I had mentioned in one post related to this issue of a "procedure" I had run across a LONG time ago ... called the Magic Median Modifier. This was used to "adjust" student grades (on a % basis) across different courses to equalize the medians for all courses at 65. Don't ask me WHY this was done ... I am just reporting on the method. But, in a way, it does have some relevance to the current discussion of scaling raters marks across examinees.
The long ago paper was: Remarks Re: Marks by, H. B, Dean ... Assistant Superintendent , Secondary Schools, Toronto Board of Education .... Ontario Sec. School Teachers' Federation Bulletin ... V 43, #4, Sept 1963 Within that paper, there was a grid that was used to "adjust" grades ... here is a quick but not too clear scan of the grid (best I could do) http://roberts.ed.psu.edu/users/droberts/mmm.jpg Now, there is a line about 1/3 of the way up from the bottom ... the 65 line ... and notice the diagonal is a dark line ... from the 65 spot along that horizontal at 65 ... Here is what you would do There would be a small strip you would slide straight up or down with 0 to 100 on it ... Say your median WAS 65 .... then, the "strip" would align itself along that horizontal at 65 ... and your scores would read off (student by student) on the same scale that has 65 as the median ... if, NO adjustment But, what if the median in your class was 75? Then, you would slide that "strip" straight down until your 75 aligned with the diagonal of 65 ... see http://roberts.ed.psu.edu/users/droberts/mmm2.jpg So, your 75 became 65 ... your 60 became about 53, etc. on up and down the scale (Note: I will try to draw a nicer grid and post later) Dennis Roberts, 208 Cedar Bldg., University Park PA 16802 <Emailto: [EMAIL PROTECTED]> WWW: http://roberts.ed.psu.edu/users/droberts/drober~1.htm AC 8148632401 . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
