ROUNDING IN REPORTED TIMES: REQUESTS FOR ACCURATE REPORTING AND THE ORDER OF QUESTIONS

This study examines how the rounding characteristics of times reported differ when accurate reporting is requested and when the order of questions asked differs. In addition to a standard order of ‘(A1) departure time followed by arrival time’, ‘(B1) arrival time followed by departure time’ and ‘(C1) travel time followed by departure time’, are adopted. Two derivations of (A1) were prepared to investigate the impact of a request for accurate reporting. (A2) explicitly requests respondents to report times accurately to the nearest minute, and (A3) further requests them to refer to tools that can be used for recalling times. Descriptive analyses, Gini coefficients, and rounding probability models were employed for the investigation. Requests for accurate reporting resulted in mitigated rounding to the nearest multiples of 30 and 60 minutes, but not to the nearest multiples of 5, 10, and 15 minutes. Departure times reportedly have more rounding than arrival times. However, the arrival times in the (B1) variant had more rounding, suggesting that items asked earlier are more likely to be rounded. The results are not inconsistent with respondents recalling arrival times by adding travel time to departure time.