Volume 83, Issue 4 p. 955-979

The impact of methodological moderators on prevalence rates of workplace bullying. A meta-analysis

Morten Birkeland Nielsen

Corresponding Author

Morten Birkeland Nielsen

Department of Psychosocial Science, University of Bergen, Norway

Correspondence should be addressed to Morten Birkeland Nielsen, Department of Psychosocial Science, University of Bergen, Christies Gate 12 (4 etg), 5055 Bergen, Norway (e-mail: [email protected]).Search for more papers by this author
Stig Berge Matthiesen

Stig Berge Matthiesen

Department of Psychosocial Science, University of Bergen, Norway

Search for more papers by this author
Ståle Einarsen

Ståle Einarsen

Department of Psychosocial Science, University of Bergen, Norway

Search for more papers by this author
First published: 01 February 2011
Citations: 350

Abstract

The aim of this study was to investigate how different measurement methods and sampling techniques contribute to the observed variation in prevalence rates of workplace bullying. A total of 102 prevalence estimates of bullying from 86 independent samples (N=130,973) were accumulated and compared by means of meta-analysis. At an average, the statistically independents samples provided an estimate of 14.6%. Yet, the findings show that methodological moderators influence the estimated rates. As for measurement method, a rate of 11.3% was found for studies investigating self-labelled victimization from bullying based on a given definition of the concept, whereas a rate of 14.8% was found for behavioural measure studies, and 18.1% for self-labelling studies without a given definition. A difference of 8.7% points was found between randomly sampled and non-randomly sampled studies. When controlling for geographical differences, the findings show that geographical factors also influence findings on bullying. Hence, findings from different studies on workplace bullying cannot be compared without taking moderator variables into account.