# Can I use an ANOVA?

#### Kyra Dawbarn

##### New Member
Hi everyone,
I’m doing an assignment where I was training my dog to sit. I want to compare 4 different levels (baseline, calm environments, excited environments & all environments) the last 3 levels had primary reinforcement and the baseline did not. I did a FR1-FR3 for 2 levels and VR5 for the last level.

I believe I can not use a one way anova as they are dependent values (done on the same subject). However I also don’t think I can use repeated measure anova as each level has different number of values (7,15,7,8). Is that correct? Does that mean that I can’t do any anylisis as not able to do chi square as not independent values or multiple t tests as it increases type 1 error. Also I was concerned with the small sample size. I remember reading that the minimum sample size of the total number of levels + 1 so I would be ok, but I’m having trouble working out the minimum sample size from a power analysis.

Kyra

#### j58

##### Active Member
I have no idea what all the psychology jargon in your post means. But the fact that you performed your treatments on the same subject does not make the observations dependent. If you have more than one subject, then observations on the same subject are dependent; however, they are conditionally indpendent, given the subject. It's having multiple subjects that induces dependence; you only have one subject, so this doesn't apply to your experiment. So, assuming all the other requirements of ANOVA are ok, you can perform 1-way ANOVA. On the other hand, have you given consideration to the possiblily that earlier treatments contaminate the effects of later treatments?

Last edited:

#### Kyra Dawbarn

##### New Member
Thanks! Sorry for the jargon I thought it was relevant stat talk lol I get confused with what is math and what is psychology. Very helpful thanks!

#### j58

##### Active Member
Thanks! Sorry for the jargon I thought it was relevant stat talk lol I get confused with what is math and what is psychology. Very helpful thanks!
No problem. It's a lot better to give more information than not enough.