Dots in a Repeated Measures ANOVA
Can someone please help me in the following: 1. I have conducted a Repeated-Measures ANOVA using SPSS and the table of Within-subjects effects showed some results of dots in the cells of the table. I also get these dots in the alpha column of the Homogeneity of covariance test results table. Since Repeated-measures ANOVA is very sensitive to the test of Homogeneity of Covariances, I really need to understand what these dots in the results of the Homogeneity of Covariances test signify. 2. I have lost some significant effect that was present when I conducted the repeated-measures ANOVA on sub-samples of the sample separately (i.e. when I conducted the same ANOVA test on the whole sample the significant interaction effect disappeared). Is this a normal outcome? HAYTHAM = Instructions for joining and leaving this list and remarks about the problem of INAPPROPRIATE MESSAGES are available at http://jse.stat.ncsu.edu/ =
Re: Dots in a Repeated Measures ANOVA
Thanks for your response. Can you please clarify the point you made about the Homogeneity test. Does this mean that the Homogeneity of covariance test has not been violated and thus the null hypothesis of equal covariances in the groups is therefore not rejected. Can you refer me to any books or other sources of that can explain this situation? HAYTHAM Thom Baguley <[EMAIL PROTECTED]> wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... > haytham siala wrote: > > > > Can someone please help me in the following: > > > > 1. I have conducted a Repeated-Measures ANOVA using SPSS and the table of > > Within-subjects effects showed some results of dots in the cells of the > > table. I also get these dots in the alpha column of the Homogeneity of > > covariance test results table. Since Repeated-measures ANOVA is very > > sensitive to the test of Homogeneity of Covariances, I really need to > > understand what these dots in the results of the Homogeneity of Covariances > > test signify. > > Dots usually mean insufficient d.f. to compute the tests. If within > factor levels have only 1 d.f. then there is 0 d.f. to estimate the > covariances. (The assumption can't be violated). > > > 2. I have lost some significant effect that was present when I conducted the > > repeated-measures ANOVA on sub-samples of the sample separately (i.e. when I > > conducted the same ANOVA test on the whole sample the significant > > interaction effect disappeared). Is this a normal outcome? > > > > HAYTHAM > > Not enough info to answer this, but it seems perfectly reasonable > and possible. If you analyze different data you'll get different output. > > Thom = Instructions for joining and leaving this list and remarks about the problem of INAPPROPRIATE MESSAGES are available at http://jse.stat.ncsu.edu/ =
Normality in Factor Analysis
Hi, I have a question regarding factor analysis: Is normality an important precondition for using factor analysis? If no, are there any books that justify this. = Instructions for joining and leaving this list and remarks about the problem of INAPPROPRIATE MESSAGES are available at http://jse.stat.ncsu.edu/ =
Re: Normality in Factor Analysis
I have checked some of the books but I could not find this statment (for e.g. using multivariate statistics (Tabachnik 1996), Latent variable models (Loehlin 1998), Easy guide to factor analysis (Kline 1994)). Can you please give me a examples of references as I reallly need a reference because I have already conducted a factor analysis on sample of data containing some non-normal variables . - Original Message - From: "Eric Bohlman" <[EMAIL PROTECTED]> Newsgroups: sci.stat.consult,sci.stat.edu,sci.stat.math Sent: Sunday, June 17, 2001 2:08 AM Subject: Re: Normality in Factor Analysis > In sci.stat.consult haytham siala <[EMAIL PROTECTED]> wrote: > > I have a question regarding factor analysis: Is normality an important > > precondition for using factor analysis? > > It's necessary for testing hypotheses about factors extracted by > Joreskog's maximum-likelihood method. Otherwise, no. > > > If no, are there any books that justify this. > > Any book on factor analysis or multivariate statistics in general. > "Eric Bohlman" <[EMAIL PROTECTED]> wrote in message 9ggvug$451$[EMAIL PROTECTED]">news:9ggvug$451$[EMAIL PROTECTED]... > In sci.stat.consult haytham siala <[EMAIL PROTECTED]> wrote: > > I have a question regarding factor analysis: Is normality an important > > precondition for using factor analysis? > > It's necessary for testing hypotheses about factors extracted by > Joreskog's maximum-likelihood method. Otherwise, no. > > > If no, are there any books that justify this. > > Any book on factor analysis or multivariate statistics in general. > = Instructions for joining and leaving this list and remarks about the problem of INAPPROPRIATE MESSAGES are available at http://jse.stat.ncsu.edu/ =
