Critical Review of an Instrument* by Sarah E. A. Woolsey, BA, BSN, RN
Psychometric instruments help researchers gather information for scientific study. A well-constructed instrument uses items to measure constructs or latent variables in phenomenon that are not otherwise measurable (DeVilles, 2012, pp. 11-12). This paper is a critical review of a psychometric instrument in use in nursing research called the Nurse Stress Index (NSI).
The NSI was developed over a three year period by Harris, Hingley, & Cooper (Harris, 1989; Williams & Cooper, 1997).
The ability to measure social and psychological phenomenon is paramount to scientist who want to work in research with intangible constructs (DeVilles, 2012). The NSI construct of interest is occupational nursing stress (Harris, 1989; Williams & Cooper, 1997).
Application & Target Population
The developers sought to create a “diagnostic measure for identifying sources of stress” in nurses with some managerial responsibility at work (Harris, 1989, p. 343).
The NSI was developed in three stages.
In Stage 1
exploratory work qualitative, unstructured interviews were conducted. Data was analyzed and 140 original items relating to occupational stress in nursing were pooled. After 140 items were distilled to 71 by a panel of experts, factor analysis was conducted and 8 factors and 52 items emerged (Harris, 1989, pp. 336-337; Williams & Cooper, 1997, p. 246).
of the NSI included re-examination of the 52 items identified in the factor analysis and comparing the results to sum pressure scores from the 71 Stage 1
items and a repeat survey (Harris, 1989, p. 342). Repeat factor analysis resulted in 44 items in 6 factors.
developed a short form of NSI that was reliable and valid. Expert nurses assisted the developers in reducing the 44 items to 30 items in 6 sub-scales with 5 items in each. Cronbach’s alpha was computed and all included items were found to be reliable (Harris, 1989, p. 338). The final short form was not re-tested.
Reliability is how well an instrument measures a latent variable or construct without error. Without reliability an instrument cannot be valid. To test reliability researchers use factor analysis, the reliability co-efficient alpha, and other methods to ensure an instrument scale and the items within it measure the construct (DeVilles, 2012; Portney & Watkins, 2009). The construct of occupational stress in nurses was tested for reliability during NSI development primarily through factor analysis and computation of coefficient alpha.
Validity is important in instruments to demonstrate that it is measuring what it is supposed to in a reasonably error-free way and that the latent variable is the cause of any variation (DeVilles, 2012; Portney & Watkins, 2009)Validity in the NSI was achieved utilizing through content validity, concurrent validity, and discriminate validity.
Content validity is achieved when items adequately sample content domain and meet investigator intent (DeVilles, 2012, p. 59). The NSI developers met content validity by including large numbers of the target population in item development. Experts were also utilized to review items (Note: more discussion is in the longer version of this article).
Norms refer to established “typical or standard” values for a population (Portney & Watkins, 2009, p. 304). No normative data was reported by Harris (1989).
The process of developing the NSI demonstrates a systematic and thorough method of psychometric instrument development beginning with qualitative work, item/factor development, and validation of factor constructs through statistical factor analysis.
Researchers and consumers of research should not just report reliability & validity for an instrument from other studies. Statistical reliability & validity testing should also be done on the sample population being studied.
* A longer version of this article is available from the author at firstname.lastname@example.org
Berman, B. A., Read, L. L., Marcus, A. C., & Gritz, E. R. (1992). Nurses enrolled in a stop smoking program: The role of occupational stress. Journal of Women's Health, 1
Costa Pereira, M., & de Sousa, S. (2011). Stressors in nurses working in intensive care units. Revista Latino-Americana De Enfermagem (RLAE), 19
(4), 1025-1032. doi:S0104-11692011000400023
DeVilles, R. F. (2012). Scale development: Theory and applications
(3rd ed.). Los Angeles: Sage.
Flanagan, N. A. (2006). Testing the relationship between job stress and satisfaction in correctional nurses. Nursing Research, 55
Harris, P. E. (1989). The nurse stress index. Work & Stress, 3
Healy, C. M., & McKay, M. F. (2000). Nursing stress: The effects of coping strategies and job satisfaction in a sample of australian nurses [corrected] [published erratum appears in J ADV NURS 2000 apr; 31(4): 989]. Journal of Advanced Nursing, 31
McGowan, B. (2001). Self-reported stress and its effects on nurses. Nursing Standard, 15
Michie, S., Ridout, K., & Johnson, M. (1996). Clinical management. stress in nursing and patients' satisfaction with health care. British Journal of Nursing (BJN), 5
Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory
(3rd ed.). New York: McGraw-Hill, Inc.
Portney, L. G., & Watkins, M. P. (2009). Foundations of cinical research: Applications to practice
(3rd ed.). Upper Saddle River, NJ: Pearson Prentice Hall.
Segerstrom, S. C. (2010). Resources, stress, and immunity: An ecological perspective on human psychoneuroimmunology. Annals of Behavioral Medicine, 40
(1), 114-125. doi:10.1007/s12160-010-9195-3
Selye, H. (1976). The stress of life
(revised ed.). New York: McGraw-Hill.
Tagliacozzo, R., & Vaughn, S. (1982). Stress and smoking in hospital nurses. American Journal of Public Health, 72
(5), 441-448. doi:10.2105/AJPH.72.5.441
Williams, S., & Cooper, C. L. (1997). Nurse stress index. In C. P. Zalaquett, & R. J. Wood (Eds.), Evaluating stress: A book of resources
(pp. 245-276). Lanham, Maryland: The Scarecrow Press, Inc.