When most people stayed home, who did they spend time with, and how did that relate to how they felt? The team surveyed adults on a single day in April 2020.
Tibbetts, M., Epstein-Shuman, A., Leitao, M. R., & Kushlev, K. (2021). A week during COVID-19: Online social interactions are associated with greater connection and more stress. Computers in Human Behavior Reports, 4, 100133. doi:10.1016/j.chbr.2021.100133
On April 8, 2020, at the peak of social distancing, 569 adults told the research team how much time they spent with strong ties (close people) and weak ties (acquaintances), in person and online, and how they felt.
More hours with strong ties in person went with more positive affect (r = .19) and more connectedness (r = .10).
r = .19
hours with close people in person and positive affect
A small positive link.
More hours online with weak ties went with more negative affect (r = .18) and more stress (r = .20). It also went with lower connectedness (r = −.16).
These are links at one point in time. The data cannot show whether online time caused the feelings or the feelings drove the time.
r = .20
hours online with acquaintances and stress
The strongest link between time use and a feeling.
Stress and negative affect were closely tied (r = .79). Connectedness fell as stress rose (r = −.70).
Time use explained between 30% and 44% of the variation in the four feelings once the four kinds of contact were combined (R² from .30 to .44).
R² = .44
variance in connectedness explained by time use
The highest of the four models.
What to look for. Positive affect and connectedness sat above the middle of the scale. Stress sat low.
| Measure | Mean | SD |
|---|---|---|
| Positive affect | 3.33 | .92 |
| Negative affect | 2.75 | .95 |
| Stress | 1.80 | .82 |
| Connectedness | 3.25 | .86 |
From the paper's descriptive statistics.
Hours of social time and feelings (correlations)
What to look for. Strong ties in person link to better feelings. Weak ties online link to worse ones.
How the four feelings relate
What to look for. Stress and negative affect are close to one. Connectedness opposes them.
What to look for. Each model combines all four kinds of contact. Connectedness and stress are the best explained.
| Outcome | R² | Key predictors (partial r) |
|---|---|---|
| Positive affect | .30 | Strong ties in person (.14) |
| Negative affect | .35 | Weak ties online (.20) |
| Stress | .41 | Weak ties online (.16) |
| Connectedness | .44 | Strong ties in person (.16), strong ties online (.15), weak ties online (−.12) |
Final stepwise models, from the paper's Table 2.
Partial correlations of the key predictors
The kind of contact mattered. Time with close people in person went with better feelings, and online time with acquaintances went the other way.
Limits. This is one day of survey data from online panels, so the data cannot show what caused what. The links are small to moderate.
Every value is from the printed paper. I did not recompute from raw data.