Rational-Critical User Discussions: How Argument Strength and the Conditions Set by News Organizations Are Linked to (Reasoned) Disagreement

Figures & data

Table 1. Discussion norms and technical features by platform.

	Type A			Type B			Type C
	zeit.de	welt.de	sueddeutsche.de	tagesspiegel.de	huffingtonpost.de	taz.de	rp-online.de	focus.de	spiegel.de
Discussion Norms
Visible moderation	X	X	X
Policy: No unverifiable claims	X	X	X				X
Policy: Give arguments	X	X
Policy: Provide sources / quotations	X	X	X	X	X	X	X
Technical Features
Reply function	X	X	X	X	X	X		X	X
Thread visualization	X	X	X	X	X	X
N° of articles	22	14	11	25	16	15	25	34	20
N° of comments	4586	1470	739	797	503	268	238	3627	2462
Print outlet	National, weekly	National, daily	National, daily	Regional, daily	–	National, daily	Regional, daily	National, weekly	National, weekly
Online reach^a	60	98	54	13	20	5	21	165	227
Editorial line^b	Liberal	Conservative	Left-liberal	Liberal	Left-liberal	Left	Conservative	Conservative	Left-liberal

Note. ^aAverage online reach in million visits (online + mobile) over the entire period under investigation (November 2015–September 2017). Based on data provided by IVW (2017). ^bAccording to euro topics (2021).

Table 2. Descriptive statistics of independent and dependent variables.

Number of comments	Abs.	Rel.
with an unjustified claim	6449	45%
with a non-verifiable justification	3671	25%
with a verifiable justification	3993	28%
that were replied to	4795	33%
that were disagreed to	4343	30%
that were disagreed to with an argument	2988	21%

Table 3. Regression models for explaining likelihood of receiving (justified) disagreement.

	DV: Disagreement received (yes / no)			DV: Justified disagreement received (yes / no)
	Exp(B)	SE	95% Conf. (Exp(B))	Exp(B)	SE	95% Conf. (Exp(B))
Main effects
Non-verifiable justification^a^,^b	1.44***	.08	[1.23; 1.69]	1.54***	.09	[1.28; 1.85]
Verifiable justification^a^,^b	1.83***	.08	[1.57; 2.14]	2.02***	.09	[1.70; 2.41]
Medium type A	2.47***	.07	[2.16; 2.83]	1.75***	.08	[1.49; 2.05]
Medium type B	2.07***	.10	[1.71; 2.51]	1.36**	.12	[1.08; 1.71]
Non-verifiable justification * Medium type A	.96	.11	[.78; 1.18]	1.14	.12	[.90; 1.43]
Verifiable justification * Medium type A	.88	.10	[.72; 1.07]	1.09	.11	[.88; 1.36]
Non-verifiable justification * Medium type B	.98	.11	[.71; 1.33]	1.20	.18	[.84; 1.70]
Verifiable justification * Medium type B	1.00	.15	[.74; 1.35]	1.35	.17	[.97; 1.89]
Control variables
Issue (1 = pension reform)	1.34***	.04	[1.24; 1.46]	1.43***	.05	[1.31; 1.57]
Position (1 = reply comment)	2.75***	.04	[2.54; 2.98]	2.66***	.05	[2.43; 2.92]
Missing respect^b	1.05	.06	[.94; 1.18]	1.08	.07	[.95; 1.23]
Negative emotion^b	.86*	.06	[.77; .97]	.89	.07	[.78; 1.02]
Constant term	.07***	.09		.04***	.11
Model Information	−2 Log-Likelihood = 16116.68; Chi^2 = 1563.70***; R^2 (Nagelkerke) = .15			−2 Log-Likelihood = 13627.83; Chi^2 = 1109.73***; R^2 (Nagelkerke) = .12

Note. Logistic Regression Models; N = 14468; ^aReference category: Comment without justification; ^b0 = not present, 1 = present. ^***p < .001; ^**p < .01; ^*p < .05.