Date of Award

2013

Document Type

Thesis

Publisher

Edith Cowan University

Degree Name

Doctor of Philosophy

School

School of Exercise and Health Sciences

Faculty

Faculty of Health, Engineering and Science

First Supervisor

Professor Donna Cross

Second Supervisor

Winthrop Professor Stephen Zubrick (UWA)

Abstract

Bullying at school is associated with negative social, psychological and academic outcomes for both the victimised student and the perpetrator. As a consequence, national strategies to address bullying have been implemented in numerous countries, and education sectors and schools have increasingly directed resources to the problem. To ensure resources are allocated to programs and strategies that will prevent and successfully respond to bullying in schools, and therefore prevent harm to students, evidence of program effectiveness is required. Evaluations of anti-bullying programs in schools have had mixed results and there is a lack of robust evidence as to their impact. Several factors have been proposed and investigated as explanations for the varying effectiveness of programs, including the lack of strong study designs and rigorous methods in many program evaluations. The aim of this doctoral research was to investigate methodological challenges identified in the reviews of evaluations of school anti-bullying programs and to determine the implications of these challenges for such evaluation studies. Recommendations for future methodological research practice were also developed. Threats to the validity of findings from evaluations of anti-bullying programs, as identified in the literature reviewing these evaluations, were investigated within a theoretical framework for assessing the methodological quality of program evaluations. This research utilised existing large data sets from three studies, one national cross-sectional survey and two group-randomised controlled trials. Self-report measures of bullying victimisation and perpetration comprised the key outcomes. Various advanced statistical techniques were applied to investigate specific threats to the construct, internal and statistical validity of findings from an evaluation. Approval for research was obtained from the relevant authorities and conducted according to the appropriate ethical standards. This research details how the validity of findings from a program evaluation can be threatened as a result of the instruments used to measure bullying behaviours; response shift bias or raised awareness of bullying within the intervention group; active only parental consent procedures; the use of data analysis methods unsuited to bullying outcomes; and insufficient sample sizes. Strategies for minimising these threats, such as suitable approaches to sample size determination and the use of novel statistical techniques, are described. With careful planning and sufficient resources, evaluations of bullying programs in schools can provide valid evidence of their efficacy. Apart from the usual methodological considerations, the limitations of many evaluation studies can be addressed through the use of strong study designs; suitable methods to determine sufficient sample sizes; strategies to maximise response and consent rates when recruiting subjects; appropriate instruments to measure bullying behaviours; measures to assess and account for response shift; and appropriate data analysis methods.

Share

 
COinS