Original article

Scand J Work Environ Health 2003;29(6):444-451    pdf


Validity of a single-item measure of stress symptoms

by Elo A-L, Leppänen A, Jahkola A

Objectives The objective of the study was to investigate the content, criterion, and construct validity of a single-item measure of stress symptoms. Such a concise measure would be useful in monitoring stress at work. The criteria for validity were convergence with conceptionally close measures, the plausibility of associations with health and work characteristics, and the power to discriminate between groups.

Methods Four sets of independent cross-sectional data were used. The first data set, from Finland Post, comprised symptoms of ill health and mental resources (N=1014). The second, from four Nordic countries, included well-known validated scales on exhaustion, mental health, sleep, vitality, and optimism, and therefore the convergence between the measures could be studied (N=1015). The third, from a metal factory, included three indicators of health and four work characteristics (N=773). The fourth, representing the Finnish working population, described group differences in stress symptoms (N=2156) and allowed comparison with a study on emotional exhaustion in the working population. Distributions, correlations, and factor analysis were used for the study.

Results The stress-symptoms item converged with items on psychological symptoms and sleep disturbances and with validated measures of well-being. It had theoretically grounded associations with indicators of health and psychosocial work characteristics, and it discriminated between gender and age groups and industrial branches in accordance with the validated emotional exhaustion scale.

Conclusions The stress-symptoms item showed satisfactory content, criterion and construct validity for group-level analysis. It is suggested that the longer scales used to measure psychological stress can be replaced with it in survey research.

This article refers to the following texts of the Journal: 2002;28(6):418-428  2000;26(3):263-272