A Monte Carlo Study of Eight Confidence Interval Methods for Coefficient Alpha

Abstract

The purpose of this research is to examine eight of the different methods for computing confidence intervals around alpha that have been proposed to determine which of these, if any, is the most accurate and precise. Monte Carlo methods were used to simulate samples under known and controlled population conditions. In general, the differences in the accuracy and precision of the eight methods examined were negligible in many conditions. For the breadth of conditions examined in this simulation study, the methods that proved to be the most accurate were those proposed by Bonett and Fisher. Larger samples sizes and larger coefficient alphas also resulted in better interval coverage, whereas smaller numbers of items resulted in poorer interval coverage.

Supplemental Material for Confidence Intervals for Population Reliability Coefficients: Evaluation of Methods, Recommendations, and Software for Composite Measures

“…Table 4 shows η 2 for all main and interaction effects of each design condition on the coverage rates of the confidence intervals for coefficient alpha on population reliability. Note that the results from this study are different from previous studies (e.g., Cui & Li, 2012, Duhachek & Iacobucci, 2004, Padilla, Divers, & Newton, 2012, Romano, Kromrey, & Hibbard, 2010, and Maydeu-Olivares, Coffman, & Hartmann, 2007 that find the coverage rates of coefficient alpha on the population coefficient alpha.…”

Dos versiones breves del Big Five Inventory en universitarios peruanos: BFI-15p y BFI-10p

“…Se utilizó el coeficiente α y sus intervalos de confianza (IC; Dominguez-Lara & Merino-Soto, 2015) con el método de Fisher (1950). Este método es el que brinda valores más estables independientes del número de ítems (Romano, Kromrey, & Hibbard, 2010;Romano, Kromrey, Owens, & Scott, 2011). Para las medidas de dos ítems se empleó el coeficiente de Angoff-Feldt (r AF ; Dominguez-Lara, Merino-Soto, & Navarro-Loli, 2016;Feldt & Brennan, 1989), dado que reduce el impacto de la heterogeneidad de varianzas de los ítems.…”

Big Five Inventory en Universitarios Peruanos: Resultados Preliminares de su Validación

“…Se utilizó el coeficiente alfa de Cronbach y sus intervalos de confianza (IC; Dominguez-Lara & Merino-Soto, 2015) con el método de Fisher (1950). Este método es el que brinda mejores estimaciones (Romano, Kromrey & Hibbard, 2010;Romano, Kromrey, Owens & Scott, 2011), esperando magnitudes ≥ 0,70 (Merino Soto, Navarro Loli & García Ramirez, 2014). Del mismo modo, se calculó alfa ordinal (Dominguez-Lara, 2012;Elosua Oliden & Zumbo, 2008), el coeficiente Omega (ω; McDonald, 1999) y el coeficiente H para conocer la confiabilidad de los constructos (Dominguez-Lara, 2016;Hancock & Mueller, 2001).…”