If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

### Course: AP®︎/College Statistics>Unit 13

Lesson 2: Testing for the slope of a regression model

# Calculating t statistic for slope of regression line

Calculate the test statistic in a test about the slope of a regression line.

## Want to join the conversation?

• how do you calculate the SE for b? if the value hasn't been given by the computer
• Shouldn't the alternate hypothesis be that beta is not equal to 0?
• It would be if we were testing for positive or negative linear relationship. But we're testing for only positive linear relationship.
• While calculating the t statistic, why are we not dividing the standard error of b (s.e) by root n?
• It's the standard deviation of the sample that would need to be divided by sqrt(n) to get the standard error.

Here, the value for the standard error we get is already precomputed, we don't need to make any further modifications.

For example, you can see that the stdev 8.554, divided by sqrt(24) = 1.746.
(1 vote)
• Regarding notation: Shouldn't the "y" denoting the population parameter be simply "y" instead of "y^", to differentiate from the estimator "y^" (statistic)?
• Why do you use a t-statistic and not z?
(1 vote)
• Use the `t` statistic because you do not know the population standard deviation, you are simply estimating it.
• Time stamp: , What are the parameters in relation to the specific test going on? In this case would they be the population mean time for the games or the population mean time for the memory time?
(1 vote)