(PS5-B49) Trait-based Difficulty in Emotional Regulation Is Negatively Associated with Eating and Sleeping Well for First-year College Students During COVID-19 Lockdown

Background: The COVID-19 pandemic exacerbated the challenges faced by emerging adult first-year college students. Dealing with the uncertainty of early lockdown (April 5 – May 20, 2020) required the frequent use of emotional regulation (ER) skills such as ABC PLEASE skills, which are a set of DBT tools used to reduce emotional vulnerability. In this study, we investigated whether first-year college students’ (Spring 2020) baseline emotion regulation tendencies would predict the likelihood that they would engage in any of the ER skills on a momentary basis. 

Methods:  

Sample: 76 first-year college students (Mage = 18.61, SD = 0.61) from small, liberal arts colleges in southern California took part in a week-long experiential study during the first mandated COVID-19 lockdowns in California. Students represented fairly diverse genders (71% female, 24% male, 4% non-binary) and ethnic (46% White, 36% Asian or Pacific Islander, 14% Hispanic or Latinx, 13% Black or African-American, 14% mixed or other) backgrounds. 

Procedure: Ecological Momentary Assessments (EMA) were administered at semi-random time points within specific three-to-four-hour time blocks (morning, afternoon, late afternoon, and evening) for seven consecutive days.

Within-person measures: We assessed whether participants at the moment have recently (since the last signal) engaged in behaviors or used (0 = no behavior, 1 = behavior). These ER skills have been demonstrated to help cope and regulate emotions, ten items were created based on “Reducing Vulnerability to Emotion Mind” from the DBT Skills Training Manual. 

Between-person measures: We used the 18-item Difficulties in Emotion Regulation Scale (DERS-18; Victor & Klonsky, 2016) to measure dispositional levels of emotion regulation and to use as a time-invariant covariate (TIC) in our models.

Data Analysis: A multilevel binomial modeling approach was used due to nested data and the binary outcome variable of ER skills. N = 1,796 momentary assessments were analyzed in R using packages for linear mixed effects (e.g., lme4).

Results:  
Generalized linear models (using a binomial distribution) showed that higher scores on the DERS-18 predicted significantly lower logit scores (i.e., likelihood) for only the balanced Eating (E) and balanced Sleeping (S) skills from ABC PLEASE. In other words, participants who had trait-based difficulties with emotion regulation were more likely to experience sleeping difficulties and have unhealthy eating habits. However, no other ABC PLEASE skills (accumulating long and short-term positives, building mastery, coping ahead, taking care of your body, avoiding substances, and exercising) were significantly predicted by dispositional emotion regulation difficulties.