Are self-reported times rounded? Insights from times reported by an objective third party

Abstract

Data reported by respondents often include rounding errors. In travel surveys, people do not report their exact departure or arrival times, and they often round the times to nearest multiples of 5, 10, 15, 30, and 60 min. Previous studies have estimated rounding probabilities by assuming that while the actual times are evenly distributed between 0 and 59 min, the reported times are rounded. However, no studies have confirmed this assumption that actual times are evenly distributed. The present study aims to confirm the assumption by utilising times reported by a third party. The data come from the Japanese Prime Minister’s minute-by-minute activity schedule published daily in newspapers. The data covers 90 days between 27 June and 24 September 2021. The present study confirmed that times relating to travel (departure and arrival times) were evenly distributed. However, times relating to some non-travel behaviours were not evenly distributed. The study supports the assumption that departure and arrival times are evenly distributed, but researchers must consider the nature of the events.