Conference Agenda

Data quality assessment is crucial to ensure reliable results in applied research. In this study, we examine the effects of data quality by analyzing paradata, specifically focusing on response times in both the survey and the DCE tasks. Unlike previous studies that rely on less precise, section-level timings or fixed thresholds, we use detailed time measurements that allow for a more sensitive detection of both, unusually fast and unusually slow responses. In addition, we propose a multidimensional classification of respondents that combines these behavioural measures with self-assessment variables such as “not understanding the DCE” and “finding the DCE difficult to understand”, providing a cognitive dimension to the evaluation of engagement and response quality. Our findings indicate a high prevalence of inattentive behavior, with almost 30% of respondents completing the DCE too quickly, spending one second or less on at least two choice tasks and failing to read the attribute descriptions. The results also indicate that speeding is not evenly distributed across respondents. Individuals under the age of 40 and those with medium–low level of education are more likely to rush through the DCE, suggesting that inattentiveness is associated with socio-demographic characteristics. Hence, we encourage the use of paradata in future studies to complement that information obtained via survey responses.