!Paired samples t-test OR Repeated measure ANOVA?!

Hi there,

i am currently writing a research proposal but I am stuck on what statistical test to use.
My Experiment:

double-blind placebo-controlled within subjects design

34 participants are split into two groups
Group A: receive oxytocin in week one and complete attachment task, repeat this in week two but receive a placebo drug
Group B: Receive placebo drug in week one, complete attachment task, repeat this in week two but receive oxytocin next.
AIM: exploring the effects of intranasal oxytocin administration on attachment security

My initial idea would be to conduct, one, paired samples t-test per group, comparing the mean SAAM scores before and after receiving the placebo vs oxytocin drug. Is this suitable? or would a repeated measures ANOVA testing the mean (SAAM score) difference between the placebo and oxytocin condition in each group be better?

I would really appreciate some clarity


Active Member
its an AB/BA 'crossover' study? paired t-test is essentially what it comes down to, but there are other issues. SAS proc t-test has a 'crossover' option which is pretty good and will walk you through it.


TS Contributor
You could use a "mixed" analysis of variance approach, with repeated-measures
factor placebo vs. oxytocin, and between-subjects factor group (A vs. B) accounting
for effects of sequence. Due to the small sample size, it is not expected that the
interaction effect sequence*condition becomes statistically significant, though.

With kind regards