2. Reduce nonequivalence in the data. Researchers can take steps to identify equivalent and nonequivalent parts of their methods and then refocus their comparisons solely on equivalent parts (refer to Byrne & Matsumoto, 2021 for an extended discussion). For example, if researchers used a 20-item scale to measure anxiety in two cultures and found evidence for nonequivalence on the scale, they might examine each of the 20 items for equivalence and rescore the test using only equivalent items. Comparisons would then be based on the rescored items.