THE ESSENTIAL GUIDE TO WRITING RESEARCH REPORTS
SCHOOL OF PSYCHOLOGY
2
CONTENTS
GENERAL COMMENTS .......................................................................................... 3
FORMATTING YOUR REPORT ............................................................................. 4
WHAT TO INCLUDE IN THE REPORT ................................................................ 5
1. TITLE ....................................................................................................................... 5
2. ABSTRACT ............................................................................................................. 6
3. INTRODUCTION ................................................................................................... 6
4. METHODS............................................................................................................... 7
5. RESULTS ................................................................................................................. 8
6. DISCUSSION......................................................................................................... 12
7. REFERENCING.................................................................................................... 12
How to reference: In outline ............................................................................ 12
Citations in the text: The details................................................................... 13
References in the reference section: The details.............................................. 14
How to reference: Primary and secondary sources.......................................... 17
8. APPENDIX ............................................................................................................ 18
IMPORTANT NOTES ON STYLE......................................................................... 18
DEVIATIONS FROM APA STYLE ....................................................................... 20
Appendix A: Example Report ..................................................................................... 21
Printed September 2011
3
GENERAL COMMENTS
Writing laboratory reports and research papers is a fundamental part of the scientific
process. A research report should communicate what you did, why you did it, how you
did it, what you found and what you think your findings mean. It is essential to follow a
standard format, with headings, which allows the reader to extract information with
minimal fuss. There is nothing worse than wading through a poorly thought out, poorly
referenced, disorganised manuscript.
The format explained in this guide is the same as that used in most published psychology
papers, i.e., the format of the American Psychological Association (APA). The details of
how to present a report in APA style are presented in greater detail in the APA
publication manual, which is available in the library and should be your reference for
more information on report writing. The full reference for the manual is:
American Psychological Association (2001). Publication Manual of the American
Psychological Association (5th ed.). Washington, DC: American Psychological
Association.
th
APA 5 ed.
Pages x-x
At a number of points in this guide you will be provided with page numbers in the
margins, as in the example to the left. These refer to the pages in the 5th edition of
the APA manual where you can get more information about the topic being covered.
In the rare cases that our guidelines differ from that of the APA it will be clearly marked
in the “Deviations from APA style” section, and you should follow our guidelines.
An example of a well constructed report is shown in Appendix A. You should look
closely at this example when producing your own reports.
Using this guide
1. This guidebook is intended to provide a summary of the key APA recommendations
for articles. It therefore includes specific formatting conventions (e.g., size of fonts to
use) but also some suggestions on how to organise and communicate your study, be it a
small practical or a more in-depth project thesis (or, even an article). Obviously, there are
other suggestions and recommendations but please refer to this guidebook in the first
instance in cases where there are differences.
2. As usual, there may be exceptions to this guidebook for some practicals and some
projects. However, any acceptable deviations from this guidebook will clearly be stated
by the module leader and/or supervisors. It is your responsibility to be aware of these
exceptions.
3. Be aware that not all journals follow APA convention. So if you would like more
examples, please use APA journals (e.g., Journal of Experimental Psychology: General).
4. This guidebook is an evolving set of guidelines. Please always remember to use the
most recent version of the guidebook.
Printed September 2011
4
FORMATTING YOUR REPORT
th
APA 5 ed.
Pages 284-293
You must use the following formatting when producing your report:
 12-pt Times New Roman font.
 Double line spacing.
o Exceptions: Use single spacing for figure/table captions. For large
tables, 10-pt Times New Roman and single spacing is allowed to fit a
table into a single page.
 Table captions should go above tables, figure captions go below figures.
 1” (2.54 cm) margins at the top, bottom and sides of the page.
 Left alignment (this should leave a ‘ragged’ right edge).
 A header showing your student number and the page number (both right aligned).
 In all of the report except the reference section: 0.5” (1.27 cm) first line indents for
each new paragraph (you do not need to leave blank lines between paragraphs).
 In the reference section: 0.5” (1.27 cm) hanging indents.
Limitations in pages, figures, tables, etc.
It is not the purpose of this guidebook to provide prescriptions on limitations in pages,
figures, and so on. These limitations are specific to different practicals and your 3rd-year
project. Therefore, you should find out any limitations from the module leader or
from your project supervisor as soon as possible.
PLAGIARISM
Please consult your degree handbook for information.
PENALTIES
Please consult your degree handbook for information.
Printed September 2011
5
WHAT TO INCLUDE IN THE REPORT
You will usually use the following
sections when you write your practical
reports:
Title
Acknowledgements*
Contents*
Abstract
Introduction
Method
Participants
Apparatus
Materials
Design
Procedure
Data Analysis
Results
Discussion
References
Appendix
Introduction
Method
Results
Discussion
* these headings are usually for projects
The above figure is a useful way of visualising how general or specific each part of the
report should be. The narrower parts of the ‘hourglass’, the method and results sections,
should be specific to your study. The introduction should start with some general
background, then focus in, becoming more specific to your study. The discussion should
begin with the specific findings of your study, then become more general, looking at how
your findings relate to the topic area.
In what follows, each of the sections of your report is described in further detail, in the
order that you should include them.
th
APA 5 ed.
Page 10
1. TITLE
(What you did)
The title should provide a single line summary of the measures of behaviour that you
made. For example, "The effect of sleep loss on the exploratory behaviour of gerbils" is a
suitable title; "Keeping gerbils awake" is not. The title may be a question: "Does sleep
deprivation affect the exploratory behaviour of gerbils?" as opposed to "Can gerbils be
kept awake?". Thus your title should be a brief, but accurate reflection of the content of
the report.
Your title should be presented on the title page, which is page 1 of the report (see
example report in Appendix A).
Printed September 2011
6
2. ABSTRACT
th
APA 5 ed.
Page 12-15
The abstract is a summary of your report. It must contain a brief description of the
rationale behind your experiment, the methods used, the results and the conclusions.
When papers are published readers will often read an abstract to decide whether the rest
of the article is of interest to them, and in reality, an abstract may be the only piece of
information which a reader can access (e.g. from a literature search). Therefore it must
capture the essence of your work. Browsing through the abstracts of journals in the
library may be helpful to find good examples.
The abstract should be presented on a separate page of the report.
th
APA 5 ed.
Page 15-17
3. INTRODUCTION
(Why you did it)
The Introduction should present a clear account of the reasoning behind your experiment.
Once a reader has read the Introduction, he or she should feel able to predict what your
experiment will be.
Begin by providing some general background (see the ‘hourglass’ shape at the beginning
of this guide). This will provide the framework for developing the specific aims of the
experiment. It should comprise a brief review of past work in the area, including citations
of published work (see later for details about referencing), and an explanation of the
theoretical or practical reasons for doing the current study. There are a number of distinct
topics which you should consider including:
1. Describe and define the area that you wish to study, perhaps explaining why it is
interesting and/or important.
2. Describe previous work by others that is relevant to the area.
3. Explain why the previous work is inadequate. It may have methodological
problems, or perhaps there is plenty of scope for extending it, or it may simply
require replicating.
4. Show how your proposed study goes beyond and improves upon the previous
attempts to study the area. (NB Bear in mind that the practicals you will carry out
are not original research so it will obviously be difficult in practice to fulfil points
3 and 4 adequately.)
5. In the light of previous results and what you propose to do, predict the outcome of
your study. You need to state a hypothesis about the relationships you will find
(e.g. the presence of music during study will improve word learning) and a
prediction about what exactly you expect to find in terms of the measures used in
the study (e.g. participants who studied in the presence of music will recall
significantly more words than participants who studied in silence). It is often
clearest if you state the hypothesis and prediction in the final paragraph of the
introduction.
The introduction should begin on a new page of the report. The rest of the report then
follows directly from the introduction up to the discussion (i.e., do not leave blank space
after each section, simply begin the next section on the same page).
Printed September 2011
7
th
APA 5 ed.
Page 17-20
4. METHOD
(How you did it)
The Method section must contain sufficient and precise information for the reader to be
able to repeat your experiment. Avoid irrelevant details. For example, if you are giving
your participants lists of words to memorise, it is not necessary to explain that they were
seated at a desk (as opposed to standing or seated at a table) unless you were specifically
studying the effects of furniture usage on memory. However, it is important to know the
details about the word list: e.g. number of words, word type, number of letters in the
words, word frequency etc. It is also unnecessary to say that response sheets were
downloaded from Blackboard or that responses were recorded on Microsoft Excel, as we
could have chosen to hand out response sheets and note down reactions times, without it
having any effect on the outcome of the study.
Below are the subsections that you will most often need to use in your method section.
These subsections will help organise your Method section and ensure that the appropriate
information is included. Depending on the experimental details you may exclude or
combine some of these sections. Bear in mind that although some subsections are
optional, you need to ensure that there is enough detail and information for your
experiment to be reproducible and make your analyses understandable. See the report in
Appendix A for an example of how different subsections can sometimes be more
appropriate.
Participants
State how many participants were used, how they were selected and any other important
characteristics. For example: mean or median age, sex ratio, educational level. Which
characteristics are important will depend upon the task you are asking people to perform
and the kinds of conclusions you wish to draw. If you only study undergraduate students,
you may not be able to generalise your findings to elderly people. If most of your
participants are females (a common situation in psychology departments) then you may
not be able to extrapolate your findings to males. Depending on the experiment, such
details may be trivial or extremely important.
Ethical issues are important for any research with human and animal subjects. A
statement about ethics should be included in the report. Example:
“The ethics was approved by the local ethics committee at Newcastle University.”
Apparatus
Some experiments involve only pencil, paper, stop-watch etc, and so an apparatus section
may not be needed (because details like these are not critical to replicating the
experiment). But this section is required when you use more complex equipment, e.g. a
computer running a particular programme (this would not include programs like Excel,
Minitab or SPSS as they are common computer programs. See example paper in
appendix A for an example of a programme that you would include in the apparatus
section). You should describe the equipment in sufficient detail, using a figure if
necessary, to allow another researcher to construct an equivalent apparatus. If you do use
Printed September 2011
8
a figure, it should be treated in the same way as figures in the results section (see “Tables
& Figures” in the “Results” section of this guide).
Materials
Stimuli, word lists, puzzles, IQ tests, questionnaires etc are materials. This section should
describe the general criteria for how you selected the particular items which you used.
For example, if you used words as the stimuli for a memory test, as far as possible you
should describe properties like: word length, word frequency (in the English language),
part of speech (noun, verb), other psycholinguistic variables (concreteness, abstractness,
etc). If you used a set of important stimuli you can present them in a figure (see example
paper in Appendix A)
Design
The choice of statistical tests typically depends on the experimental design. You should
therefore state whether the design was: within-subject (e.g. each participant carries out all
experimental conditions), between-subject (e.g. two or more groups of participants, who
differ in some way, carry out one condition each) or mixed (e.g. two or more groups of
participants carry out two or more conditions each). In this section you should make it
clear which experimental condition was carried out by which participants and in what
order.
Procedure
This section describes how the design was actually implemented and should describe
exactly what took place during the testing session. You should include a description of
the instructions given to participants, how the independent variable was controlled and
how the dependent variable was recorded. Remember, there should be enough
information for the reader to repeat your experiment.
Data Analysis
In this section you should describe any data manipulations that were carried out before
data analysis (e.g. if you log transformed the data or used only a mean value for each
participant in the analysis, then you would say this here). You should also say how the
data were analysed (i.e. what statistical tests you used and what they were used for).
Finally, include a standard sentence, such as “An alpha value of .05 was used for all
statistical analyses.” this shows that you will only be accepting differences/relationships
as significant if they are associated with a p-value of <.05. See report in Appendix A for
an example of this section.
th
APA 5 ed.
Page 20-26
5. RESULTS
(What you found in your experiment)
This section should provide the reader with a clear, concise summary of the data you
collected (i.e. average scores on the dependent variables, across different levels of the
independent variables) and the results of any statistical analyses.
Printed September 2011
9
A good results section should comprise two things: 1) tables and/or figures, and 2) text.
Please note that tables/figures can never stand alone, but need to be described within the
text and need to be accompanied by a shorter description in the table/figure captions.
th
APA 5 ed.
Tables
Pages 147-175
Figures
Pages 177-201
Tables & Figures
Tables contain the numerical results, such as means, standard deviations in tabulated
form. The example report in Appendix A uses the appropriate format for tables, so make
sure you study this closely to understand what you are trying to achieve. You will also
find useful example tables in the APA manual.
According to the APA manual, “any type of illustration other than a table is called a
figure… …. A figure may be a chart, graph, photograph, drawing, or other depiction.” It
can also include any of these combinations. Typically, your figures will include graphs
that represent your data (e.g. box plots, histograms, scatter plots). Once again, the report
in Appendix A gives a good example of how to add different types of figures to a report.
You should use the following rules when preparing tables and figures that depict results
(e.g., graphs):
 Never repeat the same information in both a table and a figure.
 All tables/figures must be referred to and described in the text of your results (e.g.
“Table 1 shows…” / “The data are plotted in Figure 3…”). Always refer to tables
and figures by their number. Tables and figures should be labelled separately, so do
not use “Table 1, Figure 2, Table 3…”, instead use “Table 1, Figure 1, Table 2…”.
 For tables:
o Clearly label each column and row, including the appropriate units of
measurement.
o Use only horizontal lines.
o To fit large tables on a single page, you can use a minimum of 10-pt
Times New Roman font and single spacing (see Formatting above)
 For figures:
o Label both axes correctly with the appropriate units.
o Use the x axis for the independent variable and the y axis for the
dependent variable.
o Explain any symbols that are used in either a legend within the figure or, if
there are too many, in the figure caption.
 Note that figures can include multiple graphs (see Appendix A for an example).
However, it is important that all labels and illustrations (e.g., symbols for line
graphs) should be clearly visible.
Text
In the text of the results section you will both describe and statistically analyse the data.
Do not interpret the findings in the results section, you will do that in the discussion.
If you are comparing groups you should present group means or medians, usually with a
measure of variability, such as the standard error or standard deviation. This can be done
in a table if you have a number of means or in the text if you only have one or two. If you
Printed September 2011
10
are looking at the relationship between two or more variables across participants in a
single group, present a graph or scatter plot to show the relationship.
Once you have your tables/figures you need to describe the relationships in the text. You
should refer directly to the table or figure and say what it shows about the differences
between the groups or the relationship between the variables.
Whenever you carry out a statistical test it is important that you report your findings
accurately, and in a format that allows others to check the statistic that you have reported.
This usually means stating whether you found a difference, giving the direction of the
difference (if there is one), providing the statistic associated with the test itself (e.g. t in
the t-test, F in an ANOVA), the degrees of freedom and the p value. You can report the
exact p value, unless your statistics package reports “p = 0.0000” (SPSS may not record 5
digits after the comma). It is impossible to have a p value of zero, so you must report this
as “p < .0001”.
There are specific conventions used to report different tests, and examples of those that
you will use most often are shown below. Pay close attention to the format in which the
statistics are reported and the use of capitals or italics, in order to match APA style.
Note that in each of these examples the description and statistics are integrated into
the same short paragraph. You should emulate this style in your own reports. These
are only typical examples encountered in psychology articles. It is your
responsibility to determine the appropriate way to report statistics for other tests
(e.g., by looking at an APA journal article).
Common examples:
Independent samples t test
Provide the t value, degrees of freedom (df) and p value. Use the format t(“df”) = “t
value”, p = “p value”.
“An independent samples t test revealed that the performance in the end-of-course
examination was significantly higher for the group that received the positive comment (M
= 65.1, SD = 4.6) than for the group that did not receive this comment (M = 56.8, SD =
8.9), t(28) = 2.70, p = .01 (two-tailed test).”
Related samples t test
The format for the related samples t test is the same as for the independent samples t test.
“A related samples t test indicated that the performance in the end-of-course examination
was significantly higher in the course where a positive comment was received (M = 65.1,
SD = 4.6) than in the course where no positive comment was received (M = 56.8, SD =
4.6), t(14) = 2.49, p = .03 (two-tailed test).”
Analysis of variance (ANOVA)
For an ANOVA you will need to state whether you used a between subjects ANOVA or a
within subjects ANOVA. You should also make it clear what the levels of each
independent variable are. If you use a mixed ANOVA you should say which is the
Printed September 2011
11
between subjects factor and which is the within subjects factor. You then need to report
the F value, the df for the numerator (the main effect or interaction) and the df for the
denominator (the error) and the p value. Use the format F(“df numerator”, “df
denominator”) = “F value”, p = “p value”.
“The data in Table 1 were analysed using 2 x 2 ANOVA for mixed designs, with
imageability (easy to image or hard to image) as the within subjects variable and
instruction (mnemonic or no mnemonic) as the between subjects variable. There was a
statistically significant main effect of instruction, F(1, 38) = 7.20, p = .01, with those in
the mnemonic group recalling more items overall than did those in the no mnemonic
group (M = 15.65, SD = 3.97; M = 12.40, SD = 3.74, respectively). There was also a
statistically significant main effect of imageability, F(1, 38) = 145.22, p <.001, with more
items from the easily imaged list being recalled than from the hard-to-image list (M =
15.98, SD = 4.12; M = 12.08, SD = 4.48, respectively). There was no significant
interaction between instruction and imageability.”
Pearson’s correlation
Report Pearson’s r, the degrees of freedom and the p value. Use the format r(“df”) = “r
value”, p = “p value”.
“Analysis of the data displayed in the scatter plot in Figure 1 using Pearson’s r indicated
that age was significantly negatively correlated with the mean ratings of the attractiveness
of the photographs, r(38) = -.37, p = .02. The variables were thus moderately correlated,
with increases in age tending to be associated with decreases in the ratings of
attractiveness.”
Regression analysis
Report how much of the variance in the target variable was accounted for by the predictor
variables (the R² or adjusted R² value) and then report the significance, giving the
ANOVA details (in the same way as in the ANOVA description above).
“The data were analysed by multiple regression, using as predictors age, income and
gender. The regression was a rather poor fit (R2adj = 40%), but the overall relationship
was significant, F(3, 12) = 4.32, p = .04. With other variables held constant, depression
scores were negatively related to age and income, decreasing by 0.16 for every extra year
of age, and by 0.09 for every extra pound per week income. Only income was a
significant predictor of depression, t(12) = 3.18, p < .01).”
Chi-square
Report the χ² value, the degrees of freedom and the total number of observations. Use the
format χ²(“df”, N = “number of observations”) = “χ² value”.
“Analysis of the data in Table 1 using a chi-square test revealed that breaking the speed
limit was significantly associated with sex, χ²(1, N = 100) = 10.83, p = .001. The males
tended to break the speed limit, whereas the females tended not to break the speed limit.”
Printed September 2011
12
th
APA 5 ed.
Page 26
6. DISCUSSION
(What you think the data mean for psychology)
This is the section in which you interpret the findings of your experiment and discuss
what they may mean. It is important that you relate the results to the issues raised in the
Introduction. You should decide whether your data support or refute the hypothesis which
you set up in your Introduction. In the discussion, it is usual and often necessary to relate
your findings to those in published articles.
Often your results may not lead to a clear cut answer. Therefore, with the benefit of
hindsight, you should discuss what limitations of your experiment may have led to the
outcome you found. This will lead naturally to suggestions about what further
experiments should be carried out to test your hypothesis further.
It is a good idea to consider constructing your discussion in the following way:
1. Summarise the essential findings from your results section and decide whether
your predictions were met and therefore support your hypothesis. You should say
whether or not there was a significant difference between groups here, but you
should not refer to p values or exact statistics.
2. Consider the limitations of your methodology and suggest reasonable
improvements.
3. Relate your own findings to previously published findings. Here you can start
with the most similar study or studies, then look at the relevance of your findings
to the broader topic area (see the ‘hourglass’ shape at the beginning of this guide).
4. Suggest directions for future research, based on what you have found.
5. Finally, a brief conclusion may be useful, emphasising the positive points of your
study and stating its broader implications.
After the discussion, begin the reference section on a new page.
th
APA 5 ed.
Page 28
&
207-281
7. REFERENCING
When referring to published work it is important to make the source of the work clear so
that the reader can locate the work, should they wish to check it themselves, or follow up
the work in more detail. Proper use of referencing is also important in distinguishing
between your own ideas and findings, and those of others, already published. Passing off
the work of others as your own is plagiarism and, for obvious reasons, is forbidden (see
the Psychology Degree Handbook). The rest of this section of the guide describes the
conventional methods of citation and referencing, and explains how you should reference
sources that you have read directly and those you have read about in other sources.
See the example report in Appendix A for examples of appropriate citations and
referencing.
How to reference: In outline
Referencing comes in two parts:
Printed September 2011
13
(1)
A citation in the text, in brackets, immediately after the relevant work has been
described, citing the author and the year the work was published. For example:
“Rudolph and Kim (1996) found that mood improved after aerobic exercise…evidence
suggests that exercise and therapy seem to be better than exercise alone (Martinsen,
1995)”
(2)
Full reference to the work that was cited, in a section headed ‘References’ at the
end of the piece of coursework. Each reference provides full details of the work
cited, so that a reader can locate the work. For example:
Rudolph, D. L., & Kim, J. G. (1996). Mood responses to recreational sport and exercise
in a Korean sample. Journal of Social Behaviour and Personality, 11(4), 841-849.
The reference section only contains references cited in the text, so if you have read
an article but not cited it in the main text it should not be included in the reference
section.
th
APA 5 ed.
Pages 207-214
Citations in the text: The details
Single author:
“Smith (1983) compared reaction times . . .” if you read the Smith article, or “Smith
(1983, cited by Jones, 1995) compared reaction times . . .” if you read about Smith in
Jones.
“In a recent study of reaction times (Smith, 1983) . . .”.
As long as a study cannot be confused with others, it is not necessary to repeat the year
when subsequently citing it in the same paragraph. For example:
“In a recent study of reaction times, Smith (1983) described the method . . . Smith also
found . . .”.
However, you must repeat the year if you go on to discuss another author and then return
to Smith, or if you discuss Smith in another paragraph.
Two authors:
“. . . as James and Ryerson (1983) demonstrated . . .”.
“. . . as has been shown (James & Ryerson, 1983) . . .”.
Notice that when the authors are mentioned in the main text the full word ‘and’ is used,
but when they are cited in parentheses an ampersand (&) is used. When there are only
two authors you must cite both authors at every mention.
Three to Five authors:
Printed September 2011
14
Cite all authors at first mention; subsequently mention only the first, followed by “et al.”.
(“et al.” Is short for “et alii”, which is Latin for “and others”).
“William, Jones and Smith (1983) found . . .”. [first citation]
“William et al. (1983) found . . .”. [subsequent citations]
Six or more authors:
Cite only the first author, followed by “et al.”, at every citation (including the first).
Two or more works within the same parentheses:
If you make a statement like “similar results have been obtained in many studies”, it is
vital that you cite at least a few of these studies to support this assertion. To do this you
will need to cite two or more works within the same parentheses and you should use the
same order as in the reference section (see below).
Several studies have shown similar results (Bruce 1980a, 1980b; Dorow & O'Neal, 1979;
Talpers 1981) . . .
Specific parts of a source:
Always give page numbers when literally quoting text from a source (and the quotation
should be in quotation marks).
(Czapiewski & Ruby, 1978, p. 10)
(Wilmarth, 1980, chap. 3)
Web Sources:
The principle here is the same as for print-based resources: name the author and date, if
possible. If the author of a document is not identified, use the title of the document. The
date of a web page is often found at the end of the page but if none is shown write “n.d.”
for “no date”. For example:
Fredrickson (2000) or (Fredrickson, 2000)
GVU's 8th WWW user survey. (n.d.).
th
APA 5 ed.
Pages 215-231
&
examples
pages 231-281
References in the reference section: The details
In the reference section list all references, of any kind (journal articles, books, web sites
etc.), in a single, alphabetically ordered, list.
Printed September 2011
15
If there is more than one reference by the same author(s) in the same year then the year
entries (both in the reference list and the citations in the text of your work) should be
denoted by 1983a, 1983b, etc. Single-author entries precede multiple-author entries
beginning with the same first author.
Note the order of the different parts of the reference and follow the punctuation
conventions carefully.
Book:
Use the order: Author(s) (year of publication). title. place of publication: publisher. For
example:
Bernstein, T. M. (1965). The careful writer: A modern guide to English usage. New
York: Atheneum.
Strunk, W., Jr., & White, E. B. (1979). The elements of style (3rd ed.) New York:
Macmillan.
Edited book:
Letheridge, S., & Cannon, C. R. (Eds.) (1980). Bilingual education: Teaching English as
a second language. New York: Praeger.
Book chapter:
This would usually refer to a chapter in an edited book. Use the order: Author(s) (year of
publication). chapter title. “In” book editor(s) “(Ed(s))”, book title (page numbers of
chapter). place of publication: publisher.
Hartley, J. T., Harker, J. O., & Walsh, D. A. (1980). Contemporary issues and new
directions in adult development of learning and memory. In L. W. Poon (Ed.),
Aging in the 1980s: Psychological issues (pp. 239-252). Washington, DC:
American Psychological Association.
Journal article:
Use the order: Author(s) (year of publication) title of article. title of journal, volume
number (issue number if available), page numbers.
Paivio, A. (1975). Perceptual comparisons through the mind's eye. Memory & Cognition,
3, 635-647.
Becker, L. J., & Seligman, C. (1981). Welcome to the energy crisis. Journal of Social
Issues, 37(2), 1-7.
Horowitz, L. M., Post, D. L., French, R. S., Wallis, K. D., & Siegelman, E. Y. (1981).
The prototype as a construct in abnormal psychology: 2. Clarifying disagreement
in psychiatric judgments. Journal of Abnormal Psychology, 90, 575-585.
Printed September 2011
16
Journal article in press:
Corcoran, D. L., & Williamson, E. M. (in press). Unlearning learned helplessness.
Journal of Personality and Social Psychology.
Web source:
Provide a reference to the specific web page that holds the relevant information, not to a
home or menu page. Provide as much as possible of the following information, in the
order shown:







Author's name (if available) or title of document
Date of publication (i.e. date of most recent update)
Title or description of document
Title of complete work, if relevant (e.g. electronic journal or newsletter),
underlined or in italics (these are equivalent)
Further information may be necessary to describe the location of the information
(e.g. volume and page numbers for an electronic journal). If a source does not
have page numbers but has internal divisions (such as sections or paragraphs), use
these instead in your citation, making use of the abbreviations chap. and para.
(e.g. “para. 3”).
Date on which you retrieved the information (because the information on web
sites can change)
URL; i.e. the web address
Web source examples:
Article in an Internet-only journal:
Fredrickson, B. L. (2000, March 7). Cultivating positive emotions to optimize health and
well-being. Prevention & Treatment, 3, Article 0001a. Retrieved November 20,
2000, from http://journals.apa.org/prevention/ volume3/pre0030001a.html
Article in an Internet-only newsletter:
Glueckauf, R. L., Whitton, J., Baxter, J., Kain, J., Vogelgesang, S., Hudson, M., et al.
(1998, July). Videocounseling for families of rural teens with epilepsy -- Project
update. Telehealth News,2(2). Retrieved from http://www.telehealth
.net/subscribe/newslettr4a.html1
Stand-alone document, no author identified, no date:
GVU's 8th WWW user survey. (n.d.). Retrieved August 8, 2000, from
http://www.cc.gatech.edu/gvu/ usersurveys/survey1997-10/
Printed September 2011
17
Document available on university program or department Web site:
Chou, L., McClintock, R., Moretti, F., & Nix, D. H. (1993). Technology and education:
New wine in new bottles: Choosing pasts and imagining educational futures.
Retrieved August 24, 2000, from Columbia University, Institute for Learning
Technologies Web site: http://www.ilt .columbia.edu/publications/papers/
newwine1.html
How to reference: Primary and secondary sources
New knowledge - theories and findings - is generally published in journals, collectively
called the primary literature, since this is where the work first appears. An example of a
journal is the British Journal of Psychology. For work to be accepted for publication in a
journal it generally has to be deemed worthy of publication by other specialists in the
same discipline, who read the work and assess it. Each piece of work described in a
journal is called an article or a paper.
Once an article has appeared in the primary literature its contents can be disseminated
further in a number of ways: in similar, specialist articles; in broader review articles that
discuss the work carried out on some particular topic; or in a book such as a student text.
Books and review articles are examples of the secondary literature or secondary
sources, as they are second interpretations of the original work.
When citing published work it is important that your reader can judge whether you have
read the primary source (the original description of the work) or a secondary source
(someone else's description of the work). This is because in communicating with others
you have a responsibility to be clear about the level of confidence your reader can have in
the accuracy of what they are reading. Highest confidence can be had when reading what
the writer has done or thought; this means what you have done or thought. Readers will
naturally have rather less confidence in reading an account of work you have read in the
primary literature. This is because your, or anyone else’s, account, despite the best
intentions, may contain errors. Least confidence will be felt by your readers when your
description of a piece of work is gained from a secondary source, since in this case errors
may have crept into both your account of what you read, and into the secondary source’s
account of the primary source.
Here is a section from a textbook by Martin, Carlson and Buskist:
“Rudolph and Kim (1996) found that mood improved after aerobic exercise…”
If you have read the primary source - for example Rudolph and Kim (1996) in the
quotation above – then you just cite “…Rudolph and Kim (1996)…” in the text, with the
full reference to Rudolph and Kim in the ‘References’ section at the end of the piece of
work.
If you have not read Rudolph and Kim (1996) but have read the piece above on page 773
of Martin, Carlson and Buskist’s textbook and want to refer to the information about
Rudolph and Kim (1996), then you would refer to that information in your own words
and cite in the text “…Rudolph and Kim (1996, cited by Martin, Carlson & Buskist,
Printed September 2011
18
2007)…”. Note that the primary and secondary source, with dates, all come together
in the same place in the text.
When it comes to compiling the references in the reference section at the end of your
work we ask you to list only those references that you have read yourself. So, if you
have read Rudolph and Kim (1996) you add this reference to the list; if you have not read
Rudolph and Kim, but have read about it in Martin, Carlson and Buskist’s textbook then
you add the reference to Martin, Carlson and Buskist (2007) but not the reference to
Rudolph and Kim (1996). (You would only reference Martin, Carlson and Buskist (2007)
once, even if you had cited it a number of times as a secondary source.)
Do not add to the reference list works that you have read but not cited in your own
work.
If you learn about a study or theory in a lecture or handout and want to mention it in your
report, make sure you locate the same information in a textbook or (ideally) the original
article. This will not only increase your knowledge, but will allow you to provide an
appropriate reference for the study or theory. Do not attempt to cite or reference teaching
materials.
th
APA 5 ed.
Page 28
8. APPENDIX
An Appendix is not a necessary part of the report, but it can sometimes be useful.
Specifically, you may include examples of data records e.g. questionnaires, where, by
doing so, it becomes clear to the reader what questions were asked of a participant.
However, do not include examples of data record sheets, such as lists of each of the
reaction times recorded for each trial in an experiment. The logic here is that there are a
whole variety of different ways in which you could have recorded such numbers, and
showing the particular piece of paper which you happened to use in your own experiment
does not provide useful information to the reader.
th
APA 5 ed.
Pages 31-40
IMPORTANT NOTES ON STYLE
Remember, this is an objective report of what you did in an experiment, so one of the
most important things to remember is to aim for a clear report that sounds objective and
logical. The following are a few things to consider in terms of the style of your report:

Do not use footnotes.

Use formal language, rather than the informal language that one might use in
conversation. For example, use “cannot”, “is not”, “it is”, rather than “can’t”, “isn’t”,
“it’s”.

Quotes. Never use extended (paragraph length) stretches of text from another
author’s work. Use of short quotes is also discouraged and should be kept to an
absolute minimum.
Printed September 2011
19

Paragraphs. The transition from one paragraph to the next should represent a
transition from one topic or issue to the next. Make sure you use paragraphs in your
report, but do not stick to discussing one paper/article per paragraph, rather, you can
use multiple articles in the process of discussing a topic and then only move onto the
next paragraph to begin a new topic.

Be concise. Your goal is to report your experiment clearly, so going off track or
adding any material to your report that does not add any useful information will only
waste time and irritate the reader. If you have covered all the relevant points, but a
section or report still seems short, do not continue writing just for the sake of
extending the length of the work. Instead, you should double check that you have
included all of the worthwhile points and then move on.

Define all complex terms. It is quite easy to slip into the use of jargon in order to
make the report sound more impressive or to automatically start using complex
terms because you are familiar with them. However, your report should be aimed at
an ‘intelligent layperson’. This means that you can assume that they understand
statistics and the principles of science, but that they do not know anything about the
specific topic area that you are discussing. Whenever you want to use a term that the
intelligent layperson would not understand, make sure you explain what it means.

Use the active voice. It is quite common in scientific articles to read sentences that
are written using the passive voice (e.g. “the experiment was conducted in a quiet
room”), but it is now recommended that you use the active voice to communicate
more clearly (e.g. “I conducted the experiment in a quiet room”).

However… avoid overusing “I did”, “I believe”, “My experiment”, etc. It is okay to
use first person singular when it helps to make the report clear and to use the active
voice (see above), but overuse can cause the report to sound subjective. You are
likely to find that you can reasonably use the first person singular most often in the
method (e.g. “I attached the electrodes” or “I asked participants”), but should rarely
use it in the introduction or discussion (e.g. the following make the report sound
subjective, “My results”, “I believe”, “I aim to investigate”).

Use the appropriate tense. Tenses can be difficult. A useful rule of thumb is to try
and think historically. When you write up your work, your methods and results are
past events, therefore you should use the past tense, reporting what was done and
what was found. In the Introduction and Discussion, the conclusions drawn by other
experimenters are also past events and should be described accordingly. However,
the theories and models that previous workers have developed generate predictions
and ideas that are current. For example, if you want to talk about a theory developed
by Smith et al. (1974), you would discuss what was done, found and concluded in
their study in the past tense, but discuss what the theory is or predicts in the present
tense.

Distinction between results and discussion. The results section and the discussion
section serve quite different purposes. In the results section you should examine the
Printed September 2011
20
data and report your findings. In the discussion section you should discuss what your
findings tell you about your hypotheses. For example, in an experiment requiring
participants to press a button as quickly as possible in response to a visual stimulus,
you may wish to compare the reaction times of males and females. So, in your
results section you should only examine the data to establish what the findings of
your study are. For example, state that (i) males on average have faster reaction
times than females (report numbers in the correct format - see pp10-11); (ii) males
display larger standard deviations; (iii) a t test indicates that the difference is
significant etc. Then in your discussion section you should just discuss possible
interpretations of your findings. For examples, say that the reaction times in
response to a visual stimulus in males are faster because they might play more
computer games etc.
Sometimes it may be appropriate to combine the Results and Discussion sections
(into a Results and Discussion heading). For example, in qualitative analyses, you
may wish to report a finding (result) and then discuss the implication of that finding
(discussion). In studies with more than one experiment, researchers sometimes
combine these two sections and then have another section called General
Discussion to tie all the experiments together.
DEVIATIONS FROM APA STYLE
This guide is based on the APA manual, however, you may have noticed that there are a
small number of instructions in this guide that differ from the APA manual. In such cases
you should follow our guidelines rather than those of the APA manual. Below are the
instructions that you should follow, even though the APA manual may say otherwise.
 Include your student number in the header, and do not use a ‘running header’ (an
abbreviated form of the title required for APA publications)
 Title your introduction ‘Introduction’, rather than repeating the title of your report.
 Figures and tables are in the body of the text.
Printed September 2011
21
Appendix A
Example Report
The following report is adapted from:
Vuong, Q. C., & Tarr, M. J. (2006). Structural similarity and spatiotemporal noise effects on
learning dynamic novel objects. Perception, 35, 497-510.
Printed September 2011
071234567
The effect of structural similarity and spatiotemporal noise on learning dynamic novel objects
Quoc Vuong
Newcastle University
1
071234567
2
Abstract
The spatiotemporal pattern projected by a moving object is specific to that object, as it depends
on both the object’s shape and its dynamics. Previous research has shown that observers learn to
make use of this spatiotemporal signature to recognize dynamic faces and objects. In two
experiments, we assessed the extent to which the structural similarity of the objects and the
presence of spatiotemporal noise affect how these signatures are learned and subsequently used
in recognition. Observers first learned to identify novel structurally distinctive or structurally
similar objects that rotated with a particular motion. At test, each learned objects moved with its
studied motion or with a nonstudied motion. In the nonstudied motion condition we manipulated
the dynamic information. We found that changing an object’s learned motion impaired
recognition performance when three-dimensional shape was similar or when the visual input was
noisy during learning. These results are consistent with the hypothesis that observers use learned
spatiotemporal signatures and that such information becomes progressively more important as
shape information becomes less reliable.
071234567
3
Introduction
We live in a dynamic environment. The interplay between our movements relative to
other objects and illumination sources produces a continuously changing projection on our
retinas. How does our visual system make sense of this visual cacophony to recognize objects?
The conventional answer is that the visual system maps dynamic information onto structures that
do not vary over time (Marr, 1982). For example, popular theories hypothesize that objects are
represented as parts and their relations (Biederman, 1987), or as views comprised of visible
features (Tarr & Pinker, 1989).
Recently, however, several studies have underscored the need to understand how the
visual system directly uses dynamic information for recognition (e.g., Knappmeyer, Thornton, &
Bülthoff, 2003; Lander & Bruce, 2000; Liu & Cooper, 2003; Stone, 1998; Vuong & Tarr, 2004).
Many of these studies are motivated by the observation that how visible features change over
time is specific to the objects being viewed, as this change depends on both their physical
structure (shape and surface appearance) and their movements. Thus although it is well
established that the visual system recovers and refines spatial structures from dynamic visual
input (e.g., Ullman, 1984), it is plausible that the visual system also uses the dynamic pattern
produced by the movement of that object. Stone (1998) argued that this object-specific dynamic
pattern constitutes a spatiotemporal signature of the object being viewed, and can therefore
provide information that can be used for recognition, in addition to any available shape
information. Indeed, studies have shown that dynamic patterns can be used to recognize
movements (e.g., Johansson, 1973); to discriminate between male and female actors (e.g.,
Mather & Murdoch, 1994); to interpret facial expressions (e.g., Bruce & Valentine, 1988); and to
071234567
4
recognize individuals (e.g., Lander & Bruce, 2000; Knappmeyer et al., 2003) and objects (e.g.,
Liu & Cooper, 2003; Stone, 1998; Vuong & Tarr, 2004).
Given the strong evidence that observers use object-specific dynamic patterns for
recognition purposes, our goal in the present study is to investigate the conditions under which
observers learn to use these spatiotemporal signatures. This is an important issue because the
extent to which static and dynamic information are ultimately used in object recognition may
depend on how a stimulus class is learned (Wallis & Bülthoff, 1999). As highlighted above,
investigators have used a wide variety of stimuli (e.g., faces, human actions, novel shapes) and
an equally wide variety of recognition tasks (e.g., old/new discrimination, identification,
categorization) to study the role of motion in object recognition. Across these different studies,
they have consistently found that recognition performance is often affected by subtle changes to
the dynamics of the objects. For example, Stone (1998) introduced a rotation-reversal
manipulation that preserved static cues to object identity, such as three-dimensional (3D) shape
and two-dimensional (2D) image features but disrupted dynamic cues, such as the temporal
ordering of views. He reported that this manipulation impaired observers’ ability to recognize
“amoebas” rotating rigidly in depth in a complex manner (both in accuracy and response times).
Liu and Cooper (2003) subsequently reported similar costs for rotation reversal on accuracy in an
old/new discrimination task, and on response-time priming in a symmetry judgment task. In their
experiments, they used structurally distinctive novel objects rotating about the vertical axis.
In many of these studies, learning the object dynamics is an important component of the
study (either during the course of the experiment or from observers’ pre-experimental
knowledge). Thus beyond demonstrating an important role of motion in object recognition, these
previous studies also suggest that learning may shape the visual information that is ultimately
071234567
5
used in recognition. However, on this issue, investigators have not teased apart the factors that
may affect the extent to which spatiotemporal signatures are picked-up and used. Here we
examined two factors suggested by the literature in object recognition (e.g., Tarr & Bülthoff,
1995): the structural similarity between objects, and the availability of shape and motion
information. Both of these factors may make learning the objects more difficult and therefore
influence the extent to which their dynamics are used in the recognition process. That is, motion
information may be more likely to be used when objects are difficult to learn, as may be the case
when the tested objects are highly similar to each other (e.g., Williams & Hayward, 2000) or
when objects are visually degraded (e.g., Bruce & Lander, 2000).
Our goal was to examine the extent to which the rotation-reversal effect reported by
Stone (1998) and Liu and Cooper (2003) may be influenced by the difficulty of learning the
objects. For example, in Stone’s study, observers had the difficult challenge of learning
structurally similar amoebas. By comparison, in Liu and Cooper’s study, observers had the
equally difficult challenge of learning many objects (32 as compared to 4 objects in Stone’s
study) from a single exposure and without knowing that their memory for these objects would be
subsequently tested. Here we varied the difficulty of learning in two ways. First, we used either
structurally distinctive objects that were “easy” to recognize or structurally similar objects that
were “hard” to recognize (see Vuong & Tarr, 2004). Second, observers could learn either “easy”
or “hard” objects in the presence or absence of a dynamic fog that degraded both shape
(including 3D structure and 2D views) and motion information. Our working hypothesis is that
there should be a larger rotation-reversal effect when the objects are difficult to learn.
071234567
6
Method
Participants
A total of 40 observers were recruited from the Brown University community (29
females/11 males). They participated either for course credit or payment. All observers gave
informed consent and were naïve to the purposes of the study. Ethics for this study was approved
by the Ethics Committee at Brown University.
Apparatus
The experiment was run on a Windows PC using a monitor with a 1280 x 1024 pixel
resolution and a 60 Hz refresh rate. The program to present the movies and collect responses was
written in C and relied on the OpenGL 1.2 interface to the PC’s graphics hardware. Observers sat
approximately 50 cm from the monitor. At this viewing distance, each object subtended a
maximum visual angle of ~9. The dynamic fog, when present, filled the entire screen.
Responses were collected from the keyboard. The four keys used were “v”, “b”, “n”, and “m”,
which were randomly assigned to the four targets for each observer. All observers were
instructed to respond with their dominant hand.
Stimuli
Figure 1 shows the two set of novel 3D objects used in the study. These objects were a
subset of those used in our earlier study, and details of their construction can be found in that
paper (Vuong & Tarr, 2004). Each set consisted of eight objects, half of which served as targets
and half as distractors.
071234567
7
Targets
easy
Distractors
Targets
hard
Distractors
Figure 1. The set of “easy” structurally distinct and “hard” structurally similar objects used.
The first set of eight stimuli consisted of “easy” objects with structurally distinctive
shapes. Objects in this set are composed of parts that can be easily discriminated on the basis of
nonaccidental properties (e.g., straight versus curved axis of elongation; see Biederman, 1987).
In contrast, the second set of eight stimuli consisted of “hard” amoebas with structurally similar
shapes (e.g., they lacked distinctive parts or features that could be easily used as identity cues),
similar to those used in several previous studies of human object recognition (Bülthoff &
Edelman, 1992; Stone, 1998). The 3D coordinates of each object’s vertices and their associated
surface normals were imported into custom software that rendered the objects with a matte gray
surface. The objects were illuminated by several light sources. All objects were rendered against
a black background.
071234567
8
Two trajectories were used to generate 128-frame (2.8/frame) animations for both
“easy” and “hard” objects. For the first trajectory, a virtual camera was arbitrarily rotated about
the three axes controlled by a parameter t that varied from 0 to 360. When either image
sequence is presented in increasing frame order, objects appear to tumble in depth with a
coherent rotation direction. The same image sequence played in decreasing order depicts each
object rotating in the opposite direction. The animations were played at ~50 ms/frame (roughly
three screen refreshes), so that it took ~6500 ms to play one entire 360 rotation.
Finally, in some conditions, we presented objects rotating in a dynamic fog to degrade
both spatial and dynamic information. The fog consisted of a pre-computed 3D fractal noise
volume (Perlin, 1985). By presenting 2D slices of this volume on each frame, we were able to
smoothly mask random fragments of the rotating object in space and time. On trials when the
dynamic fog was presented (for both learning and test phases), we randomly selected a subset of
frames and cycled back and forth through these frames on that trial. Thus, the dynamics of the
fog was completely independent of the dynamics of the objects. Figure 2 illustrates a time
sequence of an object rotating in this fog.
071234567
9
Figure 2. An example sequence of the dynamic fog. Note that the “easy” object is difficult to see
in any particular image. However, when the sequence is animated, the object is easily seen in the
dynamic fog. Note also that the dynamics of the fog is independent of the dynamics of the object.
Design
Four main factors were tested in a mixed design with Object Type (“easy”, “hard”) and
Learning Context (Fog, No-Fog) as between-participants factors, and Test Motion (studied,
nonstudied) and Test Context (Fog, No-Fog) as within-participant factors. Ten observers were
run in each of the four between-participants conditions.
Procedure
The experiment consisted of two learning phases followed by a test phase. In the first
learning phase, observers were shown four objects individually for a full 360 rotation (~6500
ms). To eliminate any effects of seeing a new rotation direction during the test phase, two targets
rotated clockwise and the other two rotated counter-clockwise (by playing the animation
sequence either forwards or backwards). Each object’s rotation direction was randomly
determined for each observer at the beginning of the experiment, which established its particular
characteristic motion learned by that observer. The starting frame was selected randomly on each
trial. Observers were instructed to press the appropriate key for each object after seeing the
object make a complete rotation. They were informed that they could not respond until the object
was removed from the screen. If observers responded incorrectly, they heard a low 500 Hz tone,
071234567
10
and the correct response key was presented on the screen. If they responded correctly, they heard
a high 1000 Hz tone. For this phase, observers were instructed to respond as accurately as
possible. Each object was presented 30 times for a total of 120 trials. There was a short selftimed break after every 40 trials.
The second learning phase was the same as the first with the following two exceptions.
First, observers were instructed to respond as quickly and as accurately possible. Thus, in this
phase, they did not have to wait for the object to disappear from the screen before responding.
Second, if observers responded incorrectly, they only heard the low tone. As in the first learning
phase, each object was presented 30 times for a total of 120 trials, with self-timed breaks after
every 40 trials.
In the test phase, observers were presented with the four studied targets intermixed with
four unstudied distractors. They were instructed to press the space bar for all distractors, and to
continue to respond with the learned letter key associated with each target. During this phase, all
objects (both targets and distractors) appeared in both the presence and absence of the dynamic
fog and rotated in both rotation directions. Thus, on 50% of the test trials, the targets rotated with
their studied motion (established during the learning phase), and on the remaining 50% of the
trials, they rotated in a nonstudied motion (in which the studied motion was reversed). Observers
were not informed that the targets would rotate any differently than before. Targets and
distractors were shown 10 times in each condition during the test phase, for a total of 320 trials
(8 objects [4 targets/4 distractors] x 2 test contexts x 2 test motions). There was a short break
after every 40 trials. As in the second learning phase, observers were instructed to respond as
quickly and as accurately as possible. No feedback was provided during this phase. The entire
experiment took approximately 45 min.
071234567
11
Data analysis
Our main focus in the present study is the effects of rotation-reversal following learning.
Thus, only results of target trials from the test phase were analyzed. Accuracy and correct
response times (RT) for targets presented during the test phase were submitted to a mixed-design
analysis of variance (ANOVA) with Object Type (“easy”, “hard”) and Learning Context (Fog,
No-Fog) as between-participants factors, and Test Motion (studied, nonstudied) and Test Context
(Fog, No-Fog) as within-participants factors. Response times outside the range of 400 and 6500
ms in this experiment were removed to eliminate anticipatory responses and outliers. This
procedure excluded less than 4% of correct trials. A significance level of α = 0.05 was adopted
for all statistical analyses reported.
Results
Accuracy data
The mean percent correct scores are plotted in the left Figure 3a as a function of Object
Type and Test Motion. We plotted this interaction throughout this study because it was the most
robust finding. Observers performed well above chance levels (20%) in all conditions.
Furthermore, there was no indication of any speed-accuracy trade-offs in the data.
071234567
(a)
12
(b)
Figure 3. Mean accuracy (a) and response times (b) as a function of Object Type and Test
Motion. Error bars reflect +1 standard error (SE) of the mean.
We found main effects of Object Type, F(1, 36) = 21.04, p < 0.001, and Test Motion,
F(1, 36) = 20.31, p < 0.001, on observers’ accuracy. As evident in the figure, there was also a
significant interaction between Object Type and Test Motion, F(1, 36) = 15.08, p < 0.001,
suggesting that the effect of rotation reversal on observers’ accuracy was modulated by the
structural similarity of the objects. Lastly, there was no significant interaction between Learning
Context and Test Motion, and no significant three-way interaction between Object Type,
Learning Context and Test Motion, ps > 0.18. The ANOVA indicates that only Object Type
modulated the effect of rotation reversal on how accurately observers identified targets.
Response time data
The mean RTs are plotted in Figure 3b. For RTs, all main effects were significant: Object
Type, F(1, 36) = 132.52, p < 0.001; Test Motion, F(1, 36) = 18.89, p < 0.001; Learning Context,
F(1, 36) = 9.18, p < 0.01; and Test Context, F(1, 36) = 111.98, p < 0.001. Similar to the accuracy
data, there was evidence that structural similarity modulated the effect of rotation reversal on
RTs, as indicated by the significant interaction between Object Type and Test Motion, F(1, 36) =
8.74, p < 0.01. There was also some evidence that the availability of shape and motion
071234567
13
information modulated the effect of rotation reversal on RTs, but this did not reach significance
in the omnibus ANOVA (the interaction between Learning Context and Test Motion was
marginally significant, F(1, 36) = 2.55, p = 0.12, and the three-way interaction between Object
Type, Learning Context, and Test Motion was not significant, F(1, 36)<1). However, the trend in
the RT data is in this direction, as shown in Table 1.
Table 1. Mean (M) and standard error (SE) of the mean for percentage correct and RTs as a
function of Object Type, Learning Context, and Test Motion.
%
Correct
Object
Type
Learning
Context
No-Fog
Easy
Fog
No-Fog
Hard
Fog
RT (ms)
Test
Motion
M
SE
M
SE
Studied
97.0
0.8
1047
43
Nonstudied
96.1
1.1
1047
42
Studied
93.9
1.4
1150
42
Nonstudied
93.8
1.5
1213
55
Studied
81.5
4.1
2020
97
nonstudied
76.8
4.8
2147
115
Studied
77.7
3.5
2488
110
nonstudied
69.6
4.0
2699
132
Discussion
The experiment reported here was motivated by the question: How does object motion
affect observers’ ability to recognize objects? The results presented here converge with previous
studies demonstrating that observers use spatiotemporal signatures to recognize dynamic stimuli
(e.g., Knappmeyer et al., 2003; Newell et al., 2004). As we, and others, have shown, changing
071234567
14
the studied motion can severely impair observers’ recognition performance, even if 3D shape and
2D views are fully preserved (Liu & Cooper, 2003; Stone, 1998).
Our contribution to this growing literature is that we tested factors that affect how learned
spatiotemporal signatures are ultimately used for recognition purposes; namely, we examined the
structural similarity of the target objects and the availability of shape information during
learning. Both of these factors have been found to affect the recognition of object presented as
static images (e.g., Hayward & Williams, 2000), and thus we predicted that they would also
mediate the recognition of moving objects. The present results are consistent with this prediction.
We found interactions between structural similarity and availability of shape information in
modulating the rotation-reversal effect (Liu & Cooper, 2003; Stone, 1998). For “easy”
structurally distinctive objects, observers’ response times were affected by rotation-reversal only
if they had studied the objects in a noisy context. By comparison, with “hard” structurally similar
objects, accuracy and response times were affected by a rotation reversal in both learning
contexts, probably because these stimuli were already difficult to recognize (see also Vuong &
Tarr, 2004).
We focused on the learning component in the present study because previous studies
have not systematically investigated how the difficulty of learning objects may influence the
visual information observers use to recognize those objects. Indeed, as far as we can tell, the test
stimuli or learning context used in previous studies generally made it difficult to learn the
stimuli. The stimuli formed a homogeneous class (e.g., faces, amoebas, arm movements); they
were degraded in some manner (e.g., shown as point-light displays); or observers had to learn
many items from limited exposures. Our strategy to address this issue was to test qualitatively
071234567
15
different types of stimuli and learning contexts using a difficult individual-level identification
task.
It is important to point out that we used a procedure to provide observers with every
opportunity to learn both static (3D shape and 2D views) and dynamic (spatiotemporal
signatures) cues to recognize the target objects. In our learning procedure, observers were
initially required see the entire rotation sequence (first learning phase). Following that, they were
encouraged to respond as quickly and as accurately as possible (second learning phase). They
were provided with explicit feedback throughout both learning phases. Other studies have used
different learning procedures. For example, Knappmeyer et al. (2003) found that observers
learned very subtle characteristic facial movements of specific individuals incidentally. In their
study, observers were merely exposed to animated faces and asked to answer questions about
each individual (e.g., “which person is more friendly?”). Similarly, Liu and Cooper (2003) had
observers incidentally learn their set of objects by having them decide whether the object could
be used for support or as a tool. Other researchers have used famous or well-known individuals
so that observers would be familiar with the idiosyncratic movements of those individuals from
normal experience (e.g., Lander & Bruce, 2000). Indeed, a possible avenue for future research is
to test different learning procedures; for example, whether observers learn objects incidentally or
explicitly.
We acknowledge that there is one potential confound in the present study that provides an
alternative account of our data. During the test phase, it is possible that observers were surprised
when learned objects moved in a different manner, and this could have caused them to make
more errors or respond more slowly. To address this issue, we divided the accuracy and RT data
into two blocks and looked at the results only on the second block. That is, we only looked at the
071234567
16
last five (out of 10) presentations of each target object and in each test condition. We assumed
that any surprise effects should have disappeared by the second half of the test phase. A similar
pattern of results emerged with only trials from the second half of the experiment. Thus, even
after seeing learned objects moving with both studied and nonstudied motions and in the absence
and presence of the dynamic fog, observers were still sensitive to the dynamics of the objects
acquired during the learning phase.
Lastly, our data suggest that both structural similarity and availability of stimulus
information (shape and motion cues) had different effects on how observers ultimately used
spatiotemporal information, although both factors generally made the recognition task more
difficult during learning. However, as is often the case for static objects, structural similarity
seemed to be the critical factor in our study (Tarr & Bülthoff, 1995). The availability of shape
and motion information, on the other hand, had a weak effect that was evident only in comparing
differences between studied and nonstudied motion (mostly) RT distributions.
For the two types of objects we used, the results suggest that spatial and dynamic
information may be weighted differently. In our experiment, spatial and dynamic information
about structurally similar “hard” objects may have been equally weighted in the object
representation because motion information would help observers discriminate between visually
similar objects. By comparison, shape information may have been weighted more than motion
information for structurally distinct “easy” objects, as these can be accurately and quickly
identified on the basis of shape. In any case, future studies will be required to further explore this
important issue. Our data provide a starting point to further investigate precisely how learning
affects the combination of shape and motion cues for object recognition, so that appropriate cue
071234567
17
combination models may be formulated. For example, it would be interesting in the future to
systematically vary the structural similarity of targets from “easy” to “hard”.
In summary, we used rotation reversal (Liu & Cooper, 2003; Stone, 1998) as a means to
investigate the information observers use to recognize dynamic objects. Combined with earlier
results, our present findings suggest that how an object’s movements unfold over time also
contribute to the recognition of that object. That is, observers can learn to directly associate
specific dynamic information with specific objects (or classes of objects), particularly if this
information is informative with regard to object identity. Hence, the term “signature” is
appropriate: spatiotemporal signatures capture space-time structures projected onto our retinas by
a dynamic world.
071234567
18
References
Biederman, I. (1987). Recognition-by-components: A theory of human image understanding.
Psychological Review, 94, 115-147.
Bruce, V., & Valentine, T. (1988). When a nod’s as good as a wink: The role of dynamic
information in facial recognition. In N. N. Gruneberg, P. E. Morris, & R. N. Sykes (Eds.)
Practical Aspects of Memory: Current Research and Issues Vol. 1 (pp. 169-174).
Chichester, UK: Wiley.
Hayward, W. G., & Williams, P. (2000). Viewpoint dependence and object discriminability.
Psychological Science, 11, 7-12.
Johansson, G. (1973). Visual perception of biological motion and a model for its analysis.
Perception and Psychophysics, 14, 201-211.
Knappmeyer, B., Thornton, I. M., & Bülthoff, H. H. (2003). The use of facial motion and facial
form during the processing of identity. Vision Research, 43, 1921-1936.
Lander, K. & Bruce, V. (2000). Recognizing famous faces: Exploring the benefits of facial
motion. Ecological Psychology, 12, 259-272.
Liu, T., & Cooper, L. A. (2003). Explicit and implicit memory for rotating objects. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 29, 554-562.
Marr, D. (1982). Vision: A Computational Investigation into the Human Representation and
Processing of Visual Information. San Francisco: Freeman.
Mather, G., & Murdoch, L. (1994). Gender discrimination in biological motion displays based on
dynamic cues. Proceedings of the Royal Society of London B, 258, 273-279.
Newell, F. N., Wallraven, C., & Huber, S. (2004). The role of characteristic motion in object
categorization. Journal of Vision, 4, 118-129.
071234567
19
Perlin, K. (1985). An image synthesizer. Computer Graphics, 19, 287-296.
Stone, J. V. (1998). Object recognition using spatiotemporal signatures. Vision Research, 38,
947-951.
Tarr, M. J., & Bülthoff, H. H. (1995). Is human object recognition better described by geon
structural descriptions or by multiple views? Comment on Biederman and Gerhardstein
(1993). Journal of Experimental Psychology: Human Perception and Performance, 21,
1494-1505.
Tarr, M. J., & Pinker, S. (1989). Mental rotation and orientation-dependence in shape
recognition. Cognitive Psychology, 21, 233-282.
Ullman, S. (1984). Maximizing rigidity: The incremental recovery of 3-D structure from rigid
and rubbery motion. Perception, 13, 255-274.
Vuong, Q. C., & Tarr, M. J. (2004). Rotation direction affects object recognition. Vision
Research, 44, 1717-1730.
Wallis, G., & Bülthoff, H. H. (1999). Learning to recognize objects. Trends in Cognitive Science,
3, 22-31.