Hi,

I am in the 2nd year of an MSc in Forensic Psych and am having a slight problem with the stats for my diss. My study is examining the effect of prisoner age and victim age on blame attribution style and locus of control. I have 4 DV's (1 q'aire with 3 subtests, and 1 other q'aire) and 2 IV's, these being prsioner age (adult/ YO) and victim age (minor/adult).
I have identified a number of outliers from box plots but am now stuck!
Firstly, I'm not sure if I've identified them correctly. In SPSS 12 I examined the data through Explore, where I set my 4x DV's in the dependent list and my 2x IV's in the factor list. From the box plots I identified 8 univariate outliers and 2 multivariate outliers. Now I have no idea what to do. Do cases need excluding/ transforming? And if so how?!

Any advice would be gratefully received,

Katypie1
:confused:

With outliers, you can either transform the data to reduce/eliminate their effect, or you can simply remove them from the analysis.

However, transforming variables makes them more difficult to interpret, so I would run the analysis both with and without the outliers to see if it changes your conclusions.