Political persuasion on social media: A moderated moderation model of political discussion disagreement and civil reasoning

Figures & data

Figure 1. Conceptual model: three-way interaction effect of exposure to political discussion disagreement (M) and civil reasoning (W) on the relationship between social media news use (X) and political persuasion on social media (Y).

Table 1. Zero-order correlations among all independent and dependent variables in the analysis.

Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. Age	—
2. Gender (female)	–0.07*	—
3. Education	–0.08*	–0.05	—
4. Income	0.02	–0.10**	0.29***	—
5. Race (white)	0.11***	–0.02	–0.09**	–0.07*	—
6. Pol. discussion network size	–0.11**	–0.14***	0.20***	0.15***	–0.01	—
7. Pol. discussion frequency	–0.15***	–0.04	0.19***	0.21***	–0.03	0.53***	—
8. Frequency of social media use	–0.41***	0.05	0.07*	0.01	–0.05	0.24***	0.29***	—
9. Non-social media news use	–0.16***	–0.09**	0.23***	0.24***	–0.10**	0.34***	0.55***	0.36***	—
10. Strength of ideology	0.17***	–0.06	–0.03	0.11***	0.05	–0.03	0.02	–0.12***	0.04	—
11. Political efficacy	–0.04	–0.20***	0.21***	0.19**	–0.01	0.34***	0.44***	0.18***	0.43***	0.05	—
12. Political knowledge	0.22***	–0.26***	0.17***	0.15***	0.06*	0.24***	0.21***	–0.12***	0.16***	0.11**	0.26***	—
13. Political interest	0.04	–0.20***	0.18***	0.18***	–0.01	0.39***	0.54***	0.11***	0.45***	0.12**	0.62***	0.46***	—
14. Social media news use	–0.43***	0.02	0.09**	0.04	–0.07*	0.31***	0.41***	0.79***	0.54***	–0.11***	0.28***	–0.07*	0.20***	—
15. Civil reasoning	–0.17***	–0.12***	0.23***	0.15***	–0.04	0.54***	0.71***	0.26***	0.47***	0.04	0.45***	0.26***	0.51***	0.36***	—
16. Political discussion disagreement	–0.29***	–0.11***	0.20***	0.11**	–0.05	0.53***	0.70***	0.53***	0.53***	–0.03	0.44***	0.10**	0.44***	0.60***	0.81***	—
17. Social media political persuasion	–0.39***	–0.04	0.08**	0.03	–0.12**	0.36***	0.43***	0.59***	0.53***	–0.13***	0.34***	–0.11**	0.21***	0.72***	0.38***	0.58***

Note. Cell entries are zero-order Pearson correlation coefficients, with the exception of gender and race (Phi) (Total N = 1077).

*p < .05, **p < .01, ***p < .001.

Table 2. OLS regression model testing relationships among social media news use, discussion political disagreement, and civil reasoning on social media political persuasion.

	Social media political persuasion
Block 1: Demographics
Age	–0.083***
Gender (1 = Female)	–0.041*
Education	–0.038
Income	–0.046*
Race (1 = White)	–0.058**
ΔR^2	17%
Block 2: News and Discussion Network Attributes
Network size	0.097***
Political discussion frequency	0.038
Frequency of social media use	0.050
Non-social media news use	0.156***
ΔR^2	35%
Block 3: Political Orientations
Strength of ideology	–0.025
Political efficacy	0.137***
Political knowledge	–0.111***
Political interest	–0.059*
ΔR^2	3%
Block 4: Independent Variables
Social media news use	0.417***
Civil reasoning	–0.038
Discussion political disagreement	0.138**
ΔR^2	6%
Total R^2	60%

Note. Cell entries are final-entry OLS standardized beta (β) coefficients. N = 1077. *p < .05; **p < .01; ***p < .001.

Table 3. OLS moderated moderation model showing three-way interactions among social media news use, political discussion disagreement, and civil reasoning political persuasion on social media.

	B	SE	95% CI
SM news use × political discussion disagreement	0.047	0.014	0.018	0.075*
SM news use × civil reasoning	0.036	0.011	0.013	0.058*
Political discussion disagreement × civil reasoning	–0.016	0.010	–0.035	0.004
SM news use × political discussion disagreement × civil reasoning	–0.008	0.003	–0.013	–0.002*

Note. SM = Social media. Analyses performed using the PROCESS macro for SPSS (Model 3; Hayes 2013). *Denotes statistically significant relationships where p < .05. Model includes the same demographic and political orientation control variables as found in Table 2.

Figure 2. Three-way interaction plot of social media news use, political discussion disagreement, and civil reasoning on political persuasion on social media.

Table A1. Demographic profile of study survey and other comparable surveys.

	Study UK Survey March 2014 (%)	UK Census 2011 (%)
Age:
20–24	5.8	8.7
25–34	17.6	17.7
35–44	21.1	16.9
45–64	53.3	33.4
65 or more	2.2	23.1
Gender
Male	44.1	48.6
Female	55.9	51.4
Race/Ethnicity:
White	91.2	87.6
Asian/Asian British	3.6	6.9
Black/African	2.7	2.9
Mixed groups	—	2.2
Other	2.5	1.0
Qualifications:
Secondary school or below	46.1	29.3
Some college	24.5	20.5
College degree or above	19.2	27
Technical or other	7.2
Household Income:
<$49,999	85.4	—
$50,000–$99,999	11.7	—
$100,000 or more	2.9	—

Note. Population statistics are based on 2014 ONS estimates. Since education levels were not available, the 2011 ONS Census was used instead.