Dear all

This is my first post in the forum. I'm sorry if this question already came up, I used the search function but did not find anything.
I am currently doing my master thesis and conduct an event study analysis. I research how the winning probability changes on an online betting exchange during a match.

I am using the constant mean return model as there is no market model for an online betting exchange.

Now, I have my abnormal, average abnormal, cumulative abnormal and cumulative average abnormal returns.
I want to test for significance.

I have read the paper of MacKinley 1997 (Event Studies in Economics and Finance)

MacKinely says the CAAR follows a
N(0, var(CAAR))
where var(CAAR) = sum(var(AAR) over the event window)

Then we can use a normal t test to check if the cumulative average abnormal return is different from zero.

However, I find that several websites use a cross-sectional t-test.
Now, my questions is which one to use?

MacKinley assumes jointly multivariate normal and independent and identically distributed asset returns.
Does this suffice to use a normal t-test?

Thanks a lot for your answers
If something is not clear or you have questions, don't hesitate to ask.