Chi-square test (χ²)

The chi-square test checks, for two columns of groups, whether one depends on the other. It compares the counts with the counts you would expect without a relationship.

Example: Gender and phone in bed: χ²(2, N = 120) = 1.83, p = .401. The data provide no clear evidence of a relationship. The group "diverse" has only 1 person, though - that makes the test unreliable here.

Caution: The test needs enough people in each cell of the table. Rule of thumb: expected counts below 5 in at most 20% of the cells and none below 1. Otherwise, for a 2 × 2 table, Fisher's exact test applies.