Table 2.
Selected predictors in the final step of stepwise regressions with forward predictor selection for each well-being outcome.
| Positive Affect |
Negative Affect |
Stress |
Social Connectedness |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| R2 |
.30 |
.35 |
.41 |
.44 |
||||||||
| b (se) | r | rpartial | b (se) | r | rpartial | b (se) | r | rpartial | b (se) | r | rpartial | |
| 1. Time Interacting in Person | ||||||||||||
| w/Strong Ties | .04 (.01) | .21 | .14 | -.02 (.01) | -.04 | -.10 | .04 (.01) | .11 | .16 | |||
| 2. Time Interacting Online | ||||||||||||
| w/Strong Ties | .05 (.02) | -.04 | .15 | |||||||||
| w/Weak Ties | .07 (.02) | .20 | .20 | .05 (.01) | .21 | .16 | -.08 (.02) | -.18 | -.12 | |||
| 3. Interaction partners | ||||||||||||
| No one | -.31 (.13) | -.12 | -.11 | |||||||||
| Roommate | -.28 (.14) | -.04 | -.08 | |||||||||
| Family | .17 (.06) | .09 | .13 | |||||||||
| Romantic Partner | .16 (.06) | .18 | .12 | |||||||||
| Customers/Clients | .51 (.15) | .10 | .15 | |||||||||
| Coworkers/Classmates | .25 (.09) | .15 | .11 | |||||||||
| 4. Other Activities | ||||||||||||
| Exercise | .20 (.07) | .23 | .12 | -.19 (.07) | -.15 | -.12 | ||||||
| Existing Hobbies | -.12 (.06) | -.02 | -.09 | |||||||||
| Making New Friends | -.40 (.13) | -.01 | -.13 | |||||||||
| Watching Shows/Movies | .20 (.08) | .00 | .10 | |||||||||
| 5. Other Online Activities | ||||||||||||
| Meetings with Clubs | .21 (.09) | .18 | .10 | |||||||||
| Games with Strong Ties | .18 (.08) | .15 | .09 | |||||||||
| # Phone Notifications | .00 (.00) | .16 | .15 | |||||||||
| 6. Living | ||||||||||||
| w/Children | .14 (.06) | .09 | .10 | |||||||||
| 7. Personality | ||||||||||||
| Emotional Stability | .25 (.03) | .42 | .38 | -.34 (.03) | -.53 | -.49 | -.31 (.02) | -.59 | -.55 | .27 (.02) | .55 | .45 |
| Extraversion | .08 (.02) | .29 | .14 | .11 (.02) | .34 | .23 | ||||||
| 8. Demographics | ||||||||||||
| Age | -.01 (.00) | -.28 | -.10 | -.01 (.00) | -.36 | -.19 | .01 (.00) | .31 | .17 | |||
| Currently Employed | -.15 (.06) | -.04 | -.10 | |||||||||
| Can Work Remotely | -.16 (.08) | .06 | -.09 | |||||||||
Note. The final step of the stepwise regressions shown above includes only the predictors that met the criterion for entry (p < .05) at each of the preceding steps. See Tables S8–S11 for details on predictor selection step-by-step. Empty cells indicate that the predictor was not selected in the final step for a given outcome. The list of variables initially provided for each stepwise regression with forward selection included time interacting with strong and weak ties online and in-person, as well as a wide range of related variables and established predictors of well-being as described next. The weekly predictors entered into the stepwise regression are as follows: (1) time interacting in person (with strong ties and weak ties); (2) time interacting online (with strong ties and weak ties); (3) interaction partners: people with whom participants interacted during the past week, each coded as 0–No, 1–Yes (no one, friends, dating partner/potential dating partner, spouse/partner/significant other, family/relatives, one's children, roommate/housemate/living partner, coworkers/classmates, boss/supervisor, clients/customers/students/patients, acquaintances, strangers); (4) other activities over the past week, each coded as 0–No, 1–Yes (existing hobbies, new hobbies, cooking/baking, exercising, leisure activities, reading for pleasure, watching shows/movies, going for walks outside, catching up with friends/family, making new friends); (5) other online activities and behaviors over the past week, each coded as 0–No, 1–Yes (virtually co-watching shows/movies, group video chats with friends/family, remote dinners/drinking, online meetings with clubs, playing online games simultaneously with family/friends, virtual workout classes, catching up online with friends), along with % social interactions online vs. in person (0–100%), % active vs. passive social media use (coded from 1 to 11, where 1 = 100% passive, 6 = equal parts passive and active, and 11 = 100% active), and smartphone screen time (0–12 h), pickups (0–400), and notifications (0–400). We also entered a range of stable and dispositional predictors, including: (6) living arrangements, each coded as 0–No, 1–Yes (living alone, living with one's partner, one's children, parent(s), sibling(s), friend(s), grandparent(s), aunt(s)/uncle(s), cousin(s), niece(s)/nephew(s), roommate(s), and caring for children); (7) personality traits (extraversion, agreeableness, conscientiousness, emotional stability, and openness to experience, measured on a scale from 1 to 7 (TIPI;Gosling et al., 2003), and (8) demographics (age, gender, income, education, and race). Finally, the following variables were also entered but are not shown in the final model, as they did not predict any of the outcomes: (a) COVID-19 diagnoses of self and others, coded as 0–No, 1–Yes; (b) employment arrangements, each coded as 0–No, 1–Yes (currently employed, commuting to work, ability to work remotely); (c) places where participants spent time in the past week, each coded as 0–No, 1–Yes (at home, at work/school, on public transit, in a car, at a restaurant/café, at a bar/party, at the gym, out in public, out in nature). See Supplementary Online Material for measure details, descriptive statistics, and correlations.