Using Reports for Terms Prior to Fall 1992Reports prior to fall 1992 do not show section STD or SEM. These can be calculated as follows (* means "multiplied by"; **2 means "squared"):
Let QQ = [Number of A's * (4 - section mean)] **2 + [Number of B's * (3 - section mean)] **2 + [Number of C's * (2 - section mean)] **2 + [Number of D's * (1 - section mean)] **2 + [Number of F's * (0 - section mean)] **2 STD = square root of (QQ / (N - 1)), where N = number of A-F responses. SEM = STD / (square root of N).Reports from fall 1990 through spring 1992 showed percentiles rounded differently, so that "95" meant 95-99, not 92-99 as now; "90" meant 90-94 not 87-91, etc. We changed rounding methods to simulate true rounding (rather than truncation) better.
Reports for spring 1990 and before showed campus and department averages for ALL sections, not those in a single instructor group. The averages were weighted by section size (sheets returned), so that large classes had more weight. We separated instructor groups at the request of several departments and colleges. We switched to unweighted averages to capture the idea of a typical or normative section better.
* * * * * * * * Weighted and unweighted campus and department averages and standard deviations by instructor group (or with instructor groups combined) are available from R+I for longitudinal comparisons. See the entry "Requesting FCQ data" for more information.
The 12 standard questions changed in fall 1991. All published section reports show the questions used at the time.
Reports prior to fall 1990 show a statistic labelled department or campus "SEM" and defined in a footnote on the reports as "standard error of the mean." This is NOT the standard error or standard deviation of the department or campus mean as defined in this text. Instead, it's the square root of (the variance of weighted section means divided by the number of sheets returned in the one section in question). Thus this "SEM" is not the same for all sections within a group, but differs with class size. We now consider the reported statistic to be an extremely poor, essentially meaningless measure of group variability. WE RECOMMEND THAT IT NOT BE USED IN ANY COMPARISONS.
Last revision 04/08/05
PBA Home | Strategic Planning |  Institutional Research & Analysis |
Budget & Finances | Questions? Comments?
15 UCB, University of Colorado Boulder, Boulder, CO 80309-0015, (303)492-8631
© 2001, The Regents of the University of Colorado