# Thread: Doubt on statistical test

1. ## Re: Doubt on statistical test

Rightio! Let's go for it then!
Thanx a bunch

2. ## Re: Doubt on statistical test

Hi!

Does a significant p.value mean that either one or both factors (context and rate) have a significant effect on the dependent variable? If so, how can I pin it down? Is there a way to find out how much each factor explains the variability observed in the final results compared against the expected values?

Thanks!

sid2016

3. ## Re: Doubt on statistical test

Hi,
yes, the significant p tells you that somemfactor combinations do have an effect. In the output of the test you generally have the chi-squared value for each cell. Those with a high value are probably the contributors to the effect.

regards

4. ## Re: Doubt on statistical test

Thanks!! Let's see how it goes...

5. ## Re: Doubt on statistical test

Just to clarify, are these repeated measurements on each subject? As in, each person receives a sequence of all five factor combinations?

Also, is the outcome the rate at which they choose a particular response out of the three (the "right" response vs. the "wrong" response)? Or do the rates of all three responses matter?

6. ## Re: Doubt on statistical test

Hi ab-stats,

1) each listener will hear the different items five times in a random order.
2) response rate is not collected here

Best
sid2016

7. ## Re: Doubt on statistical test

OK, since you're dealing with repeated measurements here, be aware that responses within each subject might be correlated. This would violate the independence assumption of Pearson's chi-square test. One option is to fit a mixed-effects logistic model that includes a random intercept for each subject. Categorical Data Analysis by Alan Agresti is a good resource on this topic.

8. ## Re: Doubt on statistical test

Right on!! This is getting even better : ) I'll check out your suggestion!
Thanks!