Re: Chisquare and gamma coefficient

Posted by E. Bernardo on Jan 26, 2011; 3:52am
URL: http://spssx-discussion.165.s1.nabble.com/SPSS-16-0-0-error-Two-step-cluster-Options-tp3353123p3357369.html

Thank you so much for your comment, Henrik.  The sample size is n = 77. Kindly elaborate "the association is monotone"? --- On Mon, 1/24/11, Henrik Lolle <[hidden email]> wrote: From: Henrik Lolle <[hidden email]> Subject: Re: Chisquare and gamma coefficient To: "Eins Bernardo" <[hidden email]> Cc: [hidden email] Date: Monday, 24 January, 2011, 12:10 PM If the association is monotone, then Gamma is a more powerful test than a chisquare test. But it sounds a bit strange that the value of Gamma is .34 and is highly statistical significant when the p-value from the chisquare test is above .05. How many cases do you have in your table? Best, Henrik Quoting Eins Bernardo <einsbernardo@...>: > I used chisquare to test the null hypothesis that the two ordinal > variables with 4 categories each are uncorrelated.  The chisquare is > non-significant (p>.05).  The gamma coefficient is .341 and highly > significant (p<.05).  Why gamma is significant while chisquare is > insignificant.  I confused about this.  I appreciate any comments. >   > Thank you. > Eins > > ************************************************************ Henrik Lolle Department of Economics, Politics and Public Administration Aalborg University Fibigerstraede 1 9220 Aalborg Phone: [phone] ************************************************************