Skip to: search, navigation, or content.


Indiana University Bloomington

Center for the Business of Life Sciences

Email Updates

Enter email below to sign up:

The Kelley Advantage

Our Life Sciences Research Fellows are academic all-stars who focus on the business and economic issues and challenges faced by life sciences companies.

Publications

Journal Articles

Results on the Standard Error of the Coefficient Alpha Index of Reliability

2005, Marketing Science

Adam Duhachek, Anne T. Coughlan, Dawn Iacobucci

Abstract

In this research, we investigate the behavior of Cronbach's coefficient alpha and its new standard error. We systematically analyze the effects of sample size, scale length, strength of item intercorrelations, and scale dimensionality. We demonstrate the beneficial effects of sample size on alpha's standard error and of scale length and the strengths of item intercorrelations (effects that are substitutes in their benefits) on both alpha and its standard error. Our findings also speak to this adage: Heterogeneity within the item covariance matrix (e.g., through multidimensionality or poor items) negatively impacts reliability by decreasing the precision of the estimation. We also examined the question of "equilibrium" scale length, showing the conditions for which it is optimal to add no items, or one, or multiple items to a scale. In terms of "best practices," we recommend that researchers report a confidence interval or standard error along with the coefficient alpha point estimate.

Citation

Duhachek, Adam, Anne T. Coughlan and Dawn Iacobucci (2005), “Results on the Standard Error of the Coefficient Alpha Index of Reliability,” Marketing Science, Vol. 24, No. 2, pp. 294-301.

Notations

Reprinted in Recherche et Applications en Marketing, Journal of the French Marketing Association (2006).

.