Thanks for pointing them out. I think I need to learn more about categorical analysis. --Kingsley

Paul.R.Swank@uth.tmc.edu (Paul R Swank) wrote in message news:<018001c1da5b$d92ba5c0$520e6a81@PEDUCT225>... > Please note that the phi coefficient in the freq output and the Pearson > correlation from proc corr are the same. These results are in complete > agreement. Where is the problem? > > Paul R. Swank, Ph.D. > Professor, Developmental Pediatrics > Medical School > UT Health Science Center at Houston > > > > -----Original Message----- > From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU]On Behalf Of > chank > Sent: Tuesday, April 02, 2002 8:53 AM > To: SAS-L@LISTSERV.UGA.EDU > Subject: Re: Chi Square interpretation > > > I was trying to interprete a relationship on two variables. I used > proc corr and concluded they are not correlated. But when I used proc > freq; table a*b/Chisq; this morning, they are highly associated. Am I > interpreting them right? > > Here is the output: > > > The SAS System 08:53 Tuesday, April 2, 2002 26 > > TABLE OF BRDTH3 BY > OSD_DEF > > BRDTH3(If breadth > 3) > OSD_DEF(2 or More Func-OSD def) > Frequency > Percent > Row Pct > Col Pct 0 1 Total > 0 1878 257 2135 > 84.75 11.60 96.34 > 87.96 12.04 > 99.89 76.49 > 1 2 79 81 > 0.09 3.56 3.66 > 2.47 97.53 > 0.11 23.51 > Total 1880 336 2216 > 84.84 15.16 100.00 > The SAS System 08:53 Tuesday, April 2, 2002 27 > > STATISTICS FOR TABLE OF BRDTH3 BY > OSD_DEF > > Statistic DF Value Prob > Chi-Square 1 443.425 0.001 > Likelihood Ratio Chi-Square 1 297.180 0.001 > Continuity Adj. Chi-Square 1 436.804 0.001 > Mantel-Haenszel Chi-Square 1 443.225 0.001 > Fisher's Exact Test (Left) 1.000 > Right) 9.08E-66 > (2-Tail) 9.08E-66 > Phi Coefficient 0.447 > Contingency Coefficient 0.408 > Cramer's V 0.447 > > Sample Size = 2216 > > > > > The SAS System 08:53 Tuesday, April 2, 2002 28 > > Correlation Analysis > > 2 'VAR' Variables: BRDTH3 OSD_DEF > > > Simple Statistics > > Variable N Mean Std Dev Sum Minimum Maximum > Label > > BRDTH3 2216 0.036552 0.187702 81.000000 0 > 1.000000 If breadth > 3 > OSD_DEF 2216 0.151625 0.358738 336.000000 0 > 1.000000 2 or More Func-OSD def > > > Pearson Correlation Coefficients / Prob > |R| > under Ho: Rho=0 / N = 2216 > > BRDTH3 > OSD_DEF BRDTH3 > 1.00000 0.44733 > If breadth > 3 0.0 0.0001 > OSD_DEF 0.44733 1.00000 > 2 or More Func-OSD def 0.0001 0.0 > > > What is the difference between correlation and association? > > > Thanks > --Kingsley > >