Discovering Statistics; Chapter 8

Question 1

What is ordinary least squares (OLS)?

Answer 1

Parameters that are estimated by calculated the value that has the minimum sum of squared error.

Question 2

What is the standardized residual?

Answer 2

unstandardized residual / SD.

Question 3

What is the adjusted predictor value?

Answer 3

New predictor value when outliers are excluded.

Question 4

What is leverage?

Answer 4

Found in ‘residual statistics output’. The influence of observed value of the outcome variable over predicted value;

k + 1 / n

k = no. predictors.

Question 5

What is Mahalonobis distances?

Answer 5

Distance of cases from the mean(s) to the predictor variable(s).

Question 6

What is the DFBeta?

Answer 6

Difference between an estimated parameter; uses all cases and same parameter once influential case has been removed.

Question 7

What is the DFFit?

Answer 7

Difference between predicted value for a case when the model is calculated including and excluding a specific case.

Question 8

What is covariance ratio (CVR)?

Answer 8

A measure of whether a case influences the variance of the regression parameters.

Question 9

What is Akaike Information Criterion (AIC)?

Answer 9

Fit which penalizes the model for having more variables (big AIC = bad fit).

Question 10

Where can you find the individual contribution of variables to the regression model in SPSS?

Answer 10

In the coefficients table in the output from SPSS. The ‘sig.’ column tells us whether the variable maks a significant contribution.