Interpreting evidence from structural equation modeling in nursing practice

摘要

Structural equation modeling is a statistical technique that allows researchers to examine multiple hypotheses while simultaneously controlling for error. It can consist of a variety of observed and latent independent, mediator, and dependent variables. Owing to it being confirmatory in nature, this statistical approach can be used quite readily to test theoretical models. Likewise, it provides overall fit indices that determine whether the model tested actually fits the observed data. This article provides nurses with the basics on structural equation modeling so they can assimilate evidence from studies that use this statistical tool and be able to incorporate such findings into practice. Nursing research uses a variety of well-reasoned conceptual models to explain a number of phenomena ranging from agitation in nursing home residents (Vance et al., 2003) to predicting falls (Vance et al., 2006a), the effects of physical activity on cognition (Vance et al., 2005), and driving cessation in community-dwelling older adults (Vance et al., 2006b).