Sharing political news: the balancing act of intimacy and socialization in selective exposure

EPJ Data Science

Table 4 Logistic regression results for retweeting news

Predictive variables	Original	Revised
Gratification
F1 numexposures	0.93^∗∗∗	0.93^∗∗∗
F2 topic-interesting-me	−0.09	-
Selective exposure
F3 political	−0.80^∗∗∗	−0.89^∗∗∗
F4 leaning-matched-me	0.72^∗∗∗	0.91^∗∗∗
Socialization
F5 topic-interesting-followers	0.57^∗∗∗	0.49^∗∗∗
F6 leaning-matched-followers	−0.14	-
Trust and intimacy
F7 fromfriend	−2.09^∗∗∗	−1.47^∗∗∗
F8 mutualfriend	0.13^∗	-
F9 difference-in-followers	0.05^∗∗∗	-
F10 sharedfollowers	−0.00	-
F11 sharedfollowees	−0.00^∗∗∗	-
F12 sharedleaning	−0.37^∗∗∗	−0.38^∗∗∗
F13 politicsfromfriend	0.31^∗∗∗	0.35^∗∗∗

Both models pass the goodness-of-fit tests: Original ( $χ^{2} = 11.1629$ , p = 0.45) and Revised ( $χ^{2} = 13.2977$ , p = 0.39). Signif. codes: 0 (^∗∗∗) 0.001 (^∗∗) 0.01 (^∗) 0.05 (.).