Using Cronbach's Alpha in Research

Cronbach’s Alpha: Use In Research

According to Taber (2017), when used in educational research, Cronbach’s Alpha is often cited to validate findings as internally reliable. However, a systematic review of studies in the field often finds significant deficits in terms of how Cronbach’s alpha is used. For example, in one study cited by the author, Cronbach alpha values were reported for each scale composing the instrument designed to measure longitudinal self-regulation of biology learning among junior high students. But Taber (2017), upon deeper analysis of the article, found statistically that the value of the alpha (and thus scale reliability) was reported as higher than what researchers calculated as part of the different scales (Taber, 2017).

Taber (2017) also noted within the systematic review that while Cronbach’s alpha is often cited as being used to validate a study’s reliability, this is often done without explaining how this is the case, or with any interpretation or commentary as if the use of the metric is self-evident. Another issue noted as particularly problematic in secondary school research was when there were different levels of reliability within categories studied within the research, based upon unique characteristics of those subcategories.

For example, in another research study of students’ levels of enthusiasm towards school, younger students showed a higher level of enthusiasm than older students, but younger students were often inclined to report their feelings pertaining to activities centered on simply being in school (versus seeing their friends), versus the educational environment itself, which called into question conclusions made generalizing the findings between such groups.

In short, Cronbach’s alpha, the author concluded, is often cited in educational research as a way of ensuring a lack of bias and other influences that can affect reliability, but it alone cannot ensure uniform scale dimensionality (Taber, 2017). Nor does its use automatically guarantee that the research findings are accurate, even when contained in peer-reviewed journals.