Types of Hypothesis Tests

Briefly

Choose Test Hypothesis from the Analyze menu, then choose what type you want in its popup menu.

You create an empty hypothesis test either by dragging one off the tool shelf or by choosing Test Hypothesis from the Analyze menu. Choose the kind of test from the popup menu. Drag attributes into the slots at the top of the test.

If you drag no attributes, you can still use most of  the tests by editing the summary statistics in the test. (They're calculated for you if you use attributes from a collection.)

Attribute types

Test types

One continuous attribute
(e.g., just height in inches)

Test Mean (one-sample t)

Two continuous attributes
(e.g., income80 and income90)

Compare Means (two-sample t)

Test Correlation

One categorical attribute
(e.g., just sex: M or F)

Goodness of Fit (chi-square)

Test Proportion (binomial)

Two categorical attributes
(e.g., sex and marital status)

Test for Independence (chi-square)

Compare Proportions (Z test)

One continuous and one categorical
(e.g., income and sex)

Analysis of Variance (ANOVA)

Compare Means (two-sample t) (if the categorical attribute has exactly two values)

Note: to do a paired t-test, create a new attribute containing the difference between the two elements, and do a Test Mean to see if that difference is equal to zero.

How to use a hypothesis test