Institutional members access full text with Ovid®

Share this article on:

A Multilevel Confirmatory Factor Analysis of the Practice Environment Scale: A Case Study

Gajewski, Byron J.; Boyle, Diane K.; Miller, Peggy A.; Oberhelman, Frances; Dunton, Nancy

doi: 10.1097/NNR.0b013e3181d1a71e

Background: Practice Environment Scale (PES) data are collected from RNs in nursing units in hospitals that are members of the National Database of Nursing Quality Indicators (NDNQI). Patient and RN information are collected to aid in quality improvement and research at the nursing unit level. The data were collected from the individual RN, but items are worded so that analyses can be conducted at the individual, unit, or hospital level. There is a need to examine the validity of the PES at both the individual and the unit level.

Objective: To describe multilevel confirmatory factor analysis via a case study for investigating the validity of the PES, a measure of the nursing practice environment.

Approach: The PES was administered to 72,889 RNs from 4,783 nursing units (16 unit types; e.g., critical care and obstetric) in 2007. The PES has 31 items in five different domains. A multilevel confirmatory factor analytic model was fit with a structure on the basis of the five domains. From this model, an estimate was sought between unit loadings and within unit loadings to investigate factorial, convergent, and discriminant validity at both the unit and the RN levels. To investigate criterion-related validity, the five PES domains were correlated with the seven job enjoyment items adapted from the National Database of Nursing Quality Indicators at the unit and RN levels (also using a multilevel model).

Results: The multilevel factor analysis provides evidence of factorial, convergent, discriminant, and criterion-related validity at both the unit and the RN levels.

Discussion: The PES is a valid instrument for use in quality improvement and research both at the unit and individual RN levels.

Byron J. Gajewski, PhD, is Associate Professor, Department of Biostatistics, University of Kansas Schools of Medicine and Nursing.

Diane K. Boyle, PhD, RN, is Associate Professor; Peggy A. Miller, PhD, RN, is Senior Scientist; Frances Oberhelman, MSN, RN, is Graduate Research Assistant; and Nancy Dunton, PhD, is Research Professor, University of Kansas School of Nursing.

Editor's Note Materials documenting the review process for this article are posted at

Accepted for publication October 31, 2009.

Partial funding for all authors comes from a contract with the American Nurses Association.

Corresponding author: Byron J. Gajewski, PhD, University of Kansas Medical Center, 3901 Rainbow Boulevard, MS 1026, Kansas City, KS 66160 (e-mail:

© 2010 Lippincott Williams & Wilkins, Inc.