Portal de Periódicos da CAPES

Formal definitions of measurement bias and explanation bias clarify measurement and conceptual perspectives on response shift

2009; Elsevier BV; Volume: 62; Issue: 11 Linguagem: Inglês

10.1016/j.jclinepi.2009.03.013

1878-5921

Frans J. Oort, Mechteld R.M. Visser, Mirjam A. G. Sprangers,

Advanced Causal Inference Techniques

Objective Response shift is generally associated with a change in the meaning of test scores, impeding the comparison of repeated measurements. Still, different researchers have different views of response shift. From a measurement perspective, response shift can be considered as bias in the measurement of change, whereas from a more conceptual perspective, it can be considered as bias in the explanation of change. We propose definitions to accommodate both interpretations of response shift. Study Design and Setting Formal definitions of measurement bias and explanation bias serve to define response shift in measurement and conceptual perspectives. Examples from the field of health-related quality of life research illustrate the definitions. Results Definitions of response shifts as special cases of either measurement bias or explanation bias clarify different interpretations of response shift and lead to different research methods. Different structural equation models are suggested to investigate biases and response shifts in each of the two perspectives. Conclusion It is important to distinguish between measurement and conceptual perspectives as they involve different ideas about response shift. Definitions from both perspectives help to resolve conceptual and methodological confusion around response shift and to further its research.