| You use a correlation to find out if two sets of data are related – they go up-and-down at about the same times in about the same ways (though maybe negatively related). You can test the significance of your correlation using a t-statitic or do it the way that's in the course notes. You use a t-test to ask if the means of two comparable data are the same or different. You can use a t-test to compare stock prices of the same stock at different times. You can't use a t-test to compare two stocks unless they could conceivably have the same price (e.g., they both split from the same company and you want to know if they really are different after the split). Does that help? Mark Guzdial |
| The t-test is a test where you compute a t-statistic comparing two sets of data to determine if they come from the same or different population means, then look up the t-statistic in a table. You can also compute the t-statistic from a correlation, and you can look up the t-statistic in a table to determine the significance of the correlation. Mark Guzdial |