p.o. analysis on a b.subjs one way anova!

I'm doing an undergraduate psychology project on how music affects pain. I have one factor (music) with 3 levels (preferred music, non-preferred music, no music). I want to know how many participants to have in each condition (my supervisor said around 6). How would I go about calculating this (i'm fairly clueless about power of analysis). My hypothesis is that preferred music will significantly increase pain thresholds, whereas non-preferred music will significantly decrease pain thresholds.
thanks in advance!

Calculating the sample size and power for ANOVA manually is difficult but you can use a free software GPower to do the same.

I've had a look at the program and find it very confusing. And all of my text books I've looked in don't explain how to actually work it out. I've asked a few of my tutors and I'm getting the impression that most of them have problems with power of analysis too...
Thankyou so much for that! I'm going to have a go at trying to get my head around this. Just a few questions, what is MSE and MST?

(I can't believe they haven't gone over this with us at Uni.... and I'm now in my final year! :rolleyes: )

Oh and also, what are the means for? As I haven't carried out the study etc.?

(sorry, fairly clueless about statistics!)