T-test for distance between 2 Gaussians

Hello,

I have an image with eye-tracking data for it for two classes of people (25 people in each class). I've calculated the mean eye position and variance for the two classes over time. So, for a given time t, I have two Gaussians with variance in the x and y direction. I'm looking for a t-test to detect when the difference between the two Gaussians is significant. Can someone direct me in the right direction for this problem?

Many thanks!

Answer 1

[the cyclist]

If you have the Statistics Toolbox at your disposal, then you can use the ttest2() function to do this. You would apply this to the raw data, without needing to calculate the mean or variance yourself. Type

doc ttest2

for details.

Answer 2

[Zbigniew]

Thanks for the response. But, if I understand correctly, the ttest2() function only works with 1D data. I have the (x,y) coordinates for each person.

[the cyclist]

Hm. A t-test is a one-dimensional test. I don't know the equivalent for two-dimensional data.