Abstract
Grid-Group Cultural Theory (CT) has attracted significant attention across the social sciences and is increasingly being used in survey research. But the reliability of CT survey scales has mostly only been assessed using Cronbach’s alpha. Unfortunately, researchers frequently misinterpret this measure and often use the wrong or incomplete statistics to assess scale reliability. We compare the scale reliability of CT and its “cultural cognition theory” (CCT) variant as operationalized in surveys. Reviewing 73 survey studies using CT and CCT measures since 1990, we find limited application of appropriate method. CT measures, which are heterogeneous across studies, are far less reliable overall. But scales of roughly equal numbers of items are about equally reliable across CT and CCT studies. We apply recommended statistical measures to our data on both CT and CCT responses from an online national U.S. panel to illustrate a better approach to assessing scale reliability. We find that relative reliability is mixed (one CCT scale is more reliable, and one less, than the specific CT scales we used). But CT measures exhibit much better model fit in confirmatory factor analysis, a prerequisite for conducting valid reliability analyses. Thus, our literature review and survey analysis show that assessing the relative reliability of these scales is more complex than simply identifying the higher Cronbach’s alpha. Better tests of scale reliability could enhance CT survey research, and social science generally.