Time Inconsistency in the Classroom

The intervention imposes soft requirements on students within treated sections to visit TA office hours in certain weeks of the semester.

2023-01-18

2023-05-12

1. Number of TA sessions attended.
2. Number of weekly Problems & Applications assignments turned in on time or prior to the relevant test.
3. Final grade in the course.

#1 will be tracked via logs of all attendees at all TA sessions.
#2 Number of weekly P&A assignments turned in on time is based on the scheduled due date. The number turned in prior to the relevant test is a slightly more forgiving measure, as someone might be a few days late but still complete the assignment while studying for the test.

1. Grade on the final exam.
2. Grade on final exam conditional on the first three weeks of homework and Test 1 grade.
3. Reenrollment at the university in the subsequent semester.

The intervention described above will be conducted in two sections out of three course sections offered this semester. Students in those sections will be asked to visit TA office hours in the first week of the semester and whenever they have failed to complete three weekly assignments by the time that they are tested on the corresponding material.

This study is conducted in the setting of an Introductory Economics course at a large public university. Three sections of the course each have 220-300 students enrolled and are taught with the same policies, course schedule, and assignments. The course has three types of assignments that are each collected weekly during 9 weeks of the semester. This study is focused on one of those assignment types, called "Problems & Applications" (henceforth "P&A") which are due on 9 Saturdays during the semester.

As part of a broader project studying students' time inconsistency and biased recall regarding homework completion, we will collect survey data at 3 points of the semester (beginning, middle, and end) to assess students' expectations (before) and recall (after) of how many course assignments they are turning in on time. Specifically, we will ask students how many upcoming P&A assignments they expect they will complete on time, and how many of the past P&A assignments they recall completing on time. In addition, we will hand-collect data on the visits made by students to any of several weekly TA sessions held in the same room in the Economics department.

The intervention described above will be conducted in two sections of the course. Students in those sections will be asked to visit TA office hours in the first week of the semester and whenever they have failed to complete three weekly assignments by the time that they are tested on the corresponding material.

The choice of sections to be treated was not randomized but assigned such that the professor teaching two sections would have one treated and one control section, while also ensuring that of the two class sections meeting in the afternoon (as opposed to the morning) there would be one treated and one control section. This allocation was chosen in order to minimize congestion and the student:TA ratio at the TA sessions, which are the same for all three sections.

Course sections.

Yes

3 course sections (2 treated; 1 control)

820 students (520 treated; 300 control)

1 control section, 2 treatment sections

Our main outcomes are (1) number of TA sessions attended per student, (2) number of P&A assignments completed on time, and (3) course grade. (1) TA sessions: We estimate based on past semester survey data that the average number of visits in the control group will be 0.44 (SD 1.44, ICC 0). We have 80% power to detect an increase of 0.29 in the mean TA sessions attended. This is a large increase in relative terms because of the existing low use of TA sessions. However, our treatment imposes a requirement to visit TA office hours once after each test if not all prior assignments were turned in before the test. Based on P&A assignment completion last semester, we estimate this requirement will fall upon 31%-50% of students after each test, raising TA visits by 1.2 if there is full compliance. Thus, even with partial compliance of around 25%, we should be powered to detect the change. (2) P&A assignments completed on time: We estimate based on past semester gradebook data that the mean in the control group will be 6.7 (SD 2.89). Accounting for intra-cluster correlation of 0.0024, as estimated from prior data, we have 80% power to detect an increase of 0.76 P&A assignments completed on time, an 11% increase. (3) Course grade: Based on last semester’s gradebook outcomes, we expect a mean grade of 79.7 (SD 17.41) in the control section. Accounting for intra-section correlation of 0.0127, we have 80% power to detect a mean grade increase of 7.6, approximately a 10% increase. If the intra-section correlation were close to zero, we would be powered to detect a smaller increase, approaching 3.5 points (4.3 percent increase) as the correlation approaches zero.