BACKGROUND: In 1980 the Waddell score, consisting of 8 non-organic or behavioural signs, was developed to measure illness behaviour in patients with low back pain. There is some debate about whether the Waddell score is a valid screening instrument for illness behaviour and psychological distress, or whether it merely reflects elevated pain levels and diminished functional physical capacities. OBJECTIVE: The purpose of this study was to examine the construct validity of the Waddell score. METHODS: In this cross-sectional study, a total of 20 hypotheses about the associations between the Waddell score and measures from different domains were formulated a priori, based on a Medline database search (1980-2010). These hypotheses were tested in a sample of 229 patients with chronic low back pain who attended an outpatient rehabilitation centre. RESULTS: The percentage of hypotheses that were confirmed for the association between the Waddell score and the domain pain was 100%, for the domain physical 80%, for the domain illness behaviour 80% and 50% for the domain psychological. Correlation coefficients and kappa values varied between 0.06 and 0.44 for the measures that were expected to be associated with the Waddell score. CONCLUSION: Most of our challenging a priori hypotheses were accepted, and the Waddell score was found to have satisfactory cross-sectional construct validity. However, the presence of Waddell signs does not indicate exactly what the specific problems are and must therefore be conceptualized and understood in the total clinical picture of the patient. The association between the Waddell score and measures from different domains is weak. The Waddell score cannot be regarded as a straightforward psychological "screener". © 2012 Lippincott Williams & Wilkins, Inc.