'Don't know' as lack of trust: Findings from developing a health research knowledge scale.

Abstract

We aimed to develop and test a scale that assesses knowledge of medical and health research for the Singaporean general adult population and examine the implications of 'don't know' (DK) responses from participants in the knowledge assessment. Two studies were conducted among Singaporean community adults. Items in the scale were selected via exploratory factor analysis in a pilot study (N = 223) and evaluated via confirmatory factor analysis and item response modelling in a larger population survey study (N = 999). We also explored characteristics of DK responses. A two-factor structure emerged and reflected two pathways of knowledge formation, including acknowledgement of facts and resistance of myths in medical and health research. The scale showed satisfactory structure validity, reliability, item-level properties, and criterion validity. The two subscales (Facts and Myths) also achieved measurement invariance across genders. Treating DK responses as missing reduced the psychometric quality. DK response tendency in both Facts and Myths subscales was primarily and significantly associated with lower trust in research, suggesting that DK and incorrect answer might not be results of the same response process. DK responses were non-ignorable in the scoring in assessing knowledge of medical research and carried implications separate from incorrect answers. Individuals answering DK may entail both those who truly lack knowledge and those who do not wish to disclose their views. Patient and community engagement strategies for improving the public knowledge of medical and health research should not only consider communicating facts but also strategies to correct misinformation. Assessing knowledge in health should consider the implications of patients' DK responses.