Close Home Forum Sign up / Log in

Stats Help - Repeated Measures ANOVAs with Unequal Sample Sizes

F

Hi, i'm after a bit of stats help from someone a little brighter than me....

So, i've run an experiment as follows (to retain anonymity the types of data/categories have been changed): There were two groups who used system A or system B (i.e. my between groups variable). And i want to investigate the effect on their output score (between 1 and 100) over time. So, i took note of the score at 4 time periods - at the start, after 5 mins, after 10 mins and after 15 mins. So, score was my repeated measure.

I need to analyse this (preferably using R, though SPSS is also available to me), but my sticking point is that i have unequal sample sizes (agh!). So, there were 21 people using system A and 24 people using system B (I didn't design it this way - unfortunately had to discard some data).

So, i think i need to do a one-way repeated measures ANOVA for unequal sample sizes. But how?!?!?

Can anyone give me any hints? I've tried this r tutorial series (http://rtutorialseries.blogspot.com/) which is very helpful, but doesn't quite help me out on the unequal sample sizes - i looked at the unequal size section and the discussion of Type I or Type III confused me no end.

Thanks in advance,
Frog

F

I think i've managed to do it in SPSS.. but i wish i could figure this out in R. I much prefer it. Dammit. Oh and the results were non-significant. Also dammit.

yeah, that's a mixed ANOVA with one independent factor which has 2 levels (system A or B) and then a second factor, which is repeated measures, which has four levels (time 1, time 2, time 3, time 4).

When books talk about unequal sample sizes, then the difference between 21 and 24 is fine, its when the difference is like 21 and 60 then its an issue - especially if all your assumption tests are ok (i.e Levene's test).

F

Thanks :-)

21114