That is a good idea. You usually think of using Spearman-Brown in the other direction (how many items you have to add to achieve a level of reliability) but you could use it to estimate how many to cut out. One problem might be the assumption that each item would count the same percentage toward the total. And I would want to decide which to cut by looking at item-total correlations (which would mean cutting most items that aren't differentially graded -- you get the points if you do the work). Of course, those could be retained as formative items under the formative/summative distinction. I think I will do this kind of an analysis of my classes this summer. I just see a lot of redundancy in my graded assignments and way too many assignments to grade.
Rick Dr. Rick Froman Professor of Psychology John Brown University 2000 W. University Siloam Springs, AR 72761 [EMAIL PROTECTED] (479) 524-7295 http://www.jbu.edu/academics/sbs/faculty/rfroman.asp -----Original Message----- From: Scott Lilienfeld [mailto:[EMAIL PROTECTED] Sent: Friday, May 05, 2006 11:18 AM To: Teaching in the Psychological Sciences (TIPS) Subject: [tips] Re: Maximizing reliability of class assessments while minimizing grading Hi Rick: I'd assume that you could use the Spearman-Brown formula to guestimate the total number of "items" (assignments) you'd need to achieve a sufficiently high Cronbach's alpha. ...Scott --- To make changes to your subscription go to: http://acsun.frostburg.edu/cgi-bin/lyris.pl?enter=tips&text_mode=0&lang=english
