Steps in Developing a Performance Assessment
1. Define the performance
-
· integrate the desired performance with course instructional objectives.
-
· "operationalize" the task and clearly define the concept of a
"quality" performance.
-
· determine if the performance should be naturally occurring or
manufactured.
2. Determine the most appropriate form for the assessment scale
-
checklist (i.e., performance is characterized by can/can't type outcomes)
-
combination of primary traits (i.e., performance is characterized by a
number of component parts, defining its critical attributes)
-
rubric scale (i.e., performance is characterized
by a hierarchical progression of quality and complexity as the performance
is mastered)
click here for more explanation of each type of scale
3. Create a scale to represent your criteria
-
develop a framework that any and all performances can be placed within
for clear and reliable scoring.
-
determine the type of score/feedback most suitable.
-
holistic (i.e., one score representing the complex elements of the
performance)
click here to
go to rubric construction page
* primary traits (i.e., a series of scores relating to each component of
the performance)
click here to see some great scales
using primary traits
* narrative (i.e., anecdotal feedback relating to salient aspects of the
performance)
communicate assessment criteria with participants, and/or have
the participants take part in the development of the criteria.
4. Prepare for sampling and technical considerations
¨ How much of student's work is necessary to represent the whole?
¨ How many students out of the whole?
¨ How can the procedure best be carried out efficiently?
5. Address issues of reliability
Problems addressed description solution
inter-rater reliability · different scores for the same level
of performance from different raters · common criteria
· practice rating together
· anchor papers/performances
intra-rater reliability · different ratings given to performances
that are similar · experience
· awareness of biases
· sample effectively
scaling reliability · too much or too little variation in the
scores.
· all too high or too low · anchor papers (papers that
have been chosen to define each level of performance)
performance reliability · ratings may be consistent, but do
not fit with what performers were trying to do. · common understanding
between raters and participants
· Well-defined criteria
· use of student self-assessment
BACK TO MAIN PAGE
BACK TO ASSESSMENT
MAIN PAGE