t test or ANOVA or both???

#1
Hey Everyone,

I'm a Newby who has just joined this forum and already I have a question for you all.

I'm wondering which analysis to conduct... I have a within subjects design that has 2 conditions, with 2 phases for each condition. The dependent variable is response rate average and Independent variable is the stimulus used.

I was thinking of doing a 2x2 ANOVA to compare the means accross the groups and then doing pairwise comparisons or planned comparisons independent t tests to establish the direction of the effect.

What does everyone think?
 
E

elnaz

Guest
#2
Hello
if possible you write statistical.
but i say you that for every project you shoud do:
1 ) test of normality
2 )if your date is normal you can use design
3 ) if not you should use nonparametric test
excuse me if possible write statistical for exaple we dont have condition but i think your intent is invoice and i think your tent about phase is level.
is it true?

Best regards
Elnaz
 
Last edited by a moderator:
E

elnaz

Guest
#3
Hello
if you have within subjects design you should not use ANOVA you should use
repeated measure.
 
#4
Repeated measures is a special type of ANOVA. IE, you should do repeated measures ANOVA. For your multiple comparisons you should do dependent t-tests, and control for the effect of multiple testing using some kind of alpha correction, e.g. the bonferroni correction - where you divide alpha by the number of pairwise comparisons.

May I ask, what are your conditions and phases?
 
#5
Mightyphil said:
Repeated measures is a special type of ANOVA. IE, you should do repeated measures ANOVA. For your multiple comparisons you should do dependent t-tests, and control for the effect of multiple testing using some kind of alpha correction, e.g. the bonferroni correction - where you divide alpha by the number of pairwise comparisons.

May I ask, what are your conditions and phases?
I totally agree.