The following discussion focuses on a fictional data set on commuting and student attendance. An online survey collected 180 student responses during one week in March. Of those respondents, 60 reported an average one way commute time of 45 minutes or more. Of that 60, 35 (or 57%) reported missing at least one morning class during the previous month. The remaining 120 respondents (of the 180 total), had a 20% response rate of class absences in the previous month. Attendance was self-reported and travel time was reported by the respondents and not validated by transportation records or electronic registers.

 A possible explanation for the difference between those percentages may be worth exploring, but it doesn't indicate a long commute caused the absence. There are many other possible reasons that are beyond the scope of the survey. For example, respondents may have a long commute because of work or caring responsibilities. A respondent may have a disability, or an irregular commute, or may have an irregular class schedule. It's also possible respondents remembered a class absence more easily than a commute. It is possible, even if there was no attempt at deceit in the responses, the responses may be biased due to the reasons mentioned above.

The method of recruitment also creates a restriction. A survey link was posted to two student email lists, and completing the survey was voluntary. Some students may have had travel problems and been more motivated to respond to the survey. Others, who may have had greater time constraints, may have chosen to not respond to the invitation. Therefore, the sample cannot represent students at the university, much less students at other institutions. The sample is sufficient only for a classroom case study. A larger sample would not remove a systematic recruitment bias.

A stronger study would recruit students from various academic units, take into consideration background factors, and follow attendance over a long period. This study could take into consideration the reasons why students pick particular classes and timetables. Until this evidence is available, the survey should inform potential research questions and practical discussions regarding morning scheduling. The survey cannot be used to support the claim that reducing all commute times to less than 45 minutes would result in a given improvement in student attendance. The data does not support this.