A spotlight on cross-sectional research: Addressing the issues of confounding and adjustment

Abstract

Purpose: This study attempted to conceptualize confounding and guide healthcare management researchers to identify and adjust for potential confounders from the point of view of their theoretical frameworks as a basis of achieving internal validity. Design/methodology/approach: A comprehensive statistical procedure was applied to demonstrate how likely it is for misleading effect sizes to be reached in a cross-sectional healthcare management study as a result of failing to adjust for potential confounders. Participants of the study were 403 volunteer health workers in selected hospitals in Accra, Ghana. A self-reported questionnaire was used to gather data. Confirmatory factor analysis and linear regression analyses were used to present findings. Findings: The study revealed that failing to adjust for potential confounders can over- or under-estimate the ‘ultimate effect’, which was under-estimated in this study by a value of β =.006. The auxiliary regression estimates (e.g. the variance) were also wrongly estimated owing to failure to adjust for potential confounders. Originality/value: This study is the first of its kind to address the issue of confounding in healthcare management research and justifies how necessary it is for every study to control for all confounding variables.