How is the 2004 calculation process different from previous years?
The diagram below gives a general sense of how the calculation process changed. In the process used prior to 2004 (orange arrows), scores for each variable were aggregated at the institution level, creating weighted means for each item. For those variables that make up a benchmark, these means were then combined to calculate the benchmark score. This process was used to minimize the number of missing values (unanswered survey questions) in the items that make up the benchmark. In the 2004 process (blue arrows), the missing values are accounted for in a different way. Student-level scale scores (see Question 4i) are created for each student that answered 3/5ths or more of the items in a particular scale. The weighted means of these student-level scale scores are the institutional benchmarks.
