Contact effects are consistent across study, sample, and test analytic levels.

Evidence

We also examined the effects at three distinct levels of analysis. Analyses conducted at the level of studies represent the overall effects for all data reported in each paper. Analyses at the level of samples represent the overall effects for each independent sample reported in each paper; since studies often include multiple samples, analyzing data at the level of samples offers larger numbers of cases for conducting more detailed comparisons of effects. Analyses conducted at the level of tests represent effects for each individual test of the relationship between intergroup contact and prejudice. Analyzing data at the test level offers even more cases for detailed comparisons. But because multiple tests from the same sample violate statistical assumptions of independence, we use tests as our unit of analysis when variables can only be measured at that level. — , Rudman, Laurie A - On the Nature of…, 1684

Categorization

• DEFINE: Clarifies the three analytic levels used throughout chapter 16 and how independence violations guide unit choice.
• XREF: connects to: A focal definition enables replication and comparison across studies. — Erik Bleich, What Is Islamophobia and How Much Is…, 71