0
$\begingroup$

I'm new to time series data, and I want to answer a question like: "On days when I eat a banana, is my average blood sugar level higher?"

That means, I have a time series where I have a bunch of blood sugar levels sampled every day, from which I can calculate a daily average. Let's say I have this data for 60 consecutive days.

I also have the information on whether I ate a banana or not for each day during the 60-day period.

Should I use a simple Students t test (assuming normal distribution), or would that be inappropriate as there may be a seasonal variation in my blood glucose levels as this is happening along a time line?

$\endgroup$

1 Answer 1

0
$\begingroup$

Maybe you could try a regression with autocorrelated errors. Let the response variable be blood sugar level, let the banana variable be a binary predictor. Or you could lag blood sugar level.

There are multiple posts here on regression with autocorrelated errors, some

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.