These are five areas that have been shown through research to be important in high-quality educational practice. Through examining empirical results of factor analytic models and with input from its Technical Advisory Panel, CCSSE grouped survey items related to each of these areas.
In order to create the benchmark scores, the survey items associated with each benchmark are first rescaled so that all items are on the same scale (0 to 1). Next, the benchmark scores are computed by averaging the scores of the related survey items. Finally, the scores are standardized around the mean of the 3-year cohort so that respondents' scores have a mean of 50, weighted by full- and part-time attendance status, and a standard deviation of 25. Benchmark scores are then computed by averaging the scores on the associated items.
Standardizing scores increases the value of the scores in the following two ways:
- 1) all of the benchmarks are on the same scale, enabling comparisons across benchmarks; and
- 2) the scores provide information about how an individual institution is doing relative to other institutions.
Having scores on the same scale makes comparisons across benchmarks more meaningful than comparisons using raw scores. The limitation of using raw scores is illustrated by the fact that there are some aspects of student engagement that will nearly always take place with greater frequency than others. For example, students may typically engage in activities such as homework more frequently than they engage in talking with instructors about career plans, so raw scores will always be higher for homework than for talking with instructors. Without knowing what a typical score is for each of these engagement areas, college faculty and staff have little basis for determining where their institutional strengths and weaknesses may lie.
Knowing that all benchmarks are on the same scale makes it immediately obvious that a particular college's scores are either above or below those for other participating schools. Knowing that the mean is at 50 across all students in the sample enhances the score's usefulness for benchmarking, as the score contains information about whether an institution's performance is better or worse than average—and how much better or worse. For example, benchmark scores of 55 on Student Effort and 42 on Student-Faculty Interaction would indicate to a college that the institution is performing relatively better in regard to Student Effort than in regard to Student-Faculty Interaction. Further, the benchmark scores allow the college to understand that the results are above the average for participating institutions on one benchmark and below average on the other. Thus the data may be used both to identify relative strengths and to zero in on areas in which the college may need to improve.
Back to top