On the Use, the Misuse, and the Very Limited Usefulness of Cronbach’s Alpha

By Klaas Sijtsma. Psychometrika, 2008

2,612 citations recorded by the source corpus.

This discussion paper argues that both the use of Cronbach’s alpha as a reliability estimate and as a measure of internal consistency suffer from major problems. First, alpha always has a value, which cannot be equal to the test score’s reliability given the interitem covariance matrix and the usual assumptions about measurement error. Second, in practice, alpha is used more often as a measure of the test’s internal consistency than as an estimate of reliability. However, it can be shown easily that alpha is unrelated to the internal structure of the test. It is further discussed that statistics based on a single test administration do not convey much information about the accuracy of individuals’ test performance. The paper ends with a list of conclusions about the usefulness of alpha.

Subfields: Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/d3187c719d5262f93a6f8297c475b917791fa5b8

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Psychological measurement and the replication crisis: Four sacred cows. (2020, background): “For example, we do not address widely shared but erroneous assumptions regarding the interpretation of specific statistics, such as the presumption that Cronbach’s alpha is an adequate index of test homogeneity (Flake, Pek, & Hehman, 2017; Sijtsma, 2009).”
One Size Doesn’t Fit All: Using Factor Analysis to Gather Validity Evidence When Using Surveys in Your Research (2019, background): “We refer interested readers to Bandalos (2018), Sijtsma (2009), and Crocker and Algina (2008). would only be affected by their actual goal endorsement.”
Interest development: Arousing situational interest affects the growth trajectory of individual interest (2017, methodology): “The coefficient H is considered a more accurate measure of reliability than the much-used Cronbach’s alpha (Hancock & Mueller, 2001; Sijtsma, 2009).”
Rethinking Internal Consistency in Cronbach's Alpha (2017, positive): “Several authors (Cortina, 1993; Green, Lissitz, & Mulaik, 1977; Ritter, 2010; Schmidt, Le, & Iles, 2003; Schmitt, 1996; Sijtsma, 2009), however, have highlighted the limitations and misuses of alpha in applied research.”
A Latent Class Approach to Estimating Test-Score Reliability (2011, background): “This misconception has persisted despite persuasive warnings by, for example, Cortina (1993), Schmitt (1996), and Sijtsma (2009).”
Motivating Employees of the Public Sector: Does Public Service Motivation Matter? (2010, background): “On the other hand, coefficient of reliability is often misleading (Sijtsma 2009) and is highly influenced by the number of items (Cortina 1993).”
ASSESSING REFLECTIVE MODELS IN MARKETING RESEARCH: A COMPARISON BETWEEN PLS AND PLSc ESTIMATES (2018, background): “Specifically, Cronbach's alpha tends to underestimates the true reliability, and should therefore be regarded as a lower boundary to the reliability (Sijtsma, 2009).”
Applying Bifactor Statistical Indices in the Evaluation of Psychological Measures (2016, background): “One candidate index is ECV (Sijtsma, 2009; ten Berge & Soĉan, 2004), which also can be used to judge the essential unidimensionality of the modeled common variance in an item set.”
The System Usability Scale: Past, Present, and Future (2018, background): “Despite some criticisms against its use (Sijtsma, 2009), it has a mathematical relationship to more direct estimates of reliability (e.g., test – retest) in that it provides a lower bound estimate of reliability.”
The Effects of Participation Rate on the Internal Reliability of Peer Nomination Measures (2013, background): “2 There are certainly a wide variety of alternatives to Cronbach’s α (see Green, 2003 and Sijtsma, 2009 for example), but the reliability estimate itself is not the focus of the current work.”
Coefficient Alpha: Interpret With Caution (2013, critical): “Many authors have pointed out that alpha is inappropriately used in studies mainly because satisfactory, or even high, values of alpha are persistently and incorrectly used as an indication of the unidimensionality of the scale (Cortina, 1993; Nunnally & Bernstein, 1994; Schmitt, 1996; Sijtsma, 2009”
HTMT2-an improved criterion for assessing discriminant validity in structural equation modeling (2021, neutral): “This requirement is also known from the calculation of Cronbach’s alpha which also requires that all indicators are positively correlated (Sijtsma, 2009).”