Can Sustainable Consumption Trigger Political Activism? An Empirical Investigation of the Crowding-in Hypothesis

Abstract

In this article, we address the question whether political activism can be triggered by sustainable consumption. Specific attention is given to the crowding-out and crowding-in hypotheses. The first hypothesis is driven by a conflict view as it assumes that sustainable consumerism displaces the willingness to act collectively. In contrast, the latter hypothesis—crowding-in—frames conscious consumption as a potential political act whereby individual sustainable consumption may trigger political acts such as signing a petition, demonstrating, and voting. To address this issue, German survey data were analyzed (n = 936). Our analysis appears to confirm the crowding-in hypothesis. However, the results of multiple logistic regression analyses also show that the relation between sustainable consumption and political activism depends on the type of political action. Particularly, sustainable consumption does not relate to traditional political actions such as voting, but it does relate positively to less conventional (e.g., attending a demonstration) and online forms of political engagement (e.g., social media activism). Our findings also indicate that the positive association between sustainable consumption and less conventional politics may be moderated by educational attainment, suggesting that it is weakest among less educated groups. The paper ends with the empirical and theoretical conclusions that can be drawn from this study, and indicates some directions for future research.

1. Introduction

Many prominent sociologists who have tried to characterize our contemporary society have focused on its internal differentiation. For Bourdieu, for example, different “fields” have developed as contingent outcomes of processes we might call modernization in the 19th and early 20th century. The emerging fields, such as art, religion, politics or science, acquire relative autonomy. They are, in Bourdieu’s view [1], shaped by competition for a specific type of symbolic capital. Luhmann has likewise put emphasis on the heterogeneity of a differentiated societal system. Much of his work looks at the internal closure (autopoiesis) of specific “function systems”, such as religion, politics, science, or the economy [2]. Quite similar lines of inquiry have, in recent years, been explored in other variants of “field theory”, “systems theory”, as well as in other sociological approaches [3,4].

An important question, which ensues from such diagnoses of the modern social (dis)order, concerns the relationships between investments in different fields or systems. To what degree do the social expectations, which exist in different settings, converge or not? How do individuals behave when they are confronted with diverging expectations, for example, in economic and political settings? How can the evolution of our society be steered in particular directions, when forms of internal differentiation dominate? By means of an empirical case-study, and in light of current environmental challenges, we explore in this paper in more depth the relationship between consumerism and political activism. The underlying question is whether both sets of expectations are mutually exclusive or whether it is possible that they both align with one another and even reinforce each other.

In the literature on ethical consumerism and political activism, two opinions can be identified in this regard. On the one hand, there are the skeptics, such as Johnston [5], who emphasize the contradictions between acting as a consumer and acting as a citizen. Consumerism would largely emphasize choices for one’s own needs, while citizenship emphasizes responsibility for the other and for a better social order [6]. Because consumerism and citizenship are seen to belong to different “value spheres”, it is at times also added that we should not inject the needs-driven, self-seeking behavior from the economy into the political sphere. The latter should be conceived of as a reasonable democracy seeking for the collective good [7,8]. No reconciliation between the spheres of the market and the politics seems imminent. In the environmental movement, such tensions between consumerism and citizenship are also becoming more acute. For example, Kenis [9] (p. 949) stated that grassroots environmental movements “have recently started to question the focus on sustainable consumption as a main strategy to tackle climate change. They prefer to address individuals as citizens rather than as consumers, and focus on collective rather than individual change.” Altogether, there is, within this line of research, little belief in the potential of sustainable consumerism to facilitate socio-ecological change.

On the other hand, a second line of research can be detected where it is assumed that an increasing erosion of institutional trust has led consumers to make use of the market to pressure both the public and private sectors to implement more sustainable policies [10]. Micheletti and Stolle [11] speak about the establishment of a sustainable citizenship in which individuals try to assume social responsibility and meet ecological challenges. They understand the concept of sustainable citizenship as “the role that people from all walks of life should play in taking daily responsibility for global sustainable development” [11] (p. 90). Shopping decisions may illustrate this expanded notion of sustainable citizenship. In addition to boycotting, reference is thereby made to buycotting, indicating the existence of both negative and positive political consumer actions not only with regard to food but also energy usage, and transport and mobility [12]. This is consistent with the evolution in research into political consumerism whereby the initial focus on buy- and boycotting has been extended to a broader definition that also includes discursive and lifestyle political consumerism [13,14,15]. The role of citizen-consumers in global environmental change can thus also be highlighted [16,17]. The underlying idea is that “systemic” couplings between consumers and citizens, or rather the economy and politics, can be (re)established by means of conscious consumption [14,18].

Societal differentiation does not exclude that some form of integration exists. Difference does not have to imply indifference; autonomous fields or systems do not cut off all links with other fields or systems. While different opinions regarding the social value of couplings can be defended, it is also useful to look for ways in which a difference (in one system) makes a difference (in another system). (The expression “a difference that makes a difference” was first used by Gregory Bateson [19] (p.465). He used it to define the way information “works” between sender and receiver). In this light, an interesting approach can be found in the work of Willis and Schor [18]. In an article, controversially titled “Does Changing a Light Bulb Lead to Changing the World? Political Action and the Conscious Consumer”, two empirical models are contrasted: the crowding-out versus the crowding-in hypotheses. In the first model, it is assumed that conscious consumption substitutes for political action [5,6,20]. In contrast, the crowding-in hypothesis states that conscious consumption should be considered as a new form of civic and political engagement. Based on the analyses of two large-scale datasets, Willis and Schor [18] conclude that conscious consumers and political activists are likely to be the same individuals. Following their analyses, sustainable consumption seems to “crowd in” current political activism, even when the relationship is controlled for prior political engagement.

In addition to using standardized items from the General Social Survey (GSS), Willis and Schor organized a survey in which a comprehensive scale was developed for the measurement of conscious consumption. In our study, we build on this by investigating how consistently respondents engage in fifteen sustainable consumption practices, such as the (less frequent) use of disposable water bottles, the purchase of local food, and the use of second hand goods (see Section 2). At the same time, we try in this article to go a step further at the level of the dependent variable by distinguishing between different types of political action. The current evidence for the crowding-in hypothesis is based on a general measure of political activism. In their classic study, Willis and Schor did not distinguish between traditional forms of political activism, such as “voting behavior”, and other political actions, such as participating in a demonstration or in online political communication. However, given the fact that the repertoire of political activism has become increasingly diversified in the past few decades [21], it is worthwhile to study whether the type of political action that individuals are engaged in matters.

In our study, we empirically examine whether the results on political consumerism also apply to sustainable consumption, given that relatively little research has been done on the relationship between sustainable consumption and political activism. In the words of Micheletti and Stolle [11] (p. 92), sustainable citizenship unfolds “outside formal political institutions more so than does traditional citizenship, which has tended to focus on the nation-state and the parliamentary sphere.” Accordingly, it makes sense to expect that sustainable consumption does relate to less conventional and new forms of activism like online political communication but not, or to a lesser extent, to traditional forms of activism such as voting. Against this background, we here aim at shedding new light on the crowding-in hypothesis by empirically investigating whether the relationship between sustainable consumption and political action depends on the type of political activism. In what follows, we give further details about the measurements and statistical techniques used in this case-study. After the presentation and discussion of the results, we briefly return to the more general reflections from which we departed.

2. Materials and Methods

An online questionnaire was organized in Germany between June and July 2014. From the 2331 persons from a panel that were invited to take part in this study, 936 people actually participated. The quota sampling strategy was used to make the sample representative for the German population averages on gender, age, education, occupation, and income. Except for people with no schooling and students less than 18 years of age (who were underrepresented by, respectively, 4.5% and 3.3% in comparison to the German population) all quotas were filled (cf. Appendix A). In the survey, three type of variables were included (cf. below for a more detailed description): (i) socio-demographic characteristics (age, gender, and education), (ii) how consistent people engage in sustainable consumption practices, and (iii) political activism (traditional, less conventional, and online political actions). As will be shown in the results section, we first compare mean scores for sustainable consumption between socio-demographic groups, based on t- and F-tests. Additionally, we use cross tables to examine whether socio-demographic groups differ in their (percentual) engagement in political activism. Significance was evaluated based on chi-square statistics. Lastly, we end our analysis with several logistic regression models in which different forms of political activism are included as dependent variables.

Outcome variables. Three different measures of activism were included in our survey: (i) traditional forms of activism; (ii) less conventional acts; and (iii) online political communication. Specifically, survey participants were asked whether they had been engaged in any of the following activities in the past two years: voting in any election and/or contacting a politician or a civil servant to express their views (traditional); signing a petition and/or taking part in a demonstration (less conventional); posting information about political issues on Facebook or other social media and/or forwarding emails or news articles about political issues (online political communication). The type of political action (cf. Figure 1) which most commonly occurs is voting. Over the last two years, eighty five percent of the survey participants went to vote. This is followed by signing a petition (36%), forwarding e-mails or news articles about political issues (30%), and posting political information on social media (27%). The least frequently occurring acts are contacting a politician or civil servant (14%) and taking part in a demonstration (9%).

Explanatory variables. Our focus on sustainable consumption practices, rather than political consumerism, is both narrower and broader [22]. It is narrower because in our study we did not estimate dispositional factors. However, it is broader in the sense that in this study we look at 15 items such as energy use, food choices, and the use of secondhand goods. Specifically, sustainable consumption was measured by questioning: “How consistently do you engage in the following sustainable consumption practices”. Answers were assessed by an 8-point Likert scale ranging from 0 = never, 1 = very inconsistently to 7 = very consistently. The items related to practices in different sustainability domains, such as energy (e.g., reducing energy use), transport (e.g., use of alternative modes of transportation), water (e.g., less use of disposable water bottles), food (e.g., buying local food), and goods (e.g., use of second hand goods). In total, 15 items were included in the analysis (cf. Figure 2). The practices which people report to be least engaged with are actions that require specific knowledge and skills such as the repairing of goods, or activities that require a certain infrastructure or place such as the growing of food. The most frequently occurring actions seem to be those where the financial and ecological benefits coincide, as in the case with the reduction of energy and the conservation of water.

In order to identify potential clusters of items, a principal component analysis was run. A scree plot suggested the presence of one underlying component, explaining 43.3% of the total variance. A reliability analysis revealed a Cronbach’s alpha value of 0.9. There was no significant improvement in the overall reliability when one of the fifteen items was removed from the scale. Accordingly, in subsequent analyses the composite component scores of sustainable consumption are used.

As the relationship between sustainable consumption and political activism is plausibly bidirectional [18], and given the use of cross-sectional data in this study, a control variable was included to measure previous political engagement: “Thinking back five years ago to 2009, how involved were you in political activities at that time?” Scores ranged from 1 = not at all politically active to 7 = very politically active.

Other measures that were included in the present study relate to the social demographic characteristics gender, age, and educational background. About half of the persons in our sample were female (49.3%). From the 936 respondents in our sample 82 aged 18–24 (8.8%), 309 aged 25–44 (33%), 342 aged 45–64 (36.5%), and 203 aged 65+ (21.7%). Differences with regard to educational background were as follows: 38.6% of the respondents hold a degree of lower secondary education or less, 45.9% attained the level of higher secondary education, and 15.5% hold a degree in higher education.

In our analysis, we first examine bivariate association between political action and the explanatory variables, on the one hand, and between the different types of political action, on the other hand. Subsequently, we use multiple logistic regressions to estimate the probability of engaging in political action given the values of explanatory variables. The analysis was conducted stepwise and parallel for each of the types of political action. In step 1 the social demographics were entered, followed by the variables on sustainable consumption (step 2) and previous political engagement (step 3). Lastly, the models were tested for interaction effects (step 4). The latter were only included in case of statistical significance, thereby reducing potential problems of multicollinearity [23].

3. Results

Results of independent t-tests and analyses of variance (cf.

Table 1. Comparing means and percentages: test statistic (p-value).

	Sustainable Consumption	Political Activism
	Overall (Mean; SD)	Traditional (%Yes)	LessConv (%Yes)	Online (%Yes)
Gender	Male: −0.12; 1.02	Male: 89.6%	Male: 38.7%	Male: 42.5%
	Female: 0.12; 0.97	Female: 81.5%	Female: 38.6%	Female: 34.9%
	t = 3.65 ***	χ2 = 12.21 ***	χ2 = 0.00	χ2 = 5.70 *
Age	18–24: −0.04; 0.90	18–24: 83.3%	18–24: 47.6%	18–24: 53.7%
	25–44: −0.07; 1.12	25–44: 79.6%	25–44: 37.2%	25–44: 39.5%
	45–64: 0.03; 0.95	45–64: 85.4%	45–64: 38.6%	45–64: 35.7%
	65+:0.08; 0.93	65+: 96.1%	65+: 37.4%	65+: 36.9%
	F = 1.11	χ2 = 27.24 ***	χ2 = 3.14	χ2 = 9.39 *
Education	LSE: 0.02; 1.01	LSE: 81.9%	LSE: 30.2%	LSE: 28.5%
	HSE: −0.02; 1.03	HSE: 86%	HSE: 40.2%	HSE: 41.9%
	HE: 0.02; 0.88	HE: 94.4%	HE: 55.2%	HE: 55.2%
	F = 0.23	χ2 = 13.09 ***	χ2 = 28.03 ***	χ2 = 34.10 ***

LSE = lower secondary education or less; HSE = higher secondary education; HE = higher education; *** p ≤ 0.001.

) first show that women reported more engagement with sustainable consumption practices (M = 0.12; SD = 0.97) than men (M = −0.12; SD = 1.02; t = 3.65, p ≤ 0.001). Age and educational background do not relate significantly to overall sustainable consumption.

The extent to which people are politically active in the present significantly correlates with previous political involvement (cf.

Table 2. Comparing the types of political action: spearman correlation matrix.

Correlates	Traditional	Less Conventional	Online	Prev. Pol. Involvement	Sustainable Consumption
Traditional	1	0.17 ***	0.15 ***	0.25 ***	0.10 **
Less conventional	0.17 ***	1	0.30 ***	0.25 ***	0.14 ***
Online	0.15 ***	0.30 ***	1	0.40 ***	0.15 ***
Prev. pol. involvement	0.25 ***	0.25 ***	0.40 ***	1	0.22 ***
Sustainable consumption	0.10 **	0.14 ***	0.15 ***	0.22 ***	1

** p ≤ 0.01; *** p ≤ 0.001.

). The correlations are weak but significant in case of traditional and less conventional politics (in both cases, the nonparametric Spearman correlations equal 0.25; p ≤ 0.001) and moderate in case of online political communication (ρ = 0.40; p ≤ 0.001). Correlations among the three dependent variables further show that online political communication is moderately associated with less conventional political actions (ρ = 0.30; p ≤ 0.001). Both variables in turn are only weakly correlated with more traditional forms of political action (ρ = 0.15; p ≤ 0.001 for online political communication and ρ = 0.17; p ≤ 0.001 for less conventional political acts). These findings support the above argument that traditional, less conventional, and online political actions are different constructs that require separate estimates.

Furthermore, political activity seems to associate positively with sustainable consumption. Traditional political involvement is least correlated with sustainable consumption (ρ = 0.10; p ≤ 0.01). Less conventional forms of political action and online political communication are more positively associated with sustainable consumption (respectively, ρ = 0.14; p ≤ 0.001 and ρ = 0.15; p ≤ 0.001). Lastly, a significant correlation between sustainable consumption and previous political engagement (ρ = 0.22; p < 0.001) indicates a positive relationship between the two.

Results of Chi-square tests (cf. Table 1) further show that there is a significant association between gender, on the one hand, and traditional political actions (χ² = 12.21; p ≤ 0.001) and online political communication, on the other hand (χ² = 5.70, p ≤ 0.05). In both cases, men are more likely to be politically involved. Age seems to be weakly but significantly associated with online activism (χ² = 9.39, p ≤ 0.05) and rather moderately associated with traditional political acts such as voting (χ² = 27.24, p < 0.001). Specifically, whereas older people are more likely to be politically involved in a traditional way, younger people are more likely to be politically involved through the use of the World Wide Web. Finally, educational background also significantly associates with political activism. In particular, higher educated groups are more likely to be involved politically in traditional, less conventional, and online ways ((χ² is resp. 13.09 (p < 0.001), 28.03 (p < 0.001), and 34.10 (p < 0.001)).

The role of social-demographic characteristics is further assessed in the logistic regression models (see

Table 3. Logistic regression: Determinants of traditional political action.

	Model 1		Model 2		Model 3
	B(SE)	Exp b	B(SE)	Exp b	B(SE)	Exp b
Gender (male)	0.56 ** (0.20)	1.75	0.62 ** (0.20)	1.86	0.45 * (0.21)	1.57
Age (ref. 65+)
18–24	−1.60 *** (0.49)	0.20	−1.54 ** (0.49)	0.21	−1.54 ** (0.50)	0.21
25–44	−1.79 *** (0.39)	0.17	−1.72 *** (0.39)	0.18	−1.68 *** (0.40)	0.19
45–64	−1.33 *** (0.40)	0.27	−1.30 *** (0.40)	0.27	−1.19 ** (0.40)	0.30
Education (ref. higher educ.)
lower secondary education or less	−1.26 *** (0.40)	0.28	−1.23 ** (0.40)	0.29	−1.07 ** (0.40)	0.34
higher secondary education	−0.81 * (0.40)	0.44	−0.77 ^(0.40)	0.46	−0.70^(0.40)	0.50
Sustainable consumption			0.24 ** (0.09)	1.27	0.10 (0.10)	1.11
Prev. pol. involvement					0.55 *** (0.10)	1.73
Constant	3.81 *** (0.52)	44.96	3.73 *** (0.52)	41.51	2.57 *** (0.55)	13.04
Nagelkerke R square	10.5%		11.7%		18.8%

^ p ≤ 0.1; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.

,

Table 4. Logistic regression: Determinants of less conventional politics.

	Model 1		Model 2		Model 3		Model 4
	B(SE)	Exp b	B(SE)	Exp b	B(SE)	Exp b	B(SE)	Exp b
Gender (male)	−0.03 (0.14)	0.98	0.06 (0.14)	1.06	−0.08 (0.15)	0.92	−0.08 (0.15)	0.92
Age (ref. 65+)
18–24	0.36 (0.28)	1.43	0.42 (0.28)	1.52	0.47 ^ (0.29)	1.60	0.46 (0.29)	1.58
25–44	0.02 (0.19)	1.02	0.07 (0.19)	1.07	0.16 (0.20)	1.18	0.15 (0.20)	1.16
45–64	0.11 (0.19)	1.12	0.14 (0.19)	1.15	0.25 (0.20)	1.28	0.24 (0.20)	1.27
Education (ref. higher educ.)
lower secondary education or less	−1.05 *** (0.20)	0.35	−1.06 *** (0.21)	0.35	−0.98 *** (0.21)	0.37	−0.96 *** (0.21)	0.38
higher secondary education	−0.65 *** (0.20)	0.52	−0.64 *** (0.20)	0.53	−0.62 ** (0.20)	0.54	−0.62 ** (0.21)	0.54
Sustainable consumption			0.34 *** (0.07)	1.41	0.25 *** (0.08)	1.28	0.39 *** (0.10)	1.48
Prev. pol. involvement					0.29 *** (0.05)	1.33	0.29 *** (0.05)	1.34
lower secondary education or less * sustainable consumption							−0.36 * (0.15)	0.70
Constant	0.16 (0.23)	1.18	0.08 (0.23)	1.08	−0.67 * (0.26)	0.52	−0.67 * (0.26)	0.51
Nagelkerke R square	4.3%		7.5%		12.6%		13.3%

^ p ≤ 0.1; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.

and

Table 5. Logistic regression: Determinants of online political communication.

	Model 1		Model 2		Model 3
	B(SE)	Exp b	B(SE)	Exp b	B(SE)	Exp b
Gender (male)	0.33 * (0.14)	1.39	0.43 ** (0.14)	1.54	0.22 (0.16)	1.25
Age (ref. 65+)
18–24	0.67 * (0.28)	1.95	0.74 ** (0.28)	2.10	0.94 ** (0.30)	2.56
25–44	0.18 (0.19)	1.20	0.24 (0.20)	1.27	0.47 * (0.21)	1.60
45–64	0.05 (0.19)	1.05	0.07 (0.19)	1.08	0.29 (0.21)	1.34
Education (ref. higher educ.)
lower secondary education or less	−1.07 *** (0.21)	0.34	−1.09 *** (0.21)	0.34	−1.00 *** (0.22)	0.37
higher secondary education	−0.54 ** (0.20)	0.58	−0.53 ** (0.20)	0.59	−0.50 * (0.22)	0.61
Sustainable consumption			0.39 *** (0.07)	1.47	0.24 ** (0.08)	1.27
Prev. pol. involvement					0.54 *** (0.05)	1.72
Constant	−0.12 (0.23)	0.89	−0.22 (0.23)	0.80	−1.66 *** (0.28)	0.19
Nagelkerke R square	6.5%		10.4%		25.6%

* p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.

). The effect of gender on traditional political action remains significant when sustainable consumption and previous political involvement enter the model. The effect of gender on online political communication remains positive but becomes non-significant in step 3, after controlling for previous political involvement. Persons aged 65 or older consistently report to be more engaged with traditional political actions in comparison to younger people. For example, individuals in the over-65 age group are approximately 5 times more likely to act politically in a more traditional way than late adolescents. Conversely, young adults are almost 2 times more likely to be politically active by using the internet than the 45–64 age group (exp b = 1.92, p ≤ 0.05, not shown in table), and 2.56 times more likely than the over-65 age group (p ≤ 0.01). Finally, and consistent with the initial results of the Chi-square tests, higher educated persons tend to be more politically active. For example, in comparison with people with lower secondary education or less, persons with higher education degrees are almost 3 times more likely to be engaged in each of the three types of political action.

Furthermore, our results suggest that sustainable consumption positively relates to political activism. For traditional political action (see Table 3), however, this effect disappears in step 3 when the control variable on previous political activism enters the model. In contrast, the positive effect of sustainable consumption on less conventional politics remains significant throughout the analysis (exp b = 1.48; p ≤ 0.001). The same holds true for the effect of sustainable consumption on online political communication (Comparable results were attained when engagement with traditional and protest politics were added to the model as additional control variables) (exp b = 1.27; p ≤ 0.01). In other words, the relationship between sustainable consumption and political activism depends on the type of political action. Being more engaged in sustainable consumption practices does not relate significantly to traditional political actions such as voting but it does relate positively to less conventional politics and online political engagement. In a last step, we tested for interaction effects. A moderated relationship was found between less conventional political activism, on the one hand, and education and sustainable consumption, on the other (exp b = 0.70; p ≤ 0.05). While sustainable consumption positively affects the probability of engaging in less conventional politics in general, this relationship is considerably less strong for people who have a lower secondary education or less. In other words, whereas sustainable consumption appears to trigger less conventional politics for moderately and higher educated groups, this may not be the case, or at least to a lesser extent, for less educated groups.

4. Discussion

The crowding-in hypothesis, i.e., the supposition that conscious consumption may trigger political actions, has met with much skepticism in the scientific literature [5,6,8,20]. According to the sceptics, we should have little faith in the potential of sustainable consumerism to facilitate socio-ecological change. While this criticism is often reinforced by strong, ideologically charged labels, such as “anti-politics machine” or “depoliticized environmentalism” [20,21,22,23,24], the strength of such claims stands in stark contrast to their empirical substantiation. Using survey data, in this study we wanted to counterbalance this by empirically testing the relationship between sustainable consumption and political actions and the ways in which different, differentiated fields or systems interact.

The first part of the analyses explored the relationship between socio-demographic characteristics, on the one hand, and sustainable consumption and political activism, on the other. Results showed that men are more often involved in traditional political actions. Women, for their part, report to be significantly more involved in sustainable consumption. This is in line with previous research that suggests that women are more inclined to be sustainable consumers [25], possibly because women are socialized to emphasize self-transcendent values such as environmental protection [26,27]. The results of our study also provide additional support for the finding that women may not participate less, but differently in political actions [28]. In particular, as noted by Stolle and Micheletti [15], women seem to engage more in political consumerism (e.g., boycotting), while men engage more in traditional political action (e.g., contacting a politician). One explanation entails that women are generally more active in domestic spheres of life (e.g., shopping) than men, and that this reflects in gendered modes of political action. Further, age and educational background did not relate significantly to sustainable consumption practices. However, both socio-demographic factors did relate significantly to political activism. Specifically, whereas older people reported to be more involved with traditional political actions than young adults, the latter reported higher scores for online political communication. This confirms previous research which indicates that it is mainly young people who are less and less involved in traditional political actions such as voting [29]. Instead, younger people increasingly turn to so-called “politics of choice” [30]. Furthermore, young people are more familiar with technology and may be more flexible in adopting new behaviors. This makes it easier for them to take online political action, while older generations are less likely to take such actions. Additionally, we found that political participation is lowest among the lower educated groups. This is consistent with previous research that has shown that social class is negatively related to political participation [31].

In the second part of the analyses, it was shown that sustainable consumption did not relate to traditional political actions, such as voting, but that it did relate positively to less conventional politics (e.g., attending a demonstration) and online political communication (e.g., through social media like Facebook). This is in line with previous research that shows that political consumers do not engage more often than others in traditional forms of political action, but rather in unconventional forms of participation [10] and in what could be termed “innovative politics” [15]. These results partially confirm the crowding-in hypothesis as described in the study by Willis and Schor [18]. On the one hand, it is overlapping in the sense that sustainable consumption triggers—rather than crowding out—political activism. On the other hand, the findings of our study deviate from this in that the relationship between sustainable consumption and political activism depends on the type of political action. The assumption that it is necessary to differentiate between types of political action was also empirically supported. Online political communication was significantly associated with less conventional political actions, yet both variables in turn were only weakly correlated with more traditional forms of political action. These findings support the arguments that traditional, less conventional, and online political actions are different constructs (cf. supra) and that sustainable consumption practices associate with actions that take place outside formal political institutions [11]. Our findings further indicate that the positive association between sustainable consumption and less conventional politics may be moderated by educational attainment, suggesting that it is weakest among less educated groups.

The core findings of our study, that sustainable consumption is only related to online political communication and less conventional politics (especially among the higher educated), can first of all be interpreted as a response to the perceived failure of traditional politics. At the same time, this finding is in line with the more recent literature on the rise of ecologically oriented cultural orientations or a so-called “eco-habitus” [32]. A new group of people seem to have emerged, who are more prone to appropriate new forms of cultural capital and invest in alternative consumption [33,34]. These alternative consumption strategies might connect more strongly to alternative political engagement, compared to traditional forms of politics. Additionally, our findings resonate with the critique that these new cultural practices can be exclusive and elitist [35], given that less-educated people have typically less access to these “emerging forms of cultural capital” [36]. While we did not find educational differences in sustainable consumption, this could explain why in this study there was an interaction effect between sustainable consumption and less conventional politics, and not between sustainable consumption and online political communication.

Although the main conclusions of this study are situated at the empirical level, the results have in this regard theoretical implications. While we started from the observation that, according to several prominent sociologists, differentiation is characteristic of modern society, we also asked for the relationships between investments in different social fields or systems. What our results point to is that inequalities arising in one field become relevant for other fields. More particularly, participation in (higher) education seems to have become a source of societal inequality that transcends the boundaries of education itself. Perhaps as a consequence of the rapid expansion of higher education, the kind of knowledge, skills and values that people acquire in higher education has become “a difference that makes a difference” [19]. One does not have to go as far as some contemporary scholars, who characterize modern society as a “schooled society” [37,38], but our results also draw attention to the ways in which educational degrees produce other socially relevant differences. We believe that not only more empirical research about different kinds of relationships between education and other social fields is necessary to acquire a good view of these highly dynamic developments, but also that its theoretical implications nowadays need to be explored and analyzed in much more detail.

Finally, there are some shortcomings in this study which in turn offer avenues for future research. First, the questionnaire and data used for this study did not include items on political preferences and motivational factors. Furthermore, we only tested interaction terms between sustainable consumption and socio-demographic factors. Further research can investigate whether the relationship between sustainable consumption and political action is moderated by other variables such as dispositional factors and political ideology. Therefore, connections can be found to recent research on sustainable citizenship [22]. Second, we used cross-sectional data which makes it difficult to establish causal relationships. This was partially addressed by controlling for previous political engagement. To a certain extent, this allowed us to empirically test the crowding-in hypothesis, where the focus is on the unidirectional relationship between conscious consumption and political activism. However, we have not measured sustainable consumption practices from the past. Consequently, we do not know whether respondents may have been sustainable consumers in the years prior to participating in this study. Accordingly, it is recommended that future research makes use of longitudinal data. In addition, future research can make use of path modelling techniques to assess reciprocal effects. Thirdly, the data for this study were collected in 2014. In just a few years, the world has changed in various areas, including politics. Future research will have to show whether the results of this study are still valid today. Finally, as this study is based on quantitative data, it is difficult to develop a more refined and contextualized understanding of motives for action. We believe that qualitative research and mixed-method designs could shed additional light on the underlying motives of both sustainable consumption and political activism, and the relationships between them.