0

I am running a regression model and assume that all my variables are following a standard normal distribution.

From this model I am getting the value of the coefficients of the predictors and their standard deviation.

Then I am dividing these two numbers (coefficients/standard deviation) to get the t-statistic (I hope that this is the right way to do this).

Therefore, I now have the t-statistic values for each of my predictors and I want to compare these values to the critical t-statistic value which corresponds to a 5% p-value.

Is the critical t-statistic value always 1.96 for the standard normally distributed variables or also it depends on the sample size or on something else etc?

Outcast
  • 155
  • It does not follow normal distribution, at least not as a general rule. See the linked thread. – Tim Jan 22 '19 at 12:55
  • @Tim, thanks you for your answer.

    However, can you be a little more specific please? What does not necessarily follow a normal distribution?

    In general it would be good if you can provide some context in your answer.

     – Outcast Jan 22 '19 at 14:34
  • You can find details in the linked thread. I didn't provide answer because it was already answered in another thread. – Tim Jan 22 '19 at 14:59

0 Answers0