You should transform it if it is non-normal and a transformation leads
to making the distribution closer to a normal distribution.
Jeremy
2009/8/6 Geraldine Hannon <[log in to unmask]>:
>
> In order to investigate group differences I have decided to treat the
> attitudinal data as continuous and to use ANOVAS. Should I transform the
> attitudinal data in order to make analysis more robust?
--
Jeremy Miles
Psychology Research Methods Wiki: www.researchmethodsinpsychology.com
|