2
$\begingroup$

I'm currently analysing some EEG data using a repeated measures general linear model, and I've run into a problem. I have three within-subject factors; Scalp area (3), condition (2) and word (2). When I run a three-way analysis I get mean values for Word (separated by condition) which don't match up to the means I've calculated for myself, for condition 1 (those for condition 2 seem fine). When I take the relevant columns of data and enter them into a two-way analysis for just one condition (so Scalp area (3) x Word (2)), I get the correct means, and those for condition 1 don't match the ones from the three-way analysis.

I've tried this in SPSS and Statistica with identical results; I've checked the data, and I'm pretty sure it's not an input error. I've checked the order in which I'm entering the variables (slowest-changing first), and that seems fine. So either I'm doing something wrong in the analysis, or I've misunderstood the output, or... something else? I may well be staring straight past an obvious error, but I've spent all afternoon tearing my hair out, so I'm hoping someone might be able to offer some insight!

Thanks in advance!

Improve this question
$\endgroup$

1 Answer 1

Reset to default
2
$\begingroup$

The first thing to look for in a situation like this is missing values. With listwise missing value handling, for example, some cases will be knocked out that would show up in a calculation not affect by missing data in a different variable.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thanks JKP, I knew it had to be something obvious I was overlooking; excluding cases with missing data seems to bring everything back in line. Much appreciated! $\endgroup$
    – Hester
    Commented Oct 12, 2011 at 9:11
  • $\begingroup$ To clarify for future readers, this doesn't mean that excluding cases with missing data is the right way to handle the analysis, only that the two methods agree when there is no missing data. The method you choose should depend on what kind of missing data you have and what kind of inferences you want to make. $\endgroup$
    – Aaron - mostly inactive
    Commented Oct 12, 2011 at 19:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.