How are post hoc tests defined?

How are post hoc tests defined?

• A. A set of comparisons between group means that were not planned before data collection.
• B. A method for imputing missing data.
• C. A model used for time-to-event data.
• D. A method for adjusting p-values when conducting many planned comparisons.

Answer: (A)

Post hoc tests are the comparisons you perform after an overall test (like ANOVA) shows that there is a difference among group means. They’re used to explore which specific group means differ, and they’re not usually specified before you collect the data. Because you’re making multiple comparisons, post hoc tests adjust the p-values to keep the chance of a false positive (Type I error) under control. In short, they are unplanned, exploratory follow-up tests to identify where differences lie after finding a significant omnibus result. These aren’t about imputing missing data, nor are they a model for time-to-event data (survival analysis). They’re also distinct from planned comparisons, which are hypotheses you state before collecting data; those often involve adjustments too, but the defining feature of post hoc tests is that the comparisons were not planned in advance.

Post hoc tests are the comparisons you perform after an overall test (like ANOVA) shows that there is a difference among group means. They’re used to explore which specific group means differ, and they’re not usually specified before you collect the data. Because you’re making multiple comparisons, post hoc tests adjust the p-values to keep the chance of a false positive (Type I error) under control. In short, they are unplanned, exploratory follow-up tests to identify where differences lie after finding a significant omnibus result.

These aren’t about imputing missing data, nor are they a model for time-to-event data (survival analysis). They’re also distinct from planned comparisons, which are hypotheses you state before collecting data; those often involve adjustments too, but the defining feature of post hoc tests is that the comparisons were not planned in advance.