Automated Serendipity

Figures & data

TABLE 1 Descriptive statistics for the independent and control variables

	UK	USA	Germany	Spain
Search engines for news (%)	12.00	25.17	18.31	25.92
Age	47.09 (16.13)	46.59 (17.42)	47.97 (15.70)	46.42 (14.17)
Gender: male (%)	48.10	49.56	48.52	48.90
Education	3.84 (1.19)	3.59 (1.32)	3.84 (1.09)	4.06 (1.17)
Political leaning	−.05 (.44)	−.01 (.43)	−.04 (.39)	−.15 (.42)
Trust in news	3.15 (.97)	2.90 (1.25)	3.33 (.99)	3.29 (1.05)
Interest in news	2.87 (.83)	2.94 (.94)	2.93 (.87)	3.08 (.71)
Internet frequency	6.97 (1.07)	6.96 (1.26)	6.72 (1.20)	7.01 (1.06)
Aggregator for news (%)	6.53	8.82	7.90	9.31
Social media for news (%)	41.71	30.28	22.06	40.70

Note. Rows showing mean following by standard deviation in parentheses unless indicated.

Figure 1 Mean number of online news sources used in the previous week

Figure 2 Proportion that used both a left-leaning and right-leaning online news source in the last week

Figure 3 Mean ratio between left-leaning and right-leaning news sources used in the last week

TABLE 2 Poisson regression models where the dependent variable is number of online news sources used in the last week

	UK	USA	Germany	Spain
Search engine for news	.36***	.30***	.28***	.23***
	(.08)	(.05)	(.07)	(.05)
Age	−.17*	−.12*	−.07	.02
	(.08)	(.06)	(.08)	(.06)
Gender (female)	.004	.03	−.05	−.23***
	(.08)	(.05)	(.07)	(.05)
Education	.18*	.16**	.15*	.15**
	(.07)	(.05)	(.07)	(.05)
Political leaning	−.02	−.32***	.03	−.002
	(.08)	(.06)	(.07)	(.05)
Trust in news	.02	.01	−.12	.11*
	(.08)	(.05)	(.07)	(.05)
Interest in news	.40***	.34***	.69***	.38***
	(.08)	(.06)	(.08)	(.06)
Internet frequency	.26**	.09	.15	.05
	(.10)	(.06)	(.08)	(.06)
Aggregator for news	.50***	.45***	.47***	.27***
	(.09)	(.06)	(.10)	(.07)
Social media for news	.31***	.36***	.36***	.32***
	(.08)	(.05)	(.08)	(.05)
Constant	1.09***	1.48***	1.05***	1.54***
	(.04)	(.03)	(.04)	(.03)
Pseudo R ^2	.30	.28	.24	.20
N	426	937	624	826

Note. *p < .05, ** p < .01, *** p < .001. Columns showing standardized β, with standard error in parentheses.

TABLE 3 Binary logistic regression models where the dependent variable is the proportion of people who used both a left-leaning and a right-leaning online news source in the last week

	UK	USA	Germany	Spain
Search engine for news	2.44***	1.92***	2.22***	1.57**
	(.11)	(.17)	(.18)	(.15
Age	.92	.86	.81	.59**
	(.24)	(.18)	(.19)	(.16)
Gender (female)	1.17	.86	.96	.97
	(.22)	(.16)	(.18)	(.15)
Education	1.30	1.26	1.09	1.21
	(.22)	(.17)	(.18)	(.15)
Political leaning	2.54***	3.12***	.88	.67**
	(.22)	(.18)	(.18)	(.15)
Trust in news	.99	.74	1.05	1.27
	(.22)	(.17)	(.18)	(.16)
Interest in news	1.45	1.37	2.23***	1.93***
	(.23)	(.17)	(.19)	(.16)
Internet frequency	1.63*	.94	1.25	1.01
	(.25)	(.17)	(.19)	(.16)
Aggregator for news	1.96*	2.02**	1.69	1.24
	(.31)	(.22)	(.30)	(.24)
Social media for news	1.11	1.43*	1.77**	1.36*
	(.24)	(.17)	(.19)	(.15)
Constant	.57***	.29***	.65***	1.46***
	(.11)	(.09)	(.09)	(.07)
Pseudo R ^2	.10	.09	.08	.06
N	426	937	624	826

Note. *p < .05, **p < .01, ***p < .001. Columns showing standardized Exp(β), with standard error in parentheses.

TABLE 4 Poisson regression models where the dependent variable is the ratio of left-leaning and right-leaning online news sources used in the last week

	UK	USA	Germany	Spain
Search engine for news	.39*	.22	.46***	.16*
	(.15)	(.15)	(.12)	(.07)
Age	.27	.05	−.17	−.27***
	(.16)	(.16)	(.12)	(.08)
Gender: female	.07	−.12	.05	.11
	(.15)	(.15)	(.11)	(.07)
Education	.12	.08	.04	.04
	(.15)	(.16)	(.11)	(.07)
Political leaning	.72***	1.21***	−.14	−.24**
	(.14)	(.18)	(.12)	(.08)
Trust in news	−.09	−.42**	−.01	.10
	(.15)	(.16)	(.11)	(.07)
Interest in news	.08	.14	.37**	.20*
	(.15)	(.15)	(.12)	(.08)
Internet frequency	.30	.02	.08	−.01
	(.17)	(.15)	(.13)	(.08)
Aggregator for news	.33	.26	.13	.14
	(.19)	(.22)	(.17)	(.10)
Social media for news	−.07	.02	.28*	.02
	(.16)	(.15)	(.12)	(.07)
Constant	−1.63***	−2.43***	−1.40***	−1.05***
	(.08)	(.09)	(.06)	(.04)
Pseudo R ^2	.12	.15	.06	.04
N	426	937	624	826

Note. *p < .05, **p < .01, ***p < .001. Columns showing standardized β, with standard error in parentheses.