The following discussion concerns a fictional teaching dataset about commuting and student attendance. An online survey received 180 responses from students at one university during a single week in March. Sixty respondents reported a journey of more than forty-five minutes each way. In that group, 35 per cent said they had missed at least one morning class in the previous month. Among the remaining 120 respondents, the corresponding figure was 20 per cent. The survey recorded self-reported travel time and attendance; it did not use transport records or electronic registers to verify either measure.

The difference between these percentages suggests an association that deserves further investigation. It does not show that a long journey caused a student to miss a class. Employment, caring responsibilities, disability, timetable design and the availability of online materials could influence both commuting patterns and attendance. None of those factors was controlled in the survey. A respondent might also classify an occasional journey as typical, or remember an absence more readily than a journey made without difficulty. These possibilities affect the interpretation even if every response was submitted in good faith.

The recruitment method creates another limitation. A link to the survey was distributed through two student mailing lists, and participation was voluntary. Students who had experienced travel problems may have been more motivated to respond. Conversely, students under the greatest time pressure may have ignored the invitation. The sample therefore cannot be assumed to represent all students at the university, much less students at other institutions. The number of responses is useful for a classroom analysis, but a larger number alone would not remove a systematic recruitment bias.

A stronger study would recruit students across departments, record relevant background factors and follow attendance over a longer period. It could compare different timetables while taking account of the reasons students choose particular classes. Until such evidence is available, the survey can inform questions for further research and practical conversations about morning scheduling. It should not be used to claim that reducing every commute below forty-five minutes would produce a particular improvement in attendance. That prediction requires evidence the current dataset does not provide.
