 |
 |
If the samples were related, it's not the distribution of the variables that matter, but the distribution of the differences. If your sample is large enough, the tests are pretty robust to this anyway. (I don't like Wilcoxon / Mann-Whitney tests, because of the difficulty interpreting them so I avoid whenever possible). Jeremy On 1 June 2012 06:17, Kelly Tate <[log in to unmask]> wrote: > Hi all, > > > > Thanks for all your advice which was far quicker than pouring through stats > text books! In the end I used a Wilcoxon signed ranks test (as the > distribution was not normal and the sample was related). > > > > Kelly > > > > Kelly Tate > > PhD Researcher > > Sustainable Consumption Institute > > School of Psychological Sciences > > University of Manchester > > > > http://www.sci.manchester.ac.uk/ > > > > From: Research of postgraduate psychologists. > [mailto:[log in to unmask]] On Behalf Of Clare Sutherland > Sent: 01 June 2012 13:38 > To: [log in to unmask] > Subject: Re: analysis help! > > > > > > .... but you can't use a chi2 test for non-independent counts. > > > > > > I think that the Wilcoxon signed-ranks test would be acceptable. In fact, I > have similar data, but in a two-way design, and I've been using Friedman's > ANOVA since they are non-normal, so any further thoughts would be really > useful. > > > > > > Clare > > > > > > Clare Sutherland > > > > PhD Researcher > > Department of Psychology > > University of York > > > > > > On 1 Jun 2012, at 12:58, Robert Teszka wrote: > > > > You can compare two proportions using a Chi-Squared Test. > > > > - r > > On 1 June 2012 12:14, Kelly Tate <[log in to unmask]> > wrote: > > Sorry, I should have said a Wilcoxon test since the samples are not > independent (non-parametric data = baffling!) > > > > Kelly > > > > Kelly Tate > > PhD Researcher > > Sustainable Consumption Institute > > School of Psychological Sciences > > University of Manchester > > > > http://www.sci.manchester.ac.uk/ > > > > > > > > -- > "Believe those who are seeking the truth. Doubt those who find it." - Andre > Gide > > Robert Teszka | Twitter @RobTeszka | > http://robteszka.wordpress.com/ | http://scienceofmagic.org | > >