
. .- .. .
-...
. -..
an --- - - - - ----
-.
C
a
J OYSTICK
KEYBOARD
mSESSION
2
JSESSION 5
SESSION 1
SESSION 4
SESSION 7
:SESSION
TRACKBALL
SESSION 3
SESSION 6
8
SubJect 0
Moyepmnt
2900
r41688
Time
----------------- ---..---
06
11200 ----------------
C O
----------------
. -----. .f - --
--.
KEYBOARD
JOYSTICK
SESSION 1
SESSION 4
SESSION 7
----------------
SESSION 2
:SESSION 5
SESSION S
TRACKBALL
SESSION 3
SESSION 6
Subject 0
Reaction
Time
[ms ]
JOYSTICK
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
SESSION
1
2
3
4
5
6
7
8
647
724
664
614
673
672
658
617
Movement
Time
[ms]
KEYBOARD TRACKBALL
JOYSTICK
654
598
634
625
622
590
516
585
984
806
1076
849
771
748
817
827
1002
866
800
757
678
732
720
673
KEYBOARD TRACKBALL
1694
1438
1381
1415
496
614
556
599
1435
506
1585
1456
1580
624
558
664
Appendix 11
Appendix 11
193
The Experimental Protocol:
The protocol developed for this experiment is perhaps more critical
than for other tasks since it will be used during space flight.
be
easy
to
use
without
an
experimenter's help, and robustly handle
possible abort contingencies in case some unforseen
experimental
session.
This
chapter
protocol by presenting essentially the
the
ground
subjects
(with minor
presented here, but the major aspects
In
keeping
with
event
disrupts the
will demonstrate the use of
same
the
user's guide refered to by
revisions).
Not
are included.
the same order as the experimental sessions
follow.
It must
occur,
every
prompt
is
It is organized in
so
it
is
easy to
established literary guidelines, the user's
guide instructions are written in the second person.
194
Appendix 11
INSTRUCTIONS FOR GRAPHIC INPUT DEVICE EXPERIMENT
The graphic input device experiment
is
designed
performance using three different devices to move a
to evaluate your
cursor on a screen.
By performing these experiments, you will help to establish
a
data set used to see which device ( joystick, keyboard, or
baseline
trackball
)
is best suited for the cursor movement task in terms of speed, accuracy,
and
ease
of
experimental
use.
The
protocol
following
for
the
is
an instructive overview of the
Graphic Input Device Experiment.
steps described below are presented
in
the
same
order
The
as they occur
during the experimental session.
Power-up and Subject Identification
The experimental session begins when the switch on the back
Grid
Compass
microcomputer is toggled to the "on" position.
initiate the "boot"
the
keys
An example of such a screen can be seen in Figure A13.1.
until
the
demonstration purposes,
Figure
A13.2.
Once
name,
highlighted
the
your
pressing the code and return
the
you
first screen of information consisting of a
In order to select the appropriate
arrow
This will
sequence, and, after approximately 40 seconds,
will be presented with
list of names.
of the
name
"DEMO"
name has been
keys
press
surrounds
has
the
the up or down
your
name.
For
been chosen as shown in
highlighted,
simultaneously.
highlighted box at the bottom of
prompting information.
box
simply
screen
select
it
by
You may also notice
which
provides
the
Appendix 11
qr
Fordyce
Jarrett
Keller
Kneller
Kromydas
McDade
Misovec
Schaffner
DEMO
Your
last
name ->
In
Figure A13.1 Initial Start-Up Screen
Adkins
Brody
Fordyce
Jarrett
Keller
Kneller
Kromydas
McDade
misovec
mcafe
Your last name
->
IUEMU
I
Figure A13.2 Selecting the Appropriate Name
195
19 C
Appendix 11
Determining the Input Device
After pressing the "code-return" keys to confirm your name you will
be presented
with a screen that tells you which device you will use for
the upcoming trials.
prompts.
trackball
Figures
A13.3
through
A13.5
For instance, Figure A13.5 shows what you
were
to connect the
the device you would use next.
show the possible
would
see
if
the
Here you are instructed
trackball and press "code-return".
You will connect the
trackball to the serial cord leading into the back of the Grid computer.
The connections will be
color
coded to make the task easy to complete.
After pressing the "code-return" keys to confirm that you have connected
the trackball you will see the
prompt shown in Figure A13.6, which lets
you know that the testing will begin
once more.
would
as soon as you press "code-return"
If the joystick were the device
follow
the
same
you
would
procedure for that device.
use
next,
you
The keyboard is an
integral part of the Grid and does not require a separate connection.
Appendix 11
The next device is the JOYSTICK.
Connect the JOYSTICK and press code-return.
Figure A13.3 Joystick Device Prompt
The next device is the KEYBOARD.
Press code-return to acknowledge.
Figure A13.4 Keyboard Device Prompt
19 7
Appendix 11
The next device is the TRACKBALL.
Connect the TRACKBALL and press code-return.
Figure A13.5 Trackball Device Prompt
Press code-return to begin testing.
Figure A13.6 Begin Testing Prompt
198
Appendix 11
1 9
Executing the trials
When you are ready to begin the timed portion of the experiment and
have
executed
the steps outlined
positioned on the
of trials.
above,
you
should
A13.7.
Figure
A13.7 lets you know that you must
for the current
trial
are also shown on the screen, you know
the first trial.
block.
Since the targets
their size and location prior to
After 5 seconds, the prompt shown in Figure A13.7 will
disappear and a screen similar to Figure A13.8 will
replace it.
shows four targets, each with a letter corresponding to
soon as
you
cursor (
the plus sign positioned at the center of the screen )
Figure
it.
As
determine which letter belongs to the memory set, move the
target box located in
the
this case the letter is "E".
into the
same direction as the memory set letter.
Figure
not
up.
light
Figure A13.10
the target box corresponding to the memory probe lights
when the cursor
is
placed
within its boundaries.
acquired, the computer will present you
with
a
up
After the target is
new
screen similar to
Figure A13.8 and you will acquire the appropriate targets for
of the trial block.
In
A13.9 shows that if the cursor is
placed in an incorrect box, the box will
that
hand
At this point you will be presented with a screen similar to
remember the letter "E"
shows
one
graphic input device to be used for the current block
that shown in Figure
A13.8
have
the
rest
Appendix 11
Mesory
set'
E
Figure A13.7 Memory Set Screen
a
c
S
E
Figure A13.8 Trial Screen
20 0
Appendix 11
B
C
S
E
Figure A13.9 Incorrect Target Acquired
a
c
S
E
Figure A13.10 Correct Target Acquired
201
202
Appendix 11
Subjective Workload Ratings
After
block,
you
you
complete
the
eighth
will be asked to give
target
your
acquisition in the trial
impressions
of
the
workload
associated with the task you just completed.
Figure A13.ll shows one of
the six workload rating screens you will see.
The cursor will initially
be positioned halfway
cursor up or down to
workload.
for
the
moved.
between
indicate
the two endpoints, and you will move the
your
judgement of the magnitude of that
The cursor is to be moved using the graphic input device used
trial block.
Once the
Figure A13.12 shows that
cursor
is
positioned
the
cursor
has
been
where you want it, confirm the
location and continue by pressing "code-return".
This
process
will be
repeated until all six subjective workload measures have been
recorded.
Upon completion of the sixth workload rating, you will see
prompt to
press
"code-return"
to
continue.
a
Then a memory set will be presented
for the next trial block and you will repeat the process.
It
Appendix 11
MENTAL UEMANU
Very High
Very Low
Press cnd-z-return to confirm
Figure A13.ll Initial
Setting for Mental Workload
MENTFAL DEI-M
Very High
-U-
Lows
to
Figure A13.12 Final Cursor Position for Mental Workload Rating
203
204
Appendix 11
Recording the Data
Once all of the
trial
blocks
have
been
devices you will be shown six screens similar to
A13.13 and A13.14.
your
trials,
those shown in Figures
These screens contain the data you generated
and they are to be photographed as a means
backup information in
the
As each screen is shown, you
event
of
providing
will photograph it (using a camera we will
on
to
the
next data
After the last screen is photographed, you will see the
shown in Figure A13.15.
during
that the data files are somehow lost.
supply) and then press "code-return" to continue
screen.
completed for the three
This lets you know that you are all
to turn the computer's power switch off to end the session.
prompt
done,
and
Appendix 11
UEM
Monday
31-Mar-56
10:49
BLOCK I
KDHl
87865567
BLOCK 2
KCE4
31342241
BLOCK 3
KOEl
66578587
BLOCK 4
KCH4
41241332
648
630
639
543
731
666
578
483
8
17
23
99
11
14
1116
1427
478
826
562
1683
1896
547
56
33
33
99
13
19
576
546
658
569
391
630
1865
593
13
10
13
99
18
5
3341
652
1829
858
961
1164
615
940
42
5
28
85
45
29
1454
1395
1415
1454
1395
1395
1434
1454
1082
1083
1083
1883
1082
1883
1883
1883
965
925
1023
926
965
965
965
945
1591
1591
1591
1591
1591
1592
1591
1591
Figure A13.13 Data for Blocks 1 - 4
UEM
31-Mar-86
3ona
1U:4v
BLOCK 5
KCH1
33241412
BLOCK 6
KDE4
86757865
BLOCK 7
KCE1
42431213
BLOCK 8
KOH4
67885756
527
435
472
411
383
336
465
545
1
3
1666
746
1142
866
596
1594
647
908
15
556
640
646
381
618
405
594
636
5
12
11
188
13
1160
713
1745
1167
861
1299
1116
1225
43
27
39
186
14
5
18
186
3
6
1591
1591
1592
1591
1591
1591
1591
1591
8
2
188
16
17
104
925
946
925
965
984
925
984
0
1083
1083
1882
1983
1883
183
182
1083
1474
1493
1454
1435
2088
1454
1395
1903
Figure A13.14 Data for Blocks 5 - 8
205
Appendix 11
This session is now complete.
Turn the computer off to exit.
Figure A13.15 The Exit Message
modm
this
Abort This
BlockO
Device
Abort This Session
Enter
Your
choice ->
Figure A13.16 The Abort Menu Form
2 0b
Appendix 11
207
Abort Contingencies
At
some
point
in the experimental session it may be necessary to
use the abort features
of
the
software.
any time by pressing the code and
escape
You can abort the session at
keys
simultaneously.
Figure
A13.16 shows the abort form menu of options available.
The
first
executing.
option
is
to redo the trial block you
This may be necessary
reason, or
forgot
the
if
you
were
are
currently
distracted
for
memory set, or any number of reasons.
select this option, the computer
will
you just began the trial block and
some
When you
take you back to the point where
give you a new memory set.
Then you
just redo the block.
The second option is to redo the device
you
are
currently using.
When this item is selected, the computer will take you back to the point
where you connected the device and resume testing from there.
The
third
option
is
to abort the device altogether.
that the remaining trial blocks
the next device, if there are
session.
This
option
would
This means
will be skipped and you will proceed to
any
only
more
be
to
be
used
done
if
in
the
the
current
device
were to
malfunction.
The last option is to abort the entire session.
you
If this
is
done,
will not perform any more trials, and the software will advance you
to the data screens to be photographed as a backup measure.
208
Appendix 12
INSTRUCTIONS FOR WORKLOAD RATING SCALES
You
are
designed to evaluate
about to take part in an experiment
various types of
computer
We are interested in
graphic input devices.
assessing both performance and your experiences resulting from different
task
conditions.
In the following
paragraphs
we
will
describe
the
techniques to be used to examine your experiences.
In the
most
general
sense we are examining the "mental workload"
incurred while in the different
task
conditions.
difficult concept to define precisely, but a
generally.
It
is
a
mental
analog
to
simple
lifting
a
50
lb
package,
difficult
not
always easy to tell which of two
workload than the
other.
Since,
by
than
tasks
be
used
to
workload
experience.
Just
inflicts
tiring than
effective
are
However, it
more
mental
as
mental
workload is
"rulers" that
efficiently and precisely measure the mental workload
resulting from different conditions.
mental
and
others.
definition,
something that occurs in the mind there are no
can
to understand
there are some tasks that
mentally more difficult or tiring to perform
is
one
physical workload.
repeatedly lifting a 100 lb package is more
repeatedly
Mental workload is a
is
to
ask
The
people to
only
effective way to assess
describe
what
feelings
they
2 9
Appendix 12
The
experience
of workload is a
particular challenge to
collect
and
feeling,
evaluate.
experiences in the different task conditions
about
the
levels
of mental workload.
usually does not provide
sufficiently
and,
as
such,
a
Simply discussing your
provides
some information
Unfortunately, such
rich
is
information
to
discussion
allow the
combination of separate individuals' experience in a careful statistical
evaluation.
task
This can cause grave problems, especially if the
conditions
are very close in the amount of mental
different
workload
they
inflict.
To overcome this
problem,
evaluate mental workload.
you
will
perceive as workload.
the
will
use a set of rating scales to
The six workload rating component scales that
be using are defined on
descriptions carefully.
in
we
Each
the
component
next
may
page.
contribute
If you have any question about
table, please ask the experimenter about it.
important that they are clear to you.
Please
to
read
the
what
you
any of the scales
It
is
extremely
210
Appendix 12
RATING SCALE DESCRIPTIONS
Title
Endpoints
Description
MENTAL DEMAND
Very High
How mentally demanding was
Very Low
the task.
Very High
How physically demanding was
Very Low
the task.
Very High
How hurried or rushed was the
Very Low
the pace of the task.
Perfect
How successful you were in
Failure
accomplishing what you were
PHYSICAL DEMAND
TEMPORAL DEMAND
PERFORMANCE
asked to do.
EFFORT
Very High
How hard did you have to
Very Low
work to accomplish your
level of performance.
FRUSTRATION
Very High
How insecure, discouraged,
Very Low
irritated and annoyed
you were.
211
Appendix 13
Technical Specifications for the Grid Compass 1109:
In
order to understand why the Grid Compass was chosen to be space
rated, it is necessary to become familiar with its features.
below are a list of the Grid's technical specifications.
Physical Characteristics
-
magnesium case
- weight : 10 lbs
( 4.5 kg )
- height : 2 in
( 5 cm)
- width
( 29 cm )
: 11.5 in
- length : 15 in
(38
cm)
Microprocessors
- Intel 8086 16 bit main microprocessor
- Intel 8087 80 bit arithmetic co-processor
Memory
- 512,000 bytes Random Access Memory (RAM)
- 384,000 bytes non-volatile bubble memory
Display Screen
- Electroluminescent flat panel
- 80 characters X 25 lines
- 320 X 240 bit mapped display
- amber color pixels
- 6 in ( 15 cm ) diagonal
- 66 Hz refresh rate
Hardware Interfaces: RS 232-C and RS422 serial
Power Requirements: 60 watts
Enumerated
Appendix 13
212
Technical Specifications for the Graphic Input Devices:
- RS232-C Serial connection
- 9600 baud rate
- 8 bit data word
-
1 stop bit
- no parity
- x value is contained in 2 eight bit words
- y value is contained in 2 eight bit words
- total data block length - 64 bits
- position update every 6.66 milliseconds
( (64 bit block) /
(9600 bits/sec) )
- built in deadband
* the graphic input devices require an external power source.
213
REFERENCES
(Albert,
Performance in a
Cursor
Factors Society -
2 6 th
[Connelly,
"The Effect of Graphic Input Devices on
Albert, A. E.
1982]
Annual Meeting - 1982.
Connelly, E. M.
1984]
- 28th Annual Meeting - 1984.
Capacity
of
1964]
Discrete
"A
1982.
Control
Model:
Alternative
An
Proceedings of the Human Factors Society
Interpretation of Fitts' Law."
[Fitts, Peterson,
Proceedings of the Human
Task."
Positioning
1984.
Fitts,
Motor
P.
M.,
Peterson, J. R.,
Responses."
Journal
of
"Information
Experimental
Psychology. 1964, 67.
[Hart, Sellers, Guthart, 1984]
Hart, S. G.,
Sellers, J. J., Guthart, G.
"The Impact of Response Selection and Response Execution
the
Subjective
Experience
of
Workload."
Factors Society - 28 h Annual Meeting - 1984.
(Hartzell, Gopher, Hart, Lee, Dunbar, 1983]
D.,
Hart,
Lee, E., Dunbar, S.
S. G.,
Impact of Memory
Load
and
Movement
Proceedings
1981)
Johnson, R. M.
Target Acquistion."
Annual Meeting - 1981.
Proceedings
Hartzell,
"The Fittsberg
Difficulty."
1981.
"An Information
of
on
of the Human
1984.
Human Factors Society - 27th Annual Meeting - 1983.
(Johnson,
Difficulty
E.
Gopher,
J.,
Law:
The
Joint
Proceedings of the
1983.
Processing
Model
the Human Factors Society -
of
2 5 th
214
REFERENCES
T.
Experimental
"An
0.
Information Processing During
Human
Analyzing
Kvalseth,
1981]
(Kvalseth,
Human
Proceedings of the
Motor
Paradigm
Control
for
Tasks."
Factors Society - 25th Annual Meeting - 1981.
1981.
[Kvalseth, 1982]
Kvalseth, T. 0.
"The Power Law of Motor Control: Some
Diverse Applications."
Proceedings of the Human Factors Society -
Annual Meeting - 1982.
1982.
[Kubis, et al.,
1977]
Kubis,
Rusnak, R., McBride, G. H.,
J. F., McLaughlin, E. J., Jackson, J. M.,
S. V.
Saxon,
Skylab Missions 2,3, and 4: Time and
Results
Biomedical
136-153.
From
2 6 th
Skylab.
"Task and Work Performance on
Motion
Study
(NASA SP-377)
- Experiment M151."
Washington, D.C. pp.
1977.
Manufacturer, GRiD Technical Specifications.
[Mehr, 1969]
Mehr, M. H.
Digital Output."
2-Axis
H.
Manual
Measurement Systems, Inc.
Rudisill, Schulze, 1984]
[Mount,
L. J.
"A
July, 1969.
Mount, F. E.,
"Cursor Controller Comparison."
Positioning Control With
Rudisill, M.,
LEMSCO #19995.
U.S. National Aeronautics and Space Administration.
Schulze,
Houston, TX.
June, 1984.
REFERENCES
[Space Physiology and
Space
Physiology
215
Medicine, 1982] Nicogossian, A. E.,
and
Medicine,
Chapter
9:
Parker, J. F.
Neurophysiology
and
Performance. (NASA SP447) Washington, D.C. pp. 156-164. 1982
[Statistics
for
Engineers,
Statistics for Engineers.
and
Scheaffer,
Staveland, L.,
Hart,
Subjective Workload Assessment."
Annual Conference on Manual Control.
[Sternberg, 1975] Sternberg,
Current
Controversies",
R. L., McClave,
J.
T.
Boston: Duxbury Press, 1982.
[Staveland, Hart, Yeh, 1985]
"Memory
1982]
S.
S.
G.,
Yeh, Y. Y.
Proceedings of the
2 1 st
1985.
"Memory
Quarterly
Scanning:
New
Findings
and
Journal of Experimental Psychology.
27, 1975.
[Yeh, Wickens, Hart, 1984]
"The
Effect
of
Varying
Yeh,
Y.
Y.,
Wickens,
[Zaleski, Moray 1985] Zaleski, M.,
Relationship
Proceedings of the
Between
2 1 st
(Zaleski,Sanderson,
Test
of
the
Moray,
Information
N.
Load
To
1985.
"Hitts' Law?
and
A Test of
Movement Precision."
Annual Conference on Manual Control. 1984.
1984]
Relationship
Zaleski,M.,
Between
Precision." Proceedings of the 20
June, 1984.
Hart, S. G.
Task Difficulty on Subjective Workload."
Appear in the 1985 Human Factors Proceedings.
the
C. D.,
Sanderson, P.
Information
"Hitts' Law?
Load
and
A
Movement
Annual Conference on Manual Control.