ASSESSMENT &
GRADING
the best and worst of teaching
the organization
Planning
Instruction
Assessment
Instructional objectives
• for yourself and your students
• help keep organization and focus
• make sure they are observable,
measureable and outcome-directed
types of assessment
• informal
• question-asking
• observations
• formal
• types: homework, exams, attendance, quizzes, essays,
performance assessments
• remember that assessment is about distinguishing
between those who know and those who do not. Don’t
confuse with those who are good test-takers, make it
look good, etc.
writing assessment tasks
• ALWAYS write then walk away. Gives you fresh eyes
• performance assessments: make as authentic as is
feasible; focus on skill-attainment/demonstration; design
(and share, if comfortable) good scoring; make sure you
have the time for design, administration, and scoring; tend
to lend themselves to higher-order skills
• essay-type assessments: allow you to assess
writing/communication skills in addition to knowledge; nice
for allowing variations in responses, but trickier for
scoring; tend to lend themselves to higher-order skills
• traditional exam-type assessments: clarity!; format to
maximize efficient and effective responding; look carefully
for unintentional cues
borderline grades
• do they exist in your class?
• if so, what is borderline? (.5%? 2%? may or may not
want to include cut-off in syllabus)
• identify what you use to assign the higher grade
• most assignments in the class are at that grade level
• effort-based components (all assignments in on time, attendance,
participation)
• beware of being influenced by personality/negotiation factors or
characteristics that you really can’t measure (effort, potential)unfair!
• apply borderline grade changes consistently
• realize that borderline situations can increase and
decrease student grades if done fairly
extra credit
• check department and other instructors’
policies
• make available to all students
• consider how the extra credit affects final
grade- usually do not want to raise all
grades one letter
• effort- or ability-based?
• keep as a separate column in gradebook
so you can separate ability from other
factors
zeroes
• consider
• effect of the zero on the final grade
• reasons for the zero
• poor quality versus non-submission
• factors that led to the zero
• policies on late work
• alternatives
• give partial credit (late penalty, half credit substitution)
• calculate the final grade without that assignment
• provide different assignment
• allow all students to drop or replace a low grade
types of scoring
• checklists
• distinguishes “done” from “not done”
• best when you have a specific list of tasks within the assignment
• can be worth multiple points, but should be used for all-or-none credit
• e.g., “Paper was double-spaced” “Included example of concept”
• rating scales
• generally 3-5 categories
•
•
•
•
more than yes/no, but not so many categories to be fine distinctions
use parallel descriptors
avoid “average”
can assign point values to descriptors
• like checklist, breaks the assignment into specific tasks, but includes
quality markers
• e.g., “Concept description was detailed.
Not Done”
Excellent-Good-Fair-Poor-
• analytic rubrics
• like the ratings scale, but each point value has a specific
description
• e.g., “5 points=fully explained description that links to theory. 4
points= somewhat vague description. 3 points= missed a theory
link...”
• holistic rubrics
• grade overall assignment with a single grade
• still have descriptors for grade/point assignments
• used when cannot be easily broken into tasks or allowing A LOT of
freedom in responses
• generally fastest scoring to use, but tough to justify and tends to be
less reliable
• e.g., “10 points= presentation was complete and accurate. 9
points= presentation well done, but could have used more evidence
checklist example
Checklist (all-or-none)
Assignment uses the designated chart
At least 3 Bloom’s are analysis or higher
PA indicated
____ (+2)
____ (+2)
____ (+1)
ratings scale example
points
4
PA/task choice
PA/task
covers the
instructional
objectives
Expert
and
standards
indicated.
PA/task
design
allows
students to Excellent
demonstrate
mastery of
content.
3
2
1
0
Apprentice
Beginner
Novice
Not done
Good
Fair
Poor
Not done
analytic rubric
4
Habit description
Behavior
Description
description clear,
behavior
measureable
Stimuli
description
Daily log
Behaviors
each day
Plan when
fail
3
2
1
0
Description
very vague or
not
measureable
Description
Not
very vague
described
and not
measureable
Clear
description,
precede
behavior,
how
(dis)associat
e
Description
mildly vague
or somewhat
not
measureable
Mildly vague
description,
or not all
precede, or
vague
(dis)associat
e
Very vague
or none
precede or
no
(dis)associat
e
Cannot
assess those
identified as
stimuli
No identified
stimuli or
misidentified
as stimuli
5 days
4 days
2-3 days
1 day
All changes
described or
Missing
changes for
No behaviors
described
No changes
described for
holistic rubric
10 points = The report is well-organized, complete, and
reported in a unique fashion.
8 points = The report needs more detail or organization.
6 points = Good content, but format was not unique
4 points = Report needs major changes.
2 points = Report is incomplete or poorly executed
student
Brady,
Marcia
Bundy, Al
Halpert,
Pam
Jefferson
,
George
Simpson,
Lisa
Stark,
Ned
White,
Walter
averages
atten
part
ex1
hw1
pres
ex2
paper
final
excr total
%
10 pts
20
40
50
25
40
85
75
(5)
10
17
24
48
21
32
79
66
3
300 89.6%
10
18
30
48
24
33
72
28
5
268 80.0%
4
14
20
27
15
29
70
52
2
233 69.9%
9
0
28
49
25
38
82
71
0
302 90.1%
7
17
18
26
16
30
80
68
4
266 79.7%
10
20
32
0
25
40
85
75
5
292 87.2%
4
7
28
47
24
39
79
72
0
300 89.6%
13.4 25.9
35.1
21.7 34.4
78.3
7.7
335
61.9 (2.7) 281.1
84.0%
process of grading
• optimize self and environment
• have time, not tired or cranky (tough when facing a stack!)
• minimize distractions (friends, significant others, kids, Facebook)
• more subjective tasks usually require more time and concentration
• try to keep student identities unknown as you grade to avoid biases
• if grading both, grade mechanics (grammar, spelling) first, then content
• aim for two positive comments for each improvement comment- think of grading
•
•
•
•
•
as a conversation between you and the student, not a punishment
read through a few assignments before grading to get a sense of content and
quality
make comments on a separate page (or version if electronic) and go back
occasionally to see if you are grading the same as you had initially
if others are also grading the assignment, have everyone grade a few to ensure
consistency
for selected-response items, grade one page at a time to give yourself a sense of
how well items are working
for multiple constructed-response items, grade one at a time for all students (e.g.,
grade everyone’s item 27, then everyone’s 28…) for greater consistency