Testing hypotheses about a single coefficient in a model

After a model has been estimated, you may wish to test hypotheses about the value of an estimated coefficient. By default, Wizard performs a two-tailed test of statistical significance (difference from zero), but you can also perform a one-tailed test for negativity or positivity.

Wizard supports three kinds of hypothesis tests for individual model coefficients. Each is associated with a null hypothesis, as described in the table below.

To test one of these hypotheses about a model coefficients:

1. In the Model view, select a coefficient in the explanatory variables table
2. Below the coefficient picture view, use the button labeled Null hypothesis to select the null hypothesis of the test you wish to perform
3. The result of the hypothesis test will appear in the coefficient picture view

Wizard performs the hypothesis test using a t-statistic for linear models, and a z-score for non-linear models.

See also:

• Viewing the p-values of a model’s explanatory variables
• Testing hypotheses about multiple coefficients in a model
• Testing the joint significance of multiple coefficients in a model
• About models

Back to Wizard Help