Skip to main content
. 2020 Jan 17;6(1):e03214. doi: 10.1016/j.heliyon.2020.e03214

Table 3.

Multinomial logistic regression to predict group membership.

News feeds and online (n = 46)
(versus “news feeds only” (n = 49))
Offline only (n = 272)
(versus “news feeds only” (n = 49))
Coefficient (SE) p Coefficient (SE) p
Intercept 0.03 (.287) .930 1.89 (.223) <.001
Age 0.05 (.262) .856 0.82 (.202) <.001
Gender -0.02 (.261) .945 0.11 (.216) .602
Extraversion -0.16 (.231) .493 -0.23 (.184) .202
Agreeableness -0.34 (.228) .132 -0.11 (.185) .551
Conscientiousness 0.22 (.234) .347 0.51 (.189) .007
Neuroticism -0.17 (.244) .495 -0.40 (.194) .039
Openness 0.22 (.211) .299 0.08 (.165) .626
B-RWA-6 -0.27 (.235) .246 0.10 (.188) .592

Note. The “news feeds only” group was coded “1” (reference group), the “news feeds and online” group was coded “2” and the “offline only” group was coded “3”. The groups for these analyses contained n = 49, n = 46, and n = 272 participants (“news feeds only”, “news feeds and online”, “offline only”). The predictors (except gender; 0 = male, 1 = female) were z-standardized (in the complete sample) before including them in the model; hence, in N = 1,681 for age and the Big Five and in n = 1,397 for the B-RWA-6 (balanced short scale on authoritarian attitudes). The "news feeds only" group is not presented in the table as it is the reference group.