The influence of various measures of health on different types of political participation

Abstract

Recent research in political behaviour suggests that poor health can be an impediment for individuals to vote. At the same time, researchers argue that health may both hinder and reinforce other forms of political participation. With respect to these ambiguous expectations, our study asks: does the relationship between health and political involvement depend on how we measure health? We answer this question for two of the most widely used health indicators, self-reported health and being hampered by illness in daily activities. We use the European Social Survey (ESS) (N = 35,000) covering 20 European countries and find that the measurement of health indeed matters: our results illustrate that bad self-reported health is an impediment to voting, but not to other forms of political activity. When it comes to our second indicator, being hampered in daily activities, we also find a negative relationship with voting. Yet, our results also indicate that most individuals, who are hampered by illness in their daily lives, have a tendency to participate more regularly in most other forms of political activity, including boycotting, contacting a politician, or signing a petition. Robustness checks including waves 1–6 of the ESS support these findings.

Keywords

Europe hampered by illness health political participation

Introduction

The question of why, when, and how individuals participate in the political process is probably one of the best-researched questions in Political Science. We know that an individual’s likelihood to actively engage in politics is determined by money, time, and civic skills (Brady et al., 1995); by socio-economic and socio-demographic factors (Smets and van Ham, 2013); by personality traits (Gallego and Oberski, 2012); as well as by feelings of deprivation and unequal treatment (Taylor et al., 1987). However, recently, a new factor, an individual’s health, has become prominent in the political participation literature (Gollust and Rahn, 2015; Wass et al., 2017). Several studies unambiguously underscore that poor health hampers an individual’s likelihood to vote (among others Mattila et al. 2013; Söderlund and Rapeli, 2015). Yet, the influence of poor health towards other forms of political action seems to be more complex. In this article, we aim at disentangling this relationship by going back to a more conceptual level, that is, the way we measure poor health. In detail, we ask the following research question: does the relationship between individual health and political involvement depend on how we assess poor health?

To answer this question, we distinguish two health indicators – self-reported health and being hampered by illness in daily activity – and test both indicators’ relationship with different forms of political engagement (Barnes, 1979; van Deth, 2009). In doing so, we combine insights from prior research on the effects of different health indicators (Burden et al., 2017; Mattila et al. 2018; Wass et al., 2017) as well as the numerous research measuring the influence of poor health on different forms of political participation (Söderlund and Rapeli, 2015). Using data from the European Social Survey (ESS) (2014; N = 35,000), we receive a nuanced picture about the relationship between different health indicators and different forms of political participation.

In sum, we can conclude that both forms of health put a drain on voting. In contrast, the effect of any of the two health measures on the six other types of political participation – contacting a politician, signing a petition, participating in a boycott, wearing a campaign badge, being a member of a political party, and contacting a politician – is more nuanced. For one, self-reported health does not have any relevant relationship with any of these types of political activities. In contrast, the second health indicator, being hampered by illness in daily activity, shows a positive, albeit small, relationship with most of these forms of political participation: individuals, who feel hampered by illness in daily activity, have a slightly higher likelihood to contact a politician, sign a petition, wear a campaign badge, and boycott. These findings hold if we test their robustness with others waves of the ESS (waves 1–6).

This article proceeds as follows: first, we briefly situate this study within the health and political participation literatures and explain our contribution; second, we explain our variables, data, and methods; third, we present our results; and finally, we conclude and provide some avenues for future research.

Theory and measurement of health

The literature on the relationship between poor health and various forms of political participation has made great advances in recent years. Theoretically, scholars have identified two rather contradicting mechanisms. The first relies on an amended form of the resource model of political participation (Brady et al., 1995). In its original form, the model considers time, money, and civic skills as essential communication and organizational capacities to engage in political action (Brady et al., 1995). Scholars in the field of health and political participation add health as an additional individual resource (Gollust and Rahn, 2015). If this resource, that is, being in good health, is not present or scarce, then individuals’ levels of participation in political activities should be lower. The assumption is that poor health can hamper political participation both directly and indirectly (Couture and Breux, 2017). Directly, ailing health might make it physically harder to go to the polls, turn out for a rally, or walk in a demonstration (Schur and Kruse, 2000).

More indirectly, dealing with illness (both mentally and physically) might consume a lot of time and energy, which makes political engagement harder or simply impossible (Mattila et al., 2013). In addition, individuals who suffer from severe health conditions might also be disadvantaged professionally, for example, they have higher rates of unemployment, a lower educational level, or, if they have a job, they might not climb the professional ladder as quickly as the more healthy ones. Therefore, they might not acquire the same amount of civic skills as healthy individuals, which, in turn, can lower their likelihood to participate politically. Finally, citizens with a physical or mental impairment are also likely to lose their political interest – a well-known indicator pushing political behaviour (Marx and Nguyen, 2016).

The second mechanism explaining the relationship between individual health and political involvement takes a rational choice perspective.¹ This approach focuses on the costs and benefits of participation and encourages the idea that even though poor health is accompanied by less resources, those suffering from poor health could be more likely to take part in various forms of political participation such as boycotting or signing a petition to make their voices heard (Taylor et al., 1987). According to Stryker et al. (2000), individuals with health issues have an increased self-interest, which makes them more likely to be activated through collective action. To overcome certain disadvantages – such as less favourable policies or unequal political representation – they might even engage in rather unconventional forms of collective action, such as demonstrating or boycotting. Partaking in these intense forms of political actions can ensure that the government provides the necessary social services they need or aim for. Alternatively, participation in these actions might allow them to make their voices heard as well as attract political and societal attention (Söderlund and Rapeli, 2015).

But which of these two perspectives prevails? There is a consensus in the academic literature on voting behaviour that poor health discourages voting by directing attention away from political matters to more personal ones (Pacheco and Fletcher, 2015). Poor health makes citizens feel less politically efficacious (Mattila et al., 2013; Schur and Kruse, 2000). Furthermore, it reduces an individual’s general interest in electoral politics (Pacheco and Fletcher, 2015). Yet, for other forms of political participation, the literature is less clear whether the resource theory or its competitor, the rational choice perspective, prevails. For example, while confirming that poor self-reported health diminishes somebody’s likelihood to vote, Söderlund and Rapeli (2015) find no influence of perceived health on contacting a politician, signing a petition, and boycotting. Mattila et al. (2018) evaluate the effect of two indicators of health – self-reported health and disability – on six different forms of political participation, that is, voting, contacting a politician or public official, working for an organization or association besides a political party, signing a petition, boycotting a product, and expressing a political opinion on social media. In a nutshell, the authors confirm that poor health, regardless of how it is measured, triggers a decreased likelihood to cast one’s vote. Yet, in their study, the relationship between health and other forms of political participation is more complex; sometimes it is positive and sometimes it is negative.

We try to advance our knowledge between health and political engagement on a more conceptual level. We ask the question, does it matter how we measure poor health in assessing the relationship between individual health and political involvement? To answer this question, we explicitly distinguish between self-reported or perceived health and some diagnosed physical or mental impairment, which we assess by the question if someone is hampered by illness in daily activities. Following the initial thoughts by Mattila et al. (2018), we assert that it is conceptually different for somebody to feel bad or sick or to have a long-standing illness. For sure, if somebody is sick, she might be less motivated to participate politically. Yet, this might be a temporary state of affairs. It might not influence an individual’s professional career, the money she has to spend to get treatment, or the time it takes to heal from the ailment. With respect to this, the relationship with different forms of political behaviour might be less pronounced. However, suffering from long-standing or chronical illnesses might influence all three factors. It can break, interrupt, or end a person’s professional career; it can be financially costly; and it can change or alter somebody’s network and friends.

Following Ojeda’s (2015) assumptions about the differentiated effect of depression, we argue that a long-standing illness can further have a stronger impact on one’s feeling of efficacy and trust in the political system. Overall, the assumption is that reporting a long-standing illness that hampers one’s daily life should have more deeply rooted long-term consequences than just a temporary state of poor health and thus may also exhibit a stronger relationship with measures of political engagement. While all these repercussions could put a further drain on political engagement of all sorts, they could also be a push factor for higher engagement, as the political stakes for such a person are much higher. For example, somebody with a long-standing illness might be dependent on public procurement of health care, a good infrastructure, and public transportation to remain mobile. She might therefore be more willing to fight for these goods, which could lead to an increase in political participation. A long-standing illness could also shift one’s perspective to support a long-term cause, such as fighting for one’s rights or the acceptance and acknowledgement within society. Accordingly, we expect to see a positive relationship between reporting a long-standing illness and engagement in more non-institutionalized forms of participation.

By explicitly measuring the relationship between both self-reported health and the more objective measure of being hampered by illness in daily life on different forms of political involvement, we test if the type of ailment, a more perceived ailment or a more objective medical condition, make any difference in citizens’ decision to partake in the political process. This question becomes the more important considering that there is very little overlap between the two indicators. To illustrate, 10,211 individuals or about 30% of all the respondents of wave 7 of the ESS indicate that they are hampered by illness in daily life. Yet, out of these 10,211 individuals, only 469 affirm that their self-reported health is bad. Another 2099 claim that that their self-reported health is fair and the overall majority of 7643 respondents assert to be in good health. This distribution of answers clearly reveals that there is no overlap between the two answers. This finding is confirmed if we conduct a correlation analysis between self-reported health and hampered by illness in daily activities. We find that there is only a small correlation between these two indicators (the Pearson Correlation Coefficient is ρ = 0.39). Given that they are both statistically and theoretically different, the two health indicators can potentially have a varying influence on different forms of political participation. In this study, we try to find out the degree to which this is the case. We do so by including 20 European countries into the analysis, thus also increasing the scope of previous analysis (e.g. Mattila et al., 2018; Söderlund and Rapeli, 2015).²

Methods

Data and variables

In an attempt to be as inclusive as possible, we gauge the relationship between both poor self-reported health and the indicator hampered by illness in daily life and seven forms of political participation (i.e. voting, contacting a politician, signing a petition, demonstrating, being a member of a political party, wearing a campaign badge, and boycotting). To do so, we rely on data from the 17th wave of the ESS (2014) and expand prior research (e.g. Couture and Breux, 2017) by a cross-national comparison of 20 European countries. Our information stems from the 2014 release of the ESS (wave 7, 2014), which includes roughly 35,000 respondents nested in 20 European countries. These data are particularly suitable for our purposes as they incorporate both health and political behaviour measures. Moreover, we test the robustness of the results with the help of waves 1–6 of the ESS.

We have seven dependent variables, which we all code as binary variables. Voting is operationalized by the question: ‘Did you vote in [your country’s] last national elections?’ which can be either (1) yes or (0) no. Another variable, taking part in demonstrations is captured by the question: ‘During the last 12 months, have you taken part in a lawful demonstration’ (0 = no; 1 = at least once). The remaining participation forms are also binary variables which are coded 1 if the survey respondent has at least once within the past 12 months contacted a politician, signed a petition, worn a badge, or boycotted a product, respectively. Finally, the variable being a member of a political party is coded 1, if the respondent answers in an affirmative way, and 0 otherwise.

We capture our first independent variable, self-rated health, by the question ‘How is your health in general?’ There are five answering categories: ‘very good’, ‘good’, ‘fair’, ‘bad’, and ‘very bad’ ranging from 0 to 4 with higher values indicating better health. This broad operationalization of health is particularly suited for this study, as it is an umbrella category which integrates both physical and mental health issues (see Mavaddat et al., 2011; Schütte et al., 2013). In the analysis, we combine the categories ‘bad’ and ‘very bad’, as well as ‘good’ and ‘very good’, because there are fewer individuals in the outer categories (i.e. only 1.3% of the polled indicate that their health situation is very bad). The second indicator asked respondents if they are hampered in daily activities by illness, disability, infirmity, and/or mental problems. In its original version, the indicator consists of three categories: (1) yes a lot, (2) yes to some extent, and (3) no. In the analyses that follow, we combine categories 1 and 2 to one category (the logic is again that few people indicate that they are hampered a lot by illness in their daily lives) and create a dummy variable. We code all individuals, who are hampered by illness in their daily activities 0 and individuals not hampered 1.

We further take confounding factors into account. First, we implement age and age squared and hypothesize that as older individuals are more civic minded, more integrated in society, and more stable in their political ideologies, they are also more likely to participate in political action (McDowell, 2006). Second, we control for somebody’s socio-economic status (SES) and hypothesize that individuals’ likelihood to politically participate should increase with education (Melo and Stockemer, 2014). Education is captured by the number of years the respondent has spent in full-time education. Third, we add the gender of the respondent to the analysis. While, in recent decades, the gender imbalance – in favour of men – in various forms of political participation seems to have weakened (Gallego, 2010), the respondents’ gender remains a standard variable in political participation models (Hill and Huskey, 2015). Gender is operationalized by a dummy variable coded 1 for women and 0 for men.

Fourth, we control for religiosity and expect not only that individuals who are more religious have a higher likelihood to vote than the less religious ones, but they should also have a lower likelihood to participate in more protest-based activities such as demonstrating (Xu, 2005). Religiosity is an 11-item ordinal variable ranging from not religious at all (coded 0) to very religious (coded 10). Fifth and finally, we control for whether the respondent was born in the country the interview was conducted. Citizens born in the country of current citizenship probably should have a stronger attachment to the country, they ought to be more familiar with the political system, and they should have become more politically socialized in their home country than individuals who have acquired citizenship rights at a later age (Campbell, 2017). As for the previous variable, the variable born in the country is a dummy variable, coded 1 for individuals born in the country they reside and 0 otherwise.

Statistical procedures

To scrutinize the influence of our two health indicators on various forms of political participation, we use several analytical steps: first, we conduct univariate statistics of the five dependent variables as well as of the two main independent variables (see Table 1). For all variables, we find that there is variation between countries. We also find that self-reported turnout in the sample, which averages 77%, is highly over-reported.³ Because there are no official figures for the other types of political behaviour, we cannot judge if there is any bias there.

Table 1.

Univariate statistics per country (Source: European Social Survey 2014).

Country	Self-rated health	Hampered by illness	Voting	Contacting politician	Signing petition	Taking part in demonstrations	Member of political party	Wearing a campaign badge	Boycott
Austria	1.94	0.79	0.78	0.20	0.28	0.06	0.06	0.05	0.25
Belgium	1.95	0.75	0.94	0.15	0.23	0.07	0.04	0.07	0.15
Czech Republic	1.94	0.73	0.59	0.10	0.16	0.04	0.02	0.06	0.09
Denmark	1.94	0.73	0.94	0.22	0.31	0.06	0.05	0.10	0.28
Estonia	1.88	0.75	0.74	0.17	0.12	0.04	0.05	0.06	0.09
Finland	1.94	0.67	0.84	0.20	0.34	0.02	0.05	0.18	0.37
France	1.90	0.75	0.70	0.16	0.38	0.14	0.05	0.13	0.37
Germany	1.89	0.69	0.84	0.18	0.37	0.09	0.04	0.06	0.36
Great Britain	1.89	0.70	0.73	0.20	0.41	0.05	0.03	0.10	0.24
Hungary	1.83	0.71	0.74	0.09	0.05	0.03	0.01	0.01	0.03
Ireland	1.96	0.80	0.79	0.20	0.26	0.13	0.05	0.09	0.14
Lithuania	1.87	0.64	0.58	0.11	0.07	0.02	0.04	0.05	0.04
Netherlands	1.93	0.67	0.80	0.19	0.29	0.03	0.04	0.05	0.15
Norway	1.94	0.74	0.90	0.23	0.33	0.09	0.07	0.33	0.26
Poland	1.88	0.70	0.69	0.08	0.14	0.03	0.02	0.05	0.06
Portugal	1.85	0.79	0.72	0.16	0.15	0.07	0.05	0.05	0.06
Slovenia	1.88	0.65	0.68	0.15	0.12	0.04	0.04	0.02	0.07
Spain	1.87	0.83	0.81	0.17	0.33	0.24	0.09	0.12	0.18
Sweden	1.95	0.69	0.93	0.20	0.44	0.11	0.06	0.20	0.48
Switzerland	1.97	0.78	0.70	0.16	0.38	0.06	0.07	0.07	0.31
Mean	1.91	0.72	0.77	0.16	0.26	0.07	0.05	0.09	0.20

Means per country.

Second, we present bivariate analyses between both the self-reported health measure and the more objective health measure and our proxies for political participation, respectively (see Table 2). Third, we present the results of 14 multi-level logit regression models. Models 1–7 in Table 3 feature self-reported health as the main independent variable (see models 1–7 in Table 4). Models 8–14 in Table 5 feature hampered by illness in daily activities as the main independent variable. We estimate all models as multilevel logit models in which individuals are nested within the 20 countries. Even though we have a relatively low number of countries, the presented results stay robust if we implement other methods, such as country-dummies and fixed-effects.

Table 2.

Probability to take part in political activity by self-reported health.

Self-reported health	Voting	Contacting politician	Signing petition	Taking part in demonstrations	Being a member of a political party	Wearing a campaign badge	Boycott
Very bad/bad	0.65	0.17	0.18	0.03	0.04	0.07	0.17
fair	0.72	0.15	0.20	0.05	0.04	0.07	0.16
Good/very good	0.77	0.17	0.26	0.07	0.05	0.09	0.20

Table 3.

Multi-level models testing the effects of self-reported health on various forms of political participation.

	Model 1Turnout	Model 2Contacting politician	Model 3Signing a petition	Model 4Demonstrating	Model 5Member of a political party	Model 6Wearing a campaign badge	Model 7Boycotting
Self-reported health (ref. cat. very bad health)
Fair health	0.37**	−0.23	0.07	0.43	−0.08	−0.05	−0.15
Fair health	(0.12)	(0.15)	(0.15)	(0.30)	(0.29)	(0.22)	(0.15)
Good/very good health	0.83***	−0.22	0.13	0.58*	0.05	−0.01	−0.06
Good/very good health	(0.11)	(0.13)	(0.14)	(0.29)	(0.27)	(0.20)	(0.14)
Age 18 years and up	0.08***	0.08***	0.03***	0.01	0.04***	0.00	0.06***
Age 18 years and up	(0.00)	(0.01)	(0.00)	(0.01)	(0.01)	(0.01)	(0.00)
Age squared	−0.00***	−0.00***	−0.00***	−0.00**	−0.00**	−0.00	−0.00***
Age squared	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Education (years)	0.12***	0.10***	0.11***	0.08***	0.11***	0.08***	0.10***
Education (years)	(0.00)	(0.00)	(0.00)	(0.01)	(0.01)	(0.01)	(0.00)
Female	−0.13***	−0.33***	0.09***	−0.14**	−0.50***	0.11**	0.16***
Female	(0.03)	(0.03)	(0.03)	(0.04)	(0.05)	(0.04)	(0.03)
Religiosity	0.06***	0.04***	0.01	−0.05***	0.03**	0.03***	−0.01*
Religiosity	(0.01)	(0.01)	(0.00)	(0.01)	(0.01)	(0.01)	(0.01)
Born in country	0.65***	0.36***	0.31***	−0.27**	0.32**	0.17	0.18**
Born in country	(0.06)	(0.07)	(0.06)	(0.09)	(0.12)	(0.09)	(0.06)
Constant	−4.46***	−5.17***	−3.49***	−3.83***	−5.86***	−3.75***	−4.55***
Constant	(0.24)	(0.21)	(0.25)	(0.38)	(0.39)	(0.32)	(0.29)
Variance component	0.57** (0.18)	0.09** (0.03)	0.58** (0.18)	0.47** (0.15)	0.23** (0.08)	0.60** (0.20)	0.90** (0.29)
N	32,546	33,748	33,683	33,756	33,750	33,742	33,667
AIC	30,540.71	29,107.36	34,889.60	16,038.63	12,198.28	18,776.46	29,655.49
BIC	30,624.62	29,191.63	34,973.85	16,122.90	12,282.55	18,860.72	29,739.73

AIC: Akaike information criterion; BIC: Bayesian information criterion.

Standard errors in parentheses. Please note that we run model 1 without Belgium as the only compulsory voting country in our data. The results from this additional specification confirm the results reported in model 1.

p < .05; **p < .01; ***p < .001.

Table 4.

Probability to take part in political activity by hampered by illness.

Hampered by illness in daily life	Voting	Contacting politician	Signing petition	Taking part in demonstrations	Being a member of a political party	Wearing a campaign badge	Boycott
Yes	0.76	0.17	0.24	0.06	0.05	0.09	0.20
No	0.77	0.16	0.26	0.07	0.05	0.09	0.19

Table 5.

Multi-level models testing the effects of the variable hampered by illness in daily life on various forms of political participation.

	Model 8Turnout	Model 9Contacting a politician	Model 10Signing a petition	Model 11Demonstrating	Model 12Member of political party	Model 13Wearing a campaign badge	Model 14Boycotting
Not hampered by illness	0.27***	−0.13***	−0.16***	−0.07	−0.09	−0.16***	−0.19***
Not hampered by illness	(0.03)	(0.03)	(0.03)	(0.05)	(0.06)	(0.05)	(0.03)
Age 18 years and up	0.09***	0.08***	0.03***	0.01	0.04***	0.00	0.06***
Age 18 years and up	(0.00)	(0.01)	(0.00)	(0.01)	(0.01)	(0.01)	(0.00)
Age squared	−0.00***	−0.00***	−0.00***	−0.00**	−0.00**	−0.00	−0.00***
Age squared	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Education (years)	0.12***	0.10***	0.11***	0.08***	0.11***	0.08***	0.11***
Education (years)	(0.00)	(0.00)	(0.00)	(0.01)	(0.01)	(0.01)	(0.00)
Female	−0.13***	−0.33***	0.09***	−0.15***	−0.50***	0.11**	0.15***
Female	(0.03)	(0.03)	(0.03)	(0.04)	(0.05)	(0.04)	(0.03)
Religiosity	0.06***	0.04***	0.01	−0.05***	0.02**	0.03***	−0.01*
Religiosity	(0.01)	(0.01)	(0.00)	(0.01)	(0.01)	(0.01)	(0.01)
Born in country	0.66***	0.36***	0.32***	−0.27**	0.32**	0.17	0.19**
Born in country	(0.06)	(0.07)	(0.06)	(0.09)	(0.12)	(0.09)	(0.06)
Constant	−3.94***	−5.30***	−3.25***	−3.21***	−5.75***	−3.64***	−4.48***
Constant	(0.22)	(0.17)	(0.21)	(0.25)	(0.29)	(0.25)	(0.26)
Variance component	0.58** (0.19)	0.09** (0.03)	0.58** (0.19)	0.47** (0.15)	0.24** (0.09)	0.60** (0.20)	0.91** (0.29)
N	32,516	33,719	33,650	33,723	33,717	33,711	33,633
AIC	30,548.80	29,063.88	34,840.26	16,031.74	12,194.09	18,752.83	29,592.07
BIC	30,624.30	29,139.71	34,916.07	16,107.57	12,269.92	18,828.67	29,667.87

AIC: Akaike information criterion; BIC: Bayesian information criterion.

p < .05; **p < .01; ***p < .001.

We also rerun model 1 in Tables 3 and 5, featuring voting as the dependent variable, and only include countries (i.e. Belgium, Hungary, Portugal, Slovenia, and Sweden), which had national elections during the time of fieldwork. Including this robustness check avoids the problem of retrospective voting in surveys (i.e. over-reporting in individual turnout might be higher if the election was 2 or 3 years ago, compared with an election that happened during the time of the survey).⁴

Finally, we replicate models 1 to 14 with waves 1 to 6 of the ESS. We do so because we want to present robust findings with regard to the effect of our two health indicators on various forms of political participation and postulate that, if we find the same results every 2 years for a 12-year period, we can be rather certain that the relationships we report exist and are not driven by some exogenous factor (see Tables A1 to A14 in the Appendix 1).

Results

The bivariate results illustrate that the effects of either of the two health indicators on our seven forms of political participation are quite distinct. For the first indicator, we find that there is a gap towards higher engagement for individuals whose self-reported heath is good for all seven forms of political participation (see Table 2). In some instances, the absolute difference is negligible as in the case of contacting a politician, but more often, we see moderate absolute differences. For example, voting is 8 percentage points higher for citizens with good self-reported health, as compared to citizens with bad self-reported health, and citizens in good health have twice the likelihood to demonstrate than citizens in bad self-reported health.

If we look at the bivariate results for our second health proxy, hampered by illness in daily life, we find only minor differences. We see that the likelihood of voting is somewhat higher for respondents, who are not hampered by daily illness. Yet, when it comes to other forms of political participation, we find that individuals, who are hampered by illness in daily life either show no difference in participation with healthy individuals or show a slightly higher likelihood to participate. For example, they have a slightly higher probability to contact a politician or participate in a boycott.

The multivariate models add further nuance to the bivariate findings (see Tables 3 and 5). Models 1 and 8 support that healthy individuals vote more frequently, regardless of how we measure health. Table 6 further highlights that the effect of self-reported health on voting is the same if we restrict the analysis to countries in which elections took place during the ESS data collection period. As such, we confirm the prior literature on health and voting (Mattila et al., 2018; Schur and Krause, 2000); that is, there is a small net negative effect of poor health – whether real and/or perceived – on voting. The marginal effects plot in Figure 1 graphically displays this small negative effect of poor health on voting. We also find a small positive relationship between good self-reported health and participation in demonstrations (see Figure 2). Yet, this significant and relevant finding is not present for the relationship between poor self-reported health and other forms of political participation. There is no statistically significant negative linkage between self-reported health and any of the five forms of political action – contacting a politician, signing a petition, being a member of a political party, wearing a campaign badge, and boycotting, even if the net difference between individuals in good self-reported health and bad self-reported health is sometimes perceptible in the bivariate realm. As such, it seems that in the complete models featuring other forms of political action than voting (see models 2–7) other factors such as age or education trump this effect. The replication of these findings with waves 1–6 of the ESS confirms these findings (see Tables A1 to A7 in the Appendix 1)

Table 6.

Turnout models in countries that held elections prior to survey (Source: European Social Survey 7).

	Turnout model self-reported health	Turnout model illness
Self-reported health (ref. cat. very bad health)
Fair health	0.48*
Fair health	(0.21)
Good/very good health	0.87***
Good/very good health	(0.19)
Not hampered by illness in daily life		0.25**
Not hampered by illness in daily life		(0.08)
Age	0.09***	0.09***
Age	(0.01)	(0.01)
Age squared	−0.00***	−0.00***
Age squared	(0.00)	(0.00)
No. of years in education	0.08***	0.08***
No. of years in education	(0.01)	(0.01)
Female	−0.02	−0.01
Female	(0.07)	(0.07)
Religiosity	0.03*	0.03*
Religiosity	(0.01)	(0.01)
Born in country of citizenship	0.74***	0.74***
Born in country of citizenship	(0.13)	(0.13)
Intercept	−3.28***	−2.73***
Intercept	(0.52)	(0.50)
Variance	0.73	0.76
Component	(0.47)	(0.49)
N	6966	6968
AIC	5849.24	5863.27
BIC	5917.72	5924.91

AIC: Akaike information criterion; BIC: Bayesian information criterion.

Standard errors in parentheses.

p < .05; **p < .01; ***p < .001.

Figure 1.

Marginal effect of self-reported health on turnout.

Figure 2.

Marginal effect of self-reported health on taking part in a demonstration.

Interpreting the effect of the variable hampered by illness in daily activities yields somewhat more surprising results. In addition to the expected finding that citizens, who report that they are hampered by illness, vote less (see model 8 and Figure 3), our results actually portray a statistically significant and positive effect of reported disability on the four forms of political action: contacting a politician, signing a petition, wearing a campaign badge, and boycotting (see models 9, 10, 13, and 14). While this reported effect in the models is substantively very small (see Figures 4 to 7), it nevertheless sheds some new light on the relationship between health and political engagement. It seems that individuals, who suffer from disability that hinders them in their daily lives, participate equally or in slightly higher numbers than healthy individuals in some forms of political actions other than voting. For them, a lack of resources and time might well be push back factors to participate. Yet, being hampered and possibly dependent on government help, they also have a lot to gain or lose from politics, which might push them towards political actions. As for the prior health indicator, the replication of these models with waves 1–6 of the ESS confirms these findings (see Tables A8 to A14 in the Appendix 1).

Figure 3.

Marginal effect of hampered by illness in daily life on turnout.

Figure 4.

Marginal effect of hampered by illness in daily life on contacting a politician.

Figure 5.

Marginal effect of hampered by illness in daily life on signing a petition.

Figure 6.

Marginal effect of hampered by illness in daily life on wearing a campaign badge.

Figure 7.

Marginal effect of hampered by illness in daily life on boycotting products.

To put these findings into perspective, it is important to note that self-reported health only slightly correlates with other resources for political participation such as education and income. The correlation between self-reported health or being hampered by illness in daily life and years of education is 0.18 and 0.16, respectively. Additional models (available on request) further illustrate that interactions between both self-reported health and being hampered by illness with years of education do not yield any significant or relevant results, which highlights that the effect of both health indicators is rather independent of other resources. These results further imply that health is a resource that has only limited influence on somebody’s likelihood to vote or to engage in other forms of political action. Yet, it is also resource with lots of intricacies. Building on the results of this study, it seems that it matters how we measure health. For example, self-reported health does not seem to have any influence on different forms of political actions, except for voting. In contrast, the more objective measure – hampered by daily illness – seems to actually slightly strengthen some forms of actions including boycotting and signing a petition.

The regression coefficients for the control variables in Tables 3 to 6 also offer some interesting results; that is, we find that middle age and senior citizens participate in any type of activity more frequently than younger citizens, except for demonstrations and wearing a campaign badge. Education consistently exercises a positive influence on all types of political activities. When it comes to gender, women are less likely to vote, contact a politician, be a member of a political party and demonstrate, but more likely to sign a petition, wear a campaign badge, and partake in a boycott. Individuals, who are born in the country of citizenship, generally display higher political engagement, except for demonstrating, whereas non-nationals by birth show a higher likelihood to partake in demonstrations. Religious individuals participate more in the four conventional forms, voting, contacting a politician, being a member of a political party, and wearing a campaign badge, but they are less likely to partake in a demonstration or boycott, than non-religious individuals.

Discussion

The aim of this article was to holistically compare two health indicators – self-reported health and hampered by illness in daily life – on different forms of political participation in a cross-national perspective. Our results are quite nuanced. Regardless of how we measure health, we find that poor health diminishes individuals’ likelihood to vote. Yet, this relationship is more complex for other forms of political participation. For example, at least in the multivariate realm, we find that poor self-reported health has no relevant influence on the six following forms of political actions: contacting a politician, signing a petition, demonstrating, being a member of a political party, wearing a campaign badge, and boycotting. However, if we take the second heath proxy, that is, hampered by illness in daily life, we actually find that individuals with disability participate more in four of the six forms, even if this effect is very small.

Hence, it matters for our analyses and for theory development how we measure health. When it comes to voting, things are quite easy, regardless how we measure health, individuals in poor health are less likely to vote. This finding supports the resource model of political participation.⁵ Yet, for other forms of political participation, there is either a null effect, if we use self-reported voting, or even a reversed effect for distinct forms of political actions, if we regress them on hampered by illness in daily life. Conceptually, this implies that we have to be explicit in what health indicator we use and justify it. Theoretically, future research should discuss the mechanisms that relate either (bad) self-reported health or manifest physical condition to different forms of political action. It might be possible that individuals with an objective physical or mental ailment vote less, because they think that they cannot achieve anything through voting. However, they might think that they could achieve some of their goals such as an appropriate health care and social protection system through other means of political participation such as boycotting. Future studies should confirm this conjecture.

We also want to be cautious about the interpretation of our results and suggest that health issues in the population are rather underestimated than overestimated. Health issues are still a taboo topic. With bad health in the fields of obesity, drug addiction, or diabetes (to name a few) comes societal stigmatization. Hence, respondents might embellish their health history in anonymized surveys and might indicate that their (perceived) health is better than their actual health is. Equally important, studies evaluating the influence of health on political participation are mainly conducted in the industrialised world, a setting where health standards are normally high and medical treatment is available for most, if not, all residents. In contrast, in many Third World countries, there are no welfare states or social benefits for the unhealthy. In such a situation, bad health, whether real or perceived, might have a much more detrimental influence on somebody’s likelihood to participate in the political process than in the highly industrialised world.

More generally, health is certainly a topic of ever increasing importance. Following demographic changes, almost all European countries are faced with an ageing population and, consequently, increasing health issues. Yet, health problems are not restricted to the elderly. Rather contrary, we can observe rising levels of mental health problems, in particular in younger generations. Young adults, in increasing numbers, mentally suffer from accelerated life styles as well as the constant fear of being less well-off than older generations, for example, due to temporary working contracts, less well-paid job opportunities, or a bad economy. These challenges might further hamper or increase young adults’ willingness and capability to participate politically.

Footnotes

Appendix 1 Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

ORCID iD

Daniel Stockemer

Author biographies

Daniel Stockemer is Associate Professor in the School of Political Studies at the University of Ottawa. Daniel’s research interests lie in the areas of political participation and representation, as well as populism. Among others, his work has appeared in the European Journal of Political Research, Governance, Political Behavior and Party Politics.

Carolin Rapp is Assistant Professor of Political Science at the Department of Political Science at the University of Copenhagen. Her research interest lies at the intersection between Political Sociology, Political Psychology, and Social Policy Research. Among others, Carolin’s research has appeared in the European Political Science Review, Social Science Research, Political Studies, and the Journal of European Social Policy.

References

Barnes

(1979) Political Action: Mass Participation in Five Western Democracies. Thousand Oaks, CA: SAGE.

Brady

Sidney

Schlozman

(1995) Beyond SES: A resource model of political participation. American Political Science Review 89(2): 271–294.

Burden

Fletcher

Herd

Jones

Moynihan

(2017) How different forms of health matter to political participation. The Journal of Politics 79(1): 166–178.

Campbell

(2017) Gender and voting. In: Arzheimer

Evans

Lewis-Beck

(eds) The Sage Handbook of Electoral Behavior. Thousand Oaks, CA: SAGE, p.159.

Couture

Breux

(2017) The differentiated effects of health on political participation. European Journal of Public Health 27(4): 599–604.

ESS (2014) Round 7: European Social Survey Round 7 Data. Data file edition 2.2. NSD, Norwegian Centre for Research Data, Norway. Data Archive and distributor of ESS data for ESS ERIC. doi:10.21338/NSD-ESS7-2014.

Gallego

(2010) Understanding unequal turnout: Education and voting in comparative perspective. Electoral Studies 29(2): 239–247.

Gallego

Oberski

(2012) Personality and political participation: The mediation hypothesis. Political Behavior 34(3): 425–451.

Gollust

Rahn

(2015) The bodies politic: Chronic health conditions and voter turnout in the 2008 election. Journal of Health Politics, Policy and Law 40(6): 1115–1155.

10.

Hill

Huskey

(2015) Electoral stakes, labor migration, and voter turnout: The 2011 presidential election in Kyrgyzstan. The Journal of Post-Soviet Democratization Demokratizatsiya 23(1): 3–30.

11.

Marx

Nguyen

(2016) Are the unemployed less politically involved? A comparative study of internal political efficacy. European Sociological Review 32(5): 634–648.

12.

Mattila

Rapeli

Wass

Söderlund

(2018) Health and Political Engagement. London: Routledge.

13.

Mattila

Söderlund

Wass

Rapeli

(2013) Healthy voting: The effect of self-reported health on turnout in 30 countries. Electoral Studies 32(4): 886–891.

14.

Mavaddat

Kinmonth

Sanderson

Surtees

Bingham

Khaw

(2011) What determines self-rated health (SRH)? A cross-sectional study of SF-36 health domains in the EPIC-Norfolk cohort. Journal of Epidemiology and Community Health 65(9): 800–806.

15.

McDowell

(2006) Measuring Health: A Guide to Rating Scales and Questionnaires, 3rd edn. Oxford: Oxford University Press.

16.

Melo

Stockemer

(2014) Age and political participation in Germany, France and the UK: A comparative analysis. Comparative European Politics 12(1): 33–53.

17.

Ojeda

(2015) Depression and political participation. Social Science Quarterly 96(5): 1226–1243.

18.

Pacheco

Fletcher

(2015) Incorporating health into studies of political behavior: Evidence for turnout and partisanship. Political Research Quarterly 68(1): 104–116.

19.

Schur

Kruse

(2000) What determines voter turnout? Lessons from citizens with disabilities. Social Science Quarterly 81(2): 571–587.

20.

Schütte

Chastang

Parent-Thirion

Vermeylen

Niedhammer

(2013) Social differences in self-reported health among men and women in 31 countries in Europe. Scandinavian Journal of Public Health 41(1): 51–57.

21.

Smets

van Ham

(2013) The embarrassment of riches? A meta-analysis of individual-level research on voter turnout. Electoral Studies 32(2): 344–359.

22.

Söderlund

Rapeli

(2015) In sickness and in health. Politics and the Life Sciences: The Journal of the Association for Politics and the Life Sciences 34(1): 28–43.

23.

Stryker

Owens

White

(eds) (2000) Self, Identity, and Social Movements: Social Movements, Protest, and Contention, Vol. 13. Minneapolis, MN: University of Minnesota Press.

24.

Taylor

Moghaddam

Gamble

Zellerer

(1987) Disadvantaged group response to perceived inequality: From passive acceptance to collective action. The Journal of Social Psychology 127(3): 259–272.

25.

van Deth

(2009) Politische partizipation. In: Kaina

Roemmele

(eds) Politische Soziologie, Berlin; Heidelberg: Springer, pp.141–161.

26.

Vissers

Stolle

(2014) The internet and new modes of political participation: Online versus offline participation. Information, Communication & Society 17(8): 937–955.

27.

Wass

Mattila

Rapeli

Söderlund

(2017) Voting while ailing? The effect of voter facilitation instruments on health-related differences in turnout. Journal of Elections, Public Opinion and Parties 100(4): 503–502.

28.

(2005) Why do minorities participate less? The effects of immigration, education, and electoral process on Asian American voter registration and turnout. Social Science Research 34(4): 682–702.

29.

Zeglovits

Kritzinger

(2014) New attempts to reduce overreporting of voter turnout and their effects. International Journal of Public Opinion Research 26(2): 224–234.