Skip to main content
. 2020 Jan 16;76(3):551–562. doi: 10.1093/geronb/gbaa007

Table 2.

Multilevel Linear Models Predicting Pleasantness of Encounters Throughout the Day From Encounters With Different Social Partners

Variable B SE
Fixed effects
 Intercept 4.69*** 0.03
 Within-person effectsa (encounter-level)
  Friend (Ref.) (Ref.)
  Romantic partner −0.05** 0.02
  Family member −0.06*** 0.02
  Other social partner −0.25*** 0.02
 Between-person effectsb
  Friend (Ref.) (Ref.)
  Romantic partner −0.01 0.18
  Family −0.11 0.17
  Other social partner −0.25 0.20
 Covariates
  Agec 0.00 0.00
  Genderc −0.04 0.06
  Educationc −0.02 0.02
  Healthc 0.10*** 0.03
  Minority statusc 0.06 0.06
  % of friends in social convoyc 0.04 0.13
  In-person contactd 0.09*** 0.01
Random effects
 Intercept VAR (Level 2: Assessment) 0.17*** 0.01
 Intercept VAR (Level 3: Participant) 0.15*** 0.01
 Residual VAR 0.32*** 0.00
 −2 log-likelihood 34276.3

Notes. Encounters n = 17,486 from 313 participants. Pleasantness of encounters was a continuous outcome 1 (unpleasant) to 5 (pleasant).

aThe predictor was a categorical variable representing different social partners which was recoded as four dummy variables.

bThe person mean variables representing the proportion of encounters across the study that each participant had encounters with different types of social partners.

cParticipant-level covariates.

dEncounter-level covariates.

*p < .05. **p < .01. ***p < .001.