Good advice on all counts. I'm curious about where you want to take this 'correlation' if found. A college admissions person could use the relationship (in the form of a regression equation), if any, to predict the score on the 'college level subject' for those students who were unable to take such a course in HS, or to predict the performance on the college level subject, once admitted. But correlation & regression eq. do not produce the same end result, especially where the independent variable contains significant measurement variance.
Donald Burrill wrote: > On Fri, 28 Dec 2001, Petrus Nel wrote: > > > I require some advice regarding the following: One set of variables is > > the grades obtained by students for different high school subjects > > (i.e. the symbols candidates obtained such as A, B, C, D, etc. for each > > subject). The other set of variables are the scores obtained for a > > college level subject (i.e. no symbols, just their percentages > > obtained). I want to determine the correlation between their grades > > for different high school subjects (A, B, C, D, etc.) and their > > percentage scores for a college level subject. > > Why do you want to? _Nobody_ just wants correlation coefficients: > there's always something more that is desired. > > > The grades obtained for their high school subjects were coded on the > > questionnaire as follows - 1=3DA, 2=3DB, 3=3DC, 4=3DD, 5=3DE, 6=3DF. > > I`ve entered the data for the grades as 1,2,3, etc. to indicate the > > grade (category) and the percentages (as the other variable) into SPSS. > > How do I proceed? > > What follows assumes that the answer to "Why?" implies that correlation > coefficients are part of the desiderata, for good reason(s). > > First, recompute each HS subject grade: e.g., > NEWGRADE = 7 - OLDGRADE > so that both grades and percentages are coded in the same direction > (higher values = better performance); else your correlation coefficients > will be negative where the relationship is positive, etc. Translating letter grades into a number implies that the letters were ordinal categorical data, and further, that they have equal increments (distance between them). Inasmuch as the grades came from a numerical scale, typically with 60 = D, one could argue that this translation is fully valid. HOWEVER.... As I recall, a correlation analysis assumes the data is Normally distributed. Grades usually ain't, and the upper limit of 'A' forces the issue. If you wish to compute confidence intervals, this may come back to haunt you. Nonetheless, good luck on this project, and don't forget Prof. Burrill's next urging, that you look carefully at some scatter plots. Those may tel you as much or more than your calculations. > > > Second, produce scatterplots of grade vs. percentage, grade vs. grade, > and percentqage vs. percentage, for all pairs whose correlations are of > interest: so that you can properly interpret correlation coefficients > when you get them (and can be prepared to deal with nonlinear > relationships, should there be any obvious from the plots). > For this purpose one needn't be too fancy: character plots will do, > high-resolution plots won't tell you anything more unless there are some > rather odd nonlinearities among the relationships. > > Third, invoke SPSS's CORREL routine to calculate all pairwise zero-order > correlation coefficients. > > Fourth, proceed with whatever your answer to "Why?" implies about > subsequent analyses and interpretation. > > -- DFB. > ------------------------------------------------------------------------ > Donald F. Burrill [EMAIL PROTECTED] > 184 Nashua Road, Bedford, NH 03110 603-471-7128 > > ================================================================= > Instructions for joining and leaving this list and remarks about > the problem of INAPPROPRIATE MESSAGES are available at > http://jse.stat.ncsu.edu/ > ================================================================= -- Jay Warner Principal Scientist Warner Consulting, Inc. 4444 North Green Bay Road Racine, WI 53404-1216 USA Ph: (262) 634-9100 FAX: (262) 681-1133 email: [EMAIL PROTECTED] web: http://www.a2q.com The A2Q Method (tm) -- What do you want to improve today? ================================================================= Instructions for joining and leaving this list and remarks about the problem of INAPPROPRIATE MESSAGES are available at http://jse.stat.ncsu.edu/ =================================================================
