The Psychology of Religion has had a tradition measuring things that are seemingly difficult to measure, such as one’s level of religiosity or spirituality – concepts that are polysemantic, so do not have a simple definition. In doing so, researchers have developed scales to measure such constructs. This trend in Psychology of Religion research will continue as researchers start to conduct studies in non-Western Educated Industrialized Rich Democratic (WEIRD) populations, as they will require new scales that are appropriate for the way that these populations conceptualize religiosity and spirituality. Scale construction takes multiple steps, which most researchers are well-acquainted with. One important step is to demonstrate both validity and reliability. In the Psychology of Religion, an overwhelming majority of researchers use Cronbach’s alpha to estimate scale reliability. However, alpha has multiple preconditions that can easily be violated in psychology research that are rarely tested for, let alone adjusted for. Much like with using parametric statistical analyses when parametric assumptions are violated, this approach may be leading to inaccurate reporting. This article will discuss; (1) whether alpha should be used at all in Psychology of Religion research and, if so, when it is appropriate; (2) introduce another estimate of scale reliability, Omega, and show how and it might be calculated; (3) provide examples of how these techniques might be taught to students studying Psychology of Religion at undergraduate and higher levels. In doing so, I hope to move the entire field of Psychology of Religion forward in terms of methodological rigor.
|Journal||Archive for the Psychology of Religion|
|Publication status||Submitted - 13 Sep 2020|
- Cronbach’s Alpha
- Correlated Error
- Scale Reliability