In this case study, we examine a novel aspect of data collected in a typical probability and a typical nonprobability panel: mobile app data. The data were collected in Great Britain in 2018, using the Innovation Panel of the UK Household Longitudinal Study and the Lightspeed online access panel. Respondents in each panel were invited to participate in a month-long study, reporting all their daily expenditures in the app. In line with most of the research on nonprobability and probability-based panel data, our results indicate differences in the data gathered from these data sources. For example, more female, middle-aged, and highly educated people with higher digital skills and a greater interest in their finances participated in the nonprobability app study. Our findings also show that resulting differences in the app spending data are difficult to eliminate by weighting. The only data quality aspect for which we do not find evidence of differences between the nonprobability and probability-based panel is behavior in using the spending app. This finding is contrary to the argument that nonprobability online panel participants try to maximize their monetary incentive at the expense of data quality. However, this finding is in line with some of the scarce existing literature on response behavior in surveys, which is inconclusive regarding the question of whether nonprobability online panel participants answer questions less conscientiously than probability-based panel respondents. Since the two panels in our case study differ in more aspects than the sample selection procedure, more research in different contexts is necessary to establish generalizability and causality.