Hi Adrian
Technically likert scales are usually ordinal - Strongly Disagreeing
is not double Disagreeing.
But statistics are usually done on the sum of the scales, which would
be data suitable for a parametric test.
Also it is usual practice to 'cheat' the rules with surveys and use
parametric tests anyway.
Many thanks
Ben Haysom-Newport BSc (Hons), MBPsS, MSc
On 2 August 2011 13:49, Adrian Swann <[log in to unmask]> wrote:
> Hi
> I have data from questionnaires that is scored on a 5-point Likert scale. I
> want to do ANOVAs and t tests on this data. Would it be classed as ordinal
> level data and if so does that mean that I must use a nonparametric tests?
> Cheers
> Adrian.
>
>
>
>
>
>
|