Requirements and minimum sizes

Before calculating, Statable checks what can be checked from the data and names the rest.

• Group comparison: a group under 20 people counts as small - the result is then less certain. Whether a departure from normality matters depends mainly on clear skewness and outliers, not on a fixed number of people: without either, Welch's test holds even with small groups.
• Scale and factor analysis: rule of thumb at least 5 people per question, ideally 10.
• Relationship: at least 3 people with answers in both columns.
• Regression: more people than predictors plus 2.
• Not checkable: whether everyone answered independently. This follows from how the data were collected.