IA IA Inferential Statistics

[Arctus98]

I have finished my Psychology HL Internals, did a study based on Bransford and Johnson's study on schema activation and memory recall. My teacher had suggested that I use Mann-Whitney U test for inferential statistics, as my data was ordinal in nature. Admittedly, I don't have a considerable understanding of inferential statistics yet. I recently saw that for recall from memory, the data collected was actually interval data, and that I should use Wilcoxon t-test.

The source is http://tdixonblog.files.wordpress.com/2012/12/detailed-internal-assessment-guidebook-hl.pdf , some IB teacher's blog. I am not following it in blind faith, as it is just a blog, but the materials there seemed good up till now. So, what should I do now? I can't simply follow my own teachers wishes, as my entire batch is being moderated by IB, as part of a random inspection, and everything MUST be absolutely correct. Can someone with a reasonable understanding of inferential statistics help me out?

[blindpet]

What was your dependent variable? Was it a repeated or independent measures design?

With small samples it is not uncommon or 'wrong' to run a non-parametric test (for ordinal or nominal data) on parametric data (interval or ratio data). I am assuming you had around 20 participants and in that case a non-parametric test would be ok and you won't be penalized as long as you mention that you are using the test for a good reason (i.e. the small sample).

[Arctus98]

Independent measures were used, with the presence of a picture (for activating a particular schema) before reading the passage that has to be recalled being the independent variable. Thanks for the advice, I am sticking with the Mann-Whitney U test for now.

[blindpet]

You have picked the right test for the design. I am still curious what your DV is though, were participants rating something about the picture?