Change in behavior during the non-treatment period
| Ate a serving | Number of servings eaten | Number of servings discarded | |
|---|---|---|---|
| A. First five days at all schools (N=75) | |||
| Time trend (days) | −0.0161 [0.0107] | −0.0134 [0.0133] | −0.0057 [0.0108] |
| B. Full set of days at control schools (N=32) | |||
| Time trend (days) | −0.0019 [0.0053] | −0.0016 [0.0074] | 0.0029 [0.0042] |
Notes: The unit of analysis is the school-day observation. Each cell is a separate regression and the reported number is the coefficient on a variable indicating the observation day number of each school which varies from one to five in panel A and from one to twelve in panel B. The coefficient on the observation day number gives an idea for the time trend effect of our presence as researchers. Each regression includes school and day of the week fixed effects and a control for whether only vegetables were offered that day.