Effects of Personal Characteristics on Temporal Response Patterns in Ecological Momentary Assessments

Response rates to Ecological Momentary Assessment (EMA) questions vary among individuals, which inevitably results in a disproportionate distribution of the volume of EMA data to personal characteristics. Previous research has explored how response rates are influenced by subjects’ characteristics; however, few studies focused on temporal patterns of responses to EMA questions. In this study, we examined the relationship of personal characteristics with temporal response patterns by using StudentLife dataset. From this dataset, we adopted scales of personality traits (neuroticism, extraversion, openness, agreeableness, conscientiousness) and mental status (depression, stress, loneliness, life satisfaction) as subjects’ personal characteristics, and analyzed how these characteristics relate to the subjects’ temporal response patterns for EMA questions. Primary results of our analyses using regression models and latent growth curve models indicated that the subjects who reported high stress levels had a higher total number of EMA responses, but the number of responses from these subjects tended to decrease strongly throughout the experiment, especially after a stressful event such as the midterm examination. We also found that the subjects who reported low extraversion had a weak tendency to decrease the number of EMA responses throughout the experiment, but the number of EMA responses of these subjects decreased strongly after a social event such as the school festival. Based on our results, we discussed how the data collected with EMA studies are biased not only in terms of personal characteristics but also with timings.