How to Estimate the Variance of a Probit Model

By Hand

• 1

  List the output of the probit model in an organized fashion.

• 2

  Calculate the mean of the data. Sum all the data points and divide by the number of data points. Call this value “m.”

• 3

  Subtract the mean from every individual data point, creating a new dataset. Rewrite this new dataset in an organized fashion.

• 4

  Square the new data, creating again a new dataset.

• 5

  Sum the data in the new dataset.

• 6

  Calculate the variance of the new dataset. Divide the sum just calculated by the number of data points. This is the estimation for the variance for the original dataset.

By Computer

• 7

  Assign a variable to the output of the probit model. For example, in R, decide on a variable (e.g. x), and write “x <- probitmodel,” where “probitmodel” is the output of the probit model.

• 8

  Call the standard deviation command for the created variable. In R, type “sd(x).” The result will be a single number, the standard deviation of the model’s output.

• 9

  Square the standard deviation. The square of the standard deviation is the estimated variance for the probit model.