Hi, i'm studying the effect of credit rating changes on the bid ask spread of bonds; so i have multiple dates representing the event date. I first computed the abnormal bid ask spread as the difference between the actual b-a spread and the estimated b-a spread( which is simply the mean of the b-a spread in the estimation window). After having all the abnormal b-a spread for each event date, can i take the mean to connect all the event dates? If yes, how can apply the t-test on those mean?
I don't quite understand the experiment. I am a biologist, not an economist, so maybe the question is perfectly clear to others with a more appropriate background.
What are you testing with the t-test?
If you have data for every day in a year, and you want to compare if July 8th is significantly different than any other date, then a t-test will give you meaningless differences driven more by the 1 in 20 probability of rejecting the null hypothesis when the null hypothesis is true.
Well yes you can use a t-test to test the significance of abnormality. You can also use the empirical distribution based on the trend period data to test the same.
- Badges
- Science topic