| Routine |
Portion size of each person’s meal was estimated based on indicator variables indicating whether the meal was a breakfast, lunch or dinner |
The concept of chronotypes—that there are time-based patterns in eating and obesity. e.g., Fleig and Randler used food logs and found differences in diet between morning vs. evening-oriented adolescents [31], and Schubert and Randler found that morningness was negatively associated with BMI in a cohort of German university students [32] |
| Energy Balance |
Portion size was estimated based on the energy expenditure at the same hour and the average of the 3 hours preceding the meal |
Exercise related to increase in carbohydrate and protein intake [33, 34] |
| Emotional |
Portion size was estimated based on the two principal components derived from the EMA questions |
Negative mood states may be associated with eating unhealthy foods [35] |
| Food Environment |
Portion size was estimated based on the average number of food establishments encountered at each person’s Staypoints |
Neighborhood food environment associated with unhealthy eating [19–21] |
