Recommendations for Using Student Ratings of Teaching

Use multiple sources of data about a faculty member's teaching if you are serious about accurately evaluating or improving teaching.
Do use student rating data as one source of data about effective teaching.
Discuss and decide upon the purpose(s) that the student rating data will be used for before any student rating form is chosen or any data are collected.

To obtain reliable student rating data collect data from at least ten raters if this is possible.
To obtain representative student rating data from at least two-thirds of the class.
To generalize from student rating data to an instructor's overall teaching effectiveness, sample across both courses and across time.
For improvement, develop a student rating system that is flexible.
Provide comparative data, preferably for all the items. Student ratings tend to be inflated.
Discuss and decide what controls for bias will be included in your system.
Do not give undue weight to: the instructor's age, sex, teaching experience, personality or research productivity; the student's age, sex, level (freshman, etc.), grade-point-average, or personality; or the class size or time of day when it was taught.
Take into consideration the students' motivation level when interpreting student rating data.
Decide how you will treat student ratings from different course levels, e.g., freshman, graduate, etc.
Decide how you will treat student ratings from different academic fields.
For improvement, develop a system that is diagnostic.
Develop a system that is interpretable.

For evaluation, use a few global item or summary items or scores.
Use the short evaluation form (or items) in every class every term.
Use a long, diagnostic form in only one course per term-in the course that the instructor wishes to focus upon for improvement.
For improvement, use items that require as little inference as possible on the part of the student rater and as little interpretation as possible on the part of the instructor.
For improvement, do not use a single standard set of items for every class. Provide a pool of items or some kind of weighting system.
Use a 5-point to 7-point scale.
In the analysis of the results, report computations only to the first decimal place.
Do not over-interpret the data, allow for a margin of error.
Use frequency distributions-what number or percent of the students rated item "1" or "2," etc. These are more understandable to most faculty.
For improvement, ask for open-ended as well as quantitative ratings.
Use open-ended comments only for improvement.

For evaluation, develop standardized procedures covering all relevant aspects of you student rating system and monitor that the procedures are followed.
For evaluation, administer the ratings about the second week to the last week of the term.
Develop standardized instructions that include the purpose(s) for which the data will be used and who will receive what information, and when.
Instruct the students not to sign their ratings.
The instructor may hand out the rating forms and read the standardized instructions, but the instructor should leave the room until the students have completed the ratings and they are collected.
The ratings should be collected by a neutral party and the data taken to a predetermined location-often to where they are scored-and they should not be available to the instructor until the grades are turned in.

Develop a written explanation of how the analyses of the student ratings are to be interpreted.
Appoint a faculty member to serve as instructional consultant to help faculty interpret their results and to improve teaching.

LEARN Center