How Political Efficacy Relates to Online and Offline Political Participation: A Multilevel Meta-analysis

Figures & data

Table 1. Sample characteristics.

Characteristics of studies and effects	Number of Studies	Number of Effects
Type of sample
Online panel of individuals adjusted to match census data	27	89
University students	10	36
Random sample RDD or cellphone	7	46
Other type of convenience sample (e.g., twitter users survey)	4	13
Total	48	184
Country
Outside the United States	27	93
United States	22	91
Total	49	184
Sample size
Less than 249 respondents	1	2
250 to 499 respondents	18	71
500 to 749 respondents	8	26
750 to 999 respondents	5	15
1000 to 1249 respondents	8	27
1250 to 1499 respondents	4	13
More than 1500 respondents	9	30
Total	53	184

Note. For country, the total number of studies documented in Table 1 (n = 49) exceeds the total number of studies in the meta-analytic dataset (n = 48) because one study (Saldaña et al., 2015) analyzed two separate samples: one of the United States, and one outside of the US (the United Kingdom). For sample size, the total number of studies by sample size (n = 53) exceed the total 48 studies in the meta-analytic dataset because several studies include multiple models based on different sample sizes (AlSalem, 2015; Guidetti et al., 2016; Jugert et al., 2013; Saldaña et al., 2015).

Figure 1a. OnPP forest plot, random effects meta-analysis (H1: efficacy hypothesis).

Figure 1b. OffPP forest plot, random effects meta-analysis (H1: efficacy hypothesis).

Table 2. Test for subgroup differences in PE-OnPP versus PE-OffPP (H1, Efficacy hypothesis).

a. Common effect estimate dataset (k = 183)
Category	S	k	b	CI	F	p-value	% var. level 1	% var. level 2	% var. level 3
Offline PP (intercept)	41	183	0.154	[0.121, 0.187]	3.400 (1,181)	0.000	6.029	53.612	40.359
Online PP			−0.019	[−0.039, 0.001]		0.067
b. Zero-order correlation dataset (k = 45)
Category	S	k	b	CI	F	p-value	% var. level 1	% var. level 2	% var. level 3
Offline PP (intercept)	14	45	0.38	[0.301, 0.460]	0.769 (1,43)	0.000	5.766	62.874	31.360
Online PP			−0.027	[−0.089, 0.035]		0.385

Note. Models estimated as multilevel random effects meta-analysis; % var. at level 1 is due to sampling error; at level 2 is due to within-cluster variance (i.e., variance within the same dataset); and at level 3 is between-cluster variance (i.e., between distinctive datasets in the meta-analytic sample). While the variance distribution results from within-cluster variance (k = 183, 53.6%; k = 45, 62.9%) and between-cluster variance (k = 183, 40.4%; k = 45, 31.4%), the test of subgroup differences for both datasets shows no statistically significant difference in the associations between political efficacy and online versus offline participation (for k = 183, p = 0.067; for k = 45, p = 0.385).

Table 3. Test of the effect of year as moderator (H2, Over-time convergence).

a. Common effect estimate dataset (k = 183)
Category	S	k	b	CI	F	p-value
Intercept (OffPP)	41	183	−11.328	[−30.415, 7.758]	1.849 (1,179)	0.243
Online PP			8.584	[−5.613, 22.780]		0.234
Year			0.006	[−0.004, 0.015]		0.237
Online PP * year			−0.004	[−0.011, 0.003]		0.233
b. Zero-order correlation dataset (k = 45)
Category	S	k	b	CI	F	p-value
Intercept (OffPP)	14	45	39.413	[−13.809, 92.636]	1.097 (1,41)	0.142
Online PP			−34.324	[−95.130, 26.482]		0.261
Year			−0.019	[−0.046, 0.007]		0.146
Online PP * year			0.017	[−0.013, 0.047]		0.261

Note. Models estimated as multilevel random effects meta-analysis

Table 4. Test of democratic contextual features as moderators (H3, Democratic context).

a. Common effect estimate dataset (k = 183)
Moderator	Participation	S	k	b	CI	F	p-value
Polyarchy	Online PP	39	78	−0.009	[−0.107, 0.089]	0.035 (1,83)	0.853
	Offline PP	39	87	−0.111	[−0.229, 0.007]	3.500 (1,92)	0.065
Freedom of association	Online PP	39	78	0.001	[−0.088, 0.090]	0.001 (1,83)	0.974
	Offline PP	39	87	−0.098	[−0.205, 0.010]	3.274 (1,92)	0.074
Clean elections	Online PP	39	78	−0.003	[−0.089, 0.083]	0.005 (1,83)	0.946
	Offline PP	39	87	−0.111	[−0.214, −0.007]	4.515 (1,92)	0.036
Freedom of expression	Online PP	39	78	−0.023	[−0.120, 0.074]	0.230 (1,83)	0.633
	Offline PP	39	87	−0.130	[−0.243, −0.017]	5.181 (1,92)	0.025
Elected officials	Online PP	39	78	−0.007	[−0.083, 0.070]	0.031 (1,83)	0.860
	Offline PP	39	87	−0.092	[−0.184, 0.000]	3.936 (1,92)	0.050
Civil society strength	Online PP	39	78	−0.032	[−0.152, 0.089]	0.271 (1,83)	0.604
	Offline PP	39	87	−0.156	[−0.296, −0.015]	4.822 (1,92)	0.031
Internet penetration	Online PP	38	77	0.000	[−0.002, 0.002]	0.072 (1,82)	0.789
	Offline PP	38	86	−0.002	[−0.004, 0.001]	1.525 (1,91)	0.220
Globalization (trade)	Online PP	38	77	0.000	[0.000, 0.001]	1.622 (1,82)	0.206
	Offline PP	38	86	0.000	[0.000, 0.001]	0.035 (1,91)	0.853
b. Zero-order correlation dataset (k = 45)
Moderator	Participation	S	k	b	CI	F	p-value
Polyarchy	Online PP	14	20	0.039	[−0.170, 0.247]	0.151 (1,19)	0.702
	Offline PP	14	23	−0.035	[−0.311, 0.241]	0.070 (1,22)	0.794
Freedom of association	Online PP	14	20	0.026	[−0.164, 0.216]	0.083 (1,19)	0.777
	Offline PP	14	23	−0.057	[−0.311, 0.197]	0.215 (1,22)	0.647
Clean elections	Online PP	14	20	0.029	[−0.144, 0.202]	0.123 (1,19)	0.730
	Offline PP	14	23	−0.043	[−0.274, 0.188]	0.147 (1,22)	0.705
Freedom of expression	Online PP	14	20	0.033	[−0.165, 0.231]	0.121 (1,19)	0.732
	Offline PP	14	23	−0.030	[−0.294, 0.235]	0.054 (1,22)	0.819
Elected officials	Online PP	14	20	0.023	[−0.138, 0.183]	0.087 (1,19)	0.771
	Offline PP	14	23	−0.053	[−0.268, 0.162]	0.259 (1,22)	0.616
Civil society strength	Online PP	14	20	0.041	[−0.206, 0.289]	0.123 (1,19)	0.730
	Offline PP	14	23	−0.023	[−0.354, 0.307]	0.022 (1,22)	0.885
Internet penetration	Online PP	14	20	0.001	[−0.004, 0.006]	0.297 (1,19)	0.592
	Offline PP	14	23	0.000	[−0.007, 0.007]	0.004 (1,22)	0.948
Globalization (trade)	Online PP	14	20	0.000	[−0.001, 0.001]	0.005 (1,19)	0.944
	Offline PP	14	23	0.000	[−0.002, 0.001]	0.251 (1,22)	0.621

Note. Models estimated as multilevel random effects meta-analysis

Saldaña, M., McGregor, S. C., & Gil de Zúñiga, H. (2015). Social media as a public space for politics. International Journal of Communication, 9. http://ijoc.org/index.php/ijoc/article/view/3238

 Google Scholar

AlSalem, F. (2015). Digital media and women’s political participation in Kuwait [Unpublished doctoral dissertation]. Indiana University.

 Google Scholar

Guidetti, M., Cavazza, N., & Graziani, A. R. (2016). Perceived disagreement and heterogeneity in social networks: Distinct effects on political participation. The Journal of Social Psychology, 156(2), 222–242. http://doi.org/10.1080/00224545.2015.1095707

 PubMed Web of Science ®Google Scholar

Jugert, P., Eckstein, K., Noack, P., Kuhn, A., & Benbow, A. (2013). Offline and online civic engagement among adolescents and young adults from three ethnic groups. Journal of Youth and Adolescence, 42(1), 123–135. http://doi.org/10.1007/s10964-012-9805-4

 PubMed Web of Science ®Google Scholar

Data Availability Statement

The data described in this article are openly available in the Open Science Framework at http://doi.org/10.17605/OSF.IO/AF5DR.