Title

Computer self-efficacy in professional nurses: An analysis of selected factors using latent variable structural modeling

Date of Completion

January 1992

Keywords

Education, Tests and Measurements|Education, Educational Psychology|Health Sciences, Nursing|Psychology, Psychometrics

Degree

Ph.D.

Abstract

The focus of this research is the formulation of a structural equation model that estimates the degree to which certain factors may influence judgments of self-efficacy related to computer use. The target group comprised 1,200 randomly-selected critical care nurses, i.e., professionals who work in highly technical patient care settings. A combination of social cognitive, educational, and attitudinal theories (Bandura, 1986; Fishbein & Ajzen, 1975; Kolb, 1984; Schunk, 1985) were employed in the creation of a structural model that depicts relationships between and among six latent variables. Specifically, the latent variables are: abstract-concrete and active-reflective cognitive style aptitudes, math ability aptitudes, life experiences, computer attitudes, and computer self-efficacy. The influence on individuals' perceptions of self-efficacy relative to using computers in the work setting is examined using multiple indicators derived from scores on three standardized measures and demographic information.^ Questionnaires were mailed to 1,200 professional nurses working in the Northeastern region of the United States who are currently members of the American Association of Critical-care Nurses. A return rate of 43 per cent was attained with 498 complete questionnaires available for data analysis. The proposed structural equation model (SEM) was specified, estimated, and tested through the use of LISREL VI (Joreskog & Sorbom, 1986).^ Thirty-seven observed variables were measured to estimate the causal parameters linking the six latent variables. After respecification of the measurement model, and subsequently the structural model, the SEM of computer self-efficacy demonstrated adequate fit ($\chi\sp2$(20, N = 470) = 33.00 $p=.034$) with a normed fit of.971 which is indicative of both a good model fit and that the data support the study hypothesis, i.e., computer self-efficacy reflects positive attitudes about the use of computer technology, previous learning and frequent use of computers, and indirectly, math ability aptitudes in this sample of critical care nurses.^ The findings indicate that the Computer Self-Efficacy Scale (CSE) is a valid and reliable instrument for the evaluation of basic, advanced, and mainframe computer skill competencies. Further, study results offer empirical foundations important in the development of and evaluation of instructional programs associated with the implementation of computer use or information systems, in addition to identifying specific factors that contribute most to one's sense of efficacy in the attainment of computer skills. ^