Skip to main content
. Author manuscript; available in PMC: 2022 Jan 6.
Published in final edited form as: Manage Sci. 2020 Oct 15;67(7):3985–4642. doi: 10.1287/mnsc.2020.3706

Table 2. Panel A. Regressions Predicting Participants’ Weekly Workouts during the Intervention Period.

This table reports a series of ordinary least squares regressions predicting a study participant’s weekly number of (a) overall workouts, (b) workouts initiated during their workout window, and (c) workouts initiated outside of their workout window during the four-week intervention period. In each column, we report the mean number of workouts completed by the control group within this period. The primary predictors included in these regressions are treatment status indicators, which indicate the size of the incentive offered for exercise ($3 versus $7) and the flexibility of the workout schedule (flexible versus routine). We report pairwise Wald tests to assess whether or not all paired regression coefficients reported differ significantly from each other.

(1) (2) (3) (4) (5) (6) (7) (8) (9)
Total Workouts Total In-Window Workouts Total Out-of-Window Workouts Total Workouts Total In-Window Workouts Total Out-of-Window Workouts Total Workouts Total In-Window Workouts Total Out-of-Window Workouts
Flexible Payment $3 0.58*** (0.14) 0.32*** (0.09) 0.27** (0.10)
Flexible Payment $7 0.89*** (0.14) 0.43*** (0.10) 0.50*** (0.10)
Routine Payment $3 0.40** (0.14) 0.57*** (0.10) −0.15 (0.09)
Routine Payment $7 0.69*** (0.14) 0.96*** (0.10) −0.21* (0.09)
$3 Interventions 0.49*** (0.13) 0.45*** (0.09) 0.06 (0.09)
$7 Interventions 0.79*** (0.13) 0.69*** (0.09) 0.15+ (0.09)
Flexible Interventions 0.74*** (0.13) 0.37*** (0.09) 0.38*** (0.09)
Routine Interventions 0.54*** (0.13) 0.77*** (0.09) −0.18* (0.09)

Mean Values of Control Group 1.11 0.59 0.59 1.11 0.59 0.59 1.11 0.59 0.59
Observations 10032 10032 10032 10032 10032 10032 10032 10032 10032
R-squared 0.07 0.07 0.10 0.07 0.05 0.04 0.07 0.06 0.10

Wald Test ($3 Flexible-$7 Flexible)
Difference in Coefficients −0.31*** (0.09) −0.11 (0.07) −0.24*** (0.07)
Wald Test ($3 Flexible-$3 Routine)
Difference in Coefficients 0.18* (0.09) −0.25*** (0.08) 0.41*** (0.05)
Wald Test ($3 Flexible-$7 Routine)
Difference in Coefficients −0.11 (0.09) −0.64*** (0.08) 0.47*** (0.05)
Wald Test ($7 Flexible-$3 Routine)
Difference in Coefficients 0.49*** (0.09) −0.14+ (0.08) 0.65*** (0.06)
Wald Test ($7 Flexible-$7 Routine)
Difference in Coefficients 0.20* (0.09) −0.53*** (0.08) 0.71*** (0.05)
Wald Test ($3 Routine-$7 Routine)
Difference in Coefficients −0.29** (0.09) −0.39*** (0.09) 0.06 (0.04)
Wald Test ($3-$7)
Difference in Coefficients −0.30*** (0.06) −0.25*** (0.06) −0.09* (0.04)
Wald Test (Flexible-Routine)
Difference in Coefficients 0.19** (0.06) −0.39*** (0.06) 0.56*** (0.04)

Note: Standard errors clustered by workout buddy pair are in parentheses.

+

p<0.10

*

p<0.05

**

p<0.01

***

p<0.001.

The control variables in the regressions are indicators for randomization strata (12 strata: three randomization dates, crossed with whether or not workout window perfectly overlapped with partner’s workout window, crossed with whether or not self-reported pair number of workouts per week was above median for randomization date), as well as an indicator for missing workout window.