Resumen

VARGAS HALABI, Tomas  y  MORA-ESQUIVEL, Ronald. Sample sizes using structural equation modeling with latent variables: A practical method. Rev. Actual. Investig. Educ [online]. 2017, vol.17, n.1, pp.25-60. ISSN 1409-4703.  http://dx.doi.org/10.15517/aie.v17i1.27294.

Structural equation modeling is a useful technique for educational researchers to testing hypotheses about relations among latent constructs and to clarify different types of effects. However, the estimation of sample size to run properly this technique has not yet provided an adequate response in the literature. Although a variety of criteria for its estimation is available, the literature has not yet provided a practical method to summarize them and to provide an argument of the appropriate size. Accordingly, this article aims to provide a practical guidance to educational researchers to use different criteria and justify the appropriate sample size calculation in structural equations modeling by planning, a priori, measurement and structural models. To calculate sample sizes a quantitative method is utilized to determine inputs based on path diagrams of a priori models of research, as well as interfaces available online. Data are summarized in tabular form for comparative purposes. The method has been useful to define a minimum number of cases to satisfy different criteria and has shown its advantage than relying on a single rule or criterion to justify the appropriate sample size. It also reveals the advantage using parcels in planning prior models with this technique.

Palabras clave : statistical analysis; models; structural equations; sample size.

        · resumen en Español     · texto en Español     · Español ( pdf )