Exploring Interest Group Mobilization in the Netherlands PDF

Acta Politica https://doi.org/10.1057/s41269-020-00185-1 ORIGINAL ARTICLE Exploring the proportionality of representation in interest group mobilization and political access: the case of the Netherlands Marcel Hanegraaff1 · Joost Berkhout1 · Jens van der Ploeg2 Accepted: 7 November...

Acta Politica https://doi.org/10.1057/s41269-020-00185-1 ORIGINAL ARTICLE Exploring the proportionality of representation in interest group mobilization and political access: the case of the Netherlands Marcel Hanegraaff1 · Joost Berkhout1 · Jens van der Ploeg2 Accepted: 7 November 2020 © Springer Nature Limited 2020 Abstract Surveys reveal that there are great differences among citizens in their membership of political associations. Such differences plausibly lead to a better representation of interests of privileged citizens compared to other citizens. We examine the demo- graphic groups (in terms of education, gender and immigration background) that tend to be members of interest groups. We also investigate the relation between the membership profile of associations and the propensity of interest groups to be rou- tinely approached by policymakers. The results of our elite survey of Dutch interest associations indicate that relatively well-educated citizens and men are better repre- sented in interest groups. Patterns of underrepresentation are not further exacerbated by the outreach of policy-makers, except that interest associations with a relatively large female membership are less likely to be consulted. Keywords Interest groups · Lobbying · Representation · Political access Electronic supplementary material The online version of this article (https​://doi.org/10.1057/s4126​ 9-020-00185​-1) contains supplementary material, which is available to authorized users. * Marcel Hanegraaff [email protected] Joost Berkhout [email protected] Jens van der Ploeg [email protected] 1 University of Amsterdam, Amsterdam, The Netherlands 2 Algemene Rekenkamer, The Hague, The Netherlands Vol.:(0123456789)  M. Hanegraaff et al. Introduction The meaningful non-electoral representation of groups of citizens before govern- ment relies upon the willingness and capacity of citizens to become members of collective action organizations and upon policymakers to provide them a fair and equitable hearing. However, citizen participation in political organizations is nei- ther automatic nor equal. Schattschneider (1960, p. 32) argues that “large areas of the population appear to be wholly outside the system” of interest representation and laments the fact that “even non-business organizations reflect an upper-class tendency.” Recent surveys also indicate a strong inequality in political participation between, among others, highly and low educated citizens (e.g., Dalton 2017; Schloz- man et al. 2013, pp. 117–147; Theocharis and van Deth 2018). However, we do not know whether Schattschneider is correct in his assumption that this difference must lead the “heavenly chorus to sing with an upper-class accent.” because scholars lack (1) a substantively relevant benchmark of the “upper-class-ness” of interest groups and (2) only sporadically relate constituency characteristics to the voices are actually listened to by policymakers. For this reason, we ask: Do interest associations dis- proportionately represent relatively privileged individuals and do associations with more privileged members receive a more favorable hearing before government than other groups? We define interest groups as associations with individuals as members whose activities potentially intersect with public policy (Jordan et al. 2006). We contribute to the literature in three ways. First, we connect the systematic findings in studies on disproportionate political participation to studies of interest groups (Strolovitch 2006; Schlozman et al. 2013). Second, we assess whether any skewness in repre- sentation matters for political outcomes, that is, whether groups that represent rela- tively privileged members are more commonly provided with access to policymak- ers. Third, we assess more than a single demographic group. We compare interest groups in terms of education level, gender, and immigration background of their members. This comparison facilitates further theory formation regarding the under- lying mechanisms of representation via interest groups. Moreover, we think that our combined approach is a first step towards an assessment of, also, the interrelation- ship of these dimensions in terms of, for instance, intersectional dynamics within interest groups and aggregate changes over time in cross-class associability. Our data stem from an extensive survey among Dutch interest group leaders. We focus on associations of citizens and ask the leaders of these associations about the demographic characteristics of their members and the frequency with which poli- cymakers reach out to them for political input. First, we compare whether interest group membership deviates from the Dutch population on three key aspects: educa- tional level, gender, and migration background. We do not assess the political pref- erences of individual members of interest groups (e.g., Jordan and Maloney 2006). We examine of descriptive similarities between group members and the population rather than the substantive representation by interest group leaders of the political Exploring the proportionality of representation in interest… views of group members.1 In the next step, we analyze whether groups with mem- bers emanating from relatively privileged sections of society (e.g., groups represent- ing mostly highly educated citizens) are more often consulted by politicians than groups representing less privileged citizens. Even though our analysis is mainly descriptive, it provides us with exploratory insights into a question that, despite its normative implications and the attention to these issues in different literatures, has been vastly understudied in the interest group literature. We therefore hope our paper sparks more research into the relationship between unequal representation and interest group politics. In the following sections, we first theoretically identify the connections between political participation, political organization, and interest representation. We sub- sequently propose three hypotheses, describe our research design, and present our descriptive analysis on proportional representation and a multivariate assessment of policy access. We conclude with some final thoughts and identify several venues for future research. Theory: political participation, political organization, and interest representation We theoretically start with the idea that individuals possess organizational resources and that interest groups survive by bringing these resources together for some collec- tive purpose. This survival strategy is commonly conceived as a mutually beneficial exchange relationship between members and associations (Salisbury 1969; Berkhout 2013). Furthermore, we assume that leaders of organizations consider membership motivations, demands, and capabilities when deciding on organizational goals, and interest representation before government, and, in so doing, also build exchange relationships with policymakers (e.g., Schmitter and Streeck 1999). Interest associa- tions, as representatives of relevant groups in society, vary in their attractiveness to policymakers, and some of these interest associations will be in higher “demand” in the policy process than others (e.g., Leech et al. 2005). This variation results in patterns of exchange relationships between, on the one hand, citizens with particular interests and interest associations, and, on the other hand, interest groups and poli- cymakers. Interest associations’ simultaneous interaction in society (aggregation of interests) and in the policy process (articulation of interests) makes them perform an intermediary function in politics and act as a transmission belt (e.g., Truman 1951; Reher and Rasmussen 2019). Existing studies on political participation identify the relevant membership resource pools of associations, and prior work on interest group engagement in the 1 Please also note that our approach is conceptually distinct from studies of descriptive representation of elected representative of legislative bodies. We do not assess the descriptive characteristics of interest group leaders or lobbyists (e.g., Junk et al 2020) nor the similarity of these with the demographic charac- teristics of their members or the general population. We therefore draw on studies on differences among citizens in (electoral) political participation rather than studies of representation in public claimsmaking or parliaments.  M. Hanegraaff et al. Step 1 Step 2 Citizens Interest groups Policymakers - Education Do interest Proporonal? Proporonal? Do policymakers - Gender groups represent - Migration grant access to all types of groups equally? citizens equally? Fig. 1  Does unequal representation lead to unequal access? policy process indicates the types of resources that interest groups typically deliver to the policy process. We initially identify the motivations of citizens to become interest group members and subsequently evaluate the incentives for policymakers to provide access. We build our arguments in two parts (see also Fig. 1). In the first part, we explore whether interest group membership seems to be reflective in terms of key demographic dimensions, or whether interest associations relatively strongly rely on individuals from particular subsections of society. In the second part, we examine the relationship between the membership of organizations and the access they gain to policymakers, again analyzing whether underrepresentation of certain groups is magnified in the political process. As such, we explore whether potential inequalities in membership lead to the (selective) provision of access by policymak- ers to particular groups. Do interest groups give voice to those with an upper‑class accent? The well-established literature on political participation reveals that individuals differ in the degree and nature of political participation (e.g., Smets and Van Ham 2013). Most classically, this variance refers to the differences among groups of dis- tinct socioeconomic status (SES), with low-income groups being less likely to vote, protest, and participate politically in other ways (e.g., Verba et al. 1995). Recent studies indicate also that education, gender, and migration affect the political capa- bilities of individuals and consequently produce differences in political participation (e.g., Marien et al. 2010; Hakhverdian and Schakel 2019). All of these differences refer to turning up to vote in elections but also include non-conventional means of political participation such as protests, online actions, politically motivated con- sumer boycotts and, critical for our study, interest group memberships (e.g., Theo- charis and van Deth 2018). These findings on individual participation are the starting point for our first hypothesis, which we label the “proportional membership hypothesis.” This hypoth- esis deals with the similarity between the membership of interest organizations and the demographic structure of society. We consider the three demographic groups that have received the most attention in the literature on political participation in Exploring the proportionality of representation in interest… their relative levels of individual political capabilities: education, gender, and migra- tion background.2 First, existing studies and academic arguments indicate that supporters of inter- est groups are more highly educated people than the full population. This works, at least, via (1) the values accrued during education, (2) the civic skills acquired, and (3) the relatively higher income levels of higher educated. To start, education has been shown to lead citizens to value democratic norms relatively highly (e.g., Dalton 1994, p. 483). The appreciation of such norms should make it more likely to enjoy the ‘expressive benefits’ associated with being member of a (political) cause group and potentially also other types of associations. Furthermore, education creates sub- stantial opportunities to develop civic skills as part of the education process and in various subsequent (working) environments. These skills include following the news to become aware of ones’ interest and “voicing” skills such as public speak- ing, social media savviness, or letter writing. This should increase the willingness to join interest groups. Last, especially for people with relatively high incomes, ‘joining decisions are below the ‘threshold of (economic) rationality’ in the sense that membership dues are similar to buying luxury products on which buyers hardly make a rational utility analysis (Marsh 1978, p. 270; Jordan and Maloney 1998, p. 390). ‘Groups deliberately market themselves among those best able to afford sup- port’ and this implies that ‘groups have overwhelmingly middle-class member- ship because this is the target audience’ (Jordan and Maloney 1998, pp. 397–398). Some also note a trend towards inequality: “the gains in voice and public leverage have mainly accrued to the top tiers of US society—while Americans who are not wealthy or higher educated now have fewer associations representing their values and interests, and enjoy dwindling opportunities for active participation” (Skocpol 2004, p. 12). These arguments should lead to an overrepresentation of highly edu- cated among the supporters of interest groups. There are contrasting arguments about whether women or men are overrepre- sented in political mobilization through interest groups. Some scholars note that gender differences are fully intertwined with class differences and, given our assess- ment of education, do not merit separate theorization or empirical assessment. This is unsatisfactory because even if ‘the “real” origins of the gender gap in civic activ- ity are rooted, in part, in socioeconomic differences, [that] does not change the fact that women are less politically active – with the consequence that public officials are hearing less from women and their political voice are muted’ (Burns et al. 2001, pp. 47–48). Furthermore, there are non-class-related theoretical arguments to expect interest groups to disproportionally mobilize either female or male members. On the one hand, interest group membership, as a low-threshold, non-conventional type of participation may be relatively attractive to women compared to other channels of political voice. This follows from the relatively unattractiveness of alternative chan- nels, where political institutional practices, for instance, the recruitment criteria of 2 One could think of other sources of inequality in political mobilization, such as age. Yet as we lack reliable data on this demographic related to the membership of the interest groups in our sample, we do not discuss this in this paper.  M. Hanegraaff et al. political parties, work in favor men compared to women. This is in line with Marien et al.’s (2010, p. 193) finding that men are more likely to participate in more con- ventional means of political participation, such as voting and running in elections and that women are more involved in non-conventional methods of political par- ticipation, including engaging in demonstrations, consumer boycotts, petitions, and donations, and plausibly interest group memberships. Furthermore, the marketing practices of major cause groups may effectively create differences in the likelihood of mobilization of men and women, with women more likely to be targeted through direct mailing campaigns and direct recruitment activities (Jordan and Maloney 1998). On the other hand, the socioeconomic status (or low education) may have different political implications for men and women. That is, particular socio-cultural expectations increase the negative effect of low SES for women to a stronger extent than for men (e.g., Burns et al. 2001, pp. 48–49). This ‘double barriers’ argument supported by theories of intersectionality (e.g., Erzeel and Mügge 2017; Strolovitch 2006). In empirical terms, we do not effectively assess such intersections and there- fore necessarily will conflate gender and socioeconomic effects, and, consequently, expect that men, due to their socioeconomic position and the absence of ‘double barriers,’ are more likely to be interest group members than women. Finally, first-generation immigrants commonly do not receive full citizen rights and consequently enjoy fewer opportunities to develop civic skills and knowhow needed to express their interests via associations in their receptive country. This reality leaves them and, through socialization effects, their descendants with rela- tively few political knowhow (Cho 1999). In addition to this legal difference, social reasons underlie the expectation of lower investment of political resources on the part of people with a migration background. Political interests and capabilities are triggered through socialization processes, for instance, in families, sport clubs, schools, work, and other social setting (Janssens and Verweel 2014). These social network interactions and the associated opportunities to mobilize politically tend to occur less frequently for people with a migration background than other individu- als (Morales and Giugni 2011). Therefore, we expect that citizens with a migrant background are less likely to be members of interest associations than other citi- zens and hence are underrepresented in the membership of interest groups. In sum, previous studies suggest that relatively privileged segments in society—in terms of education, gender, and immigration background—are better represented by interest groups. Thus, we hypothesize: H1 “Disproportional membership hypothesis”: Interest groups tend to have mem- bers that are relatively well-educated, men, and not recently migrated. Are policymakers more likely to seek input from groups with an upper‑class accent? Furthermore, the organization of interests does not automatically engender political influence. The access to the policy process determines whether inequalities in inter- est group membership are politically consequential. Various interest group studies Exploring the proportionality of representation in interest… identify the conditionality of access (e.g., Beyers 2004). These conditions include the “encompassingness” of the interests represented (Bouwen 2004), the technical or political nature of information provided (De Bruycker 2016), and the economic power of the constituents of business associations (Klüver 2013). However, none of these studies specifically and empirically examine (1) whether particular char- acteristics of the constituents represented matter or (2) whether particular citizen subgroups, as a result of the limitation of access, benefit or lose from such practices on the part of policymakers. We identify a number of theoretical arguments on this topic below. We explore whether political processes increase the hypothesized (H1) inequal- ity in membership. More precisely, we assess whether policymakers are prone to offer more access to interest groups with politically privileged members (highly edu- cated, men, and native citizens) than to organizations which have a strong member- ship across more marginalized segments of society (lower educated, women, and migrant populations). Studies of interest groups indicate several explanations for policy access, most notably related to group type (business interests or not), strate- gies (inside or outside), and resources (number of lobbyists) (Beyers 2004; Eising 2007; Dür and Mateo 2013, 2016, for an overview see Binderkrantz et al. 2017). These studies hardly differentiate among citizen associations and therefore cannot empirically relate to descriptive nature of the citizens mobilized. We further develop several theoretical arguments made in the literature on access and note that cases can be made that policymakers (1) favor the “underprivileged” and compensate for any pre-existing mobilization inequalities, (2) provide equal access to all who man- age to mobilize, or (3) prioritize the voice of more privileged citizens and exacer- bate existing inequalities in society. For starters, politicians may dampen existing biases in mobilization when grant- ing access to groups. Driven by their desire to be re-elected, politicians may empha- size the preferences of potential, un-mobilized voters in an attempt to gather “slack” votes (Dahl 2005). Therefore, politicians may seek input from interest groups that are as close as possible to these relatively disadvantaged groups. In this manner, politicians can prioritize the voice of the relatively unorganized interests because of their potential electoral relevance (Denzau and Munger 1986). Bureaucrats have been noted to adopt a similar practice under particular circumstances. The best doc- umented case is the European Commission objective to provide access and fund- ing to civil society organizations to enhance the legitimacy of its policy programs (Greenwood 2007; Mahoney and Beckstrand 2011) or with the goal to “balance” pre-existing mobilization bias (Sanchez-Salgado 2014). Second, students of public policy emphasize the independent dynamics of the policy process, which, in aggregate terms, leads policymakers to provide access in proportion to the mobilization of interests. That is, policies are largely made in relatively small policy communities or issue networks, which limit the par- ticipation of outsiders in the policy process due to various procedures (Cobb and Elder 1983; Knoke et al. 1996). Most of the institutional dynamics within such networks such as their path dependency are largely independent from the actual composition of the membership of the active actors. In some cases, these dynam- ics favors “weaker” groups, whereas in other instances, it supports the relatively  M. Hanegraaff et al. privileged. The independent dynamics of policy processes therefore could engen- der the expectation that policymakers reach out to interest groups in a manner that is “reflective” of the actual mobilization of associations. Third, exchange-theoretical views identify incentives on the part of policymak- ers that will make them more likely to increase disproportionality in representation. Although not directly related to citizen representation, a broad literature highlights that more privileged groups gain more access to policymakers, most importantly to business organizations and resourceful groups (Beyers 2004; Eising 2007; Dür and Mateo 2013, 2016). That is, from the perspective of policymakers, we should expect them to grant more access to groups that represent the more privileged members as a rational calculation of the potential preferences of voters. Considering that voter turnout tends to be higher among more privileged groups, re-election chances may be furthered by listening more carefully to their arguments rather than to the conten- tions of those groups that may not turn out to vote at all. A more benevolent variant of this argument refers to the fact that policymakers prefer “high-quality” public pol- icy and therefore focus their attention on the most informed actors, presumably those actors who manage the mobilization of relatively powerful (e.g., Bouwen 2004; also see more pessimistic US studies on money and politics, e.g.: Smith 1995). Please note that this argument is commonly used to explain the access of business inter- ests to the policy process, where the wealth of business allows for the production of high-quality (technical) information or legislative subsidies that can be effectively exchanged for policy access and, plausibly, influence (Hall and Deardorff 2006; see discussion: De Bruycker 2016). Last and distinct from a sociological perspective, we would also expect that people prefer to be in contact with demographically simi- lar people. A number of studies indicate that politicians and civil servants emanate from more privileged segments of society, that is, relatively highly educated, non- migrant, dominantly male, and relatively old (Bovens and Wille 2017). This similar- ity potentially affects the substantive positions of members of parliament in favor of these groups (e.g., Schakel and Hakhverdian 2018; Hakhverdian 2015). A plausible explanation is that policymakers also prefer input from organizations that represent the common interests of such privileged groups. In sum, given an overall lack of research on this issue (in contrast to hypothesis 1), we have three competing hypoth- eses regarding the access different groups may gain to policymakers: H2a “Compensation access hypothesis”: Policy makers more frequently seek input from interest groups that represent lower educated citizens, women, and citizens with a migrant background. H2b “Proportional access hypothesis”: Policy makers seek equal input from inter- est groups that represent citizens of various educational, gender, and migration populations. H2c “Persistence access hypothesis”: Policy makers more frequently seek input from interest groups that represent higher educated citizens, men, and citizens with a non-migrant background. Exploring the proportionality of representation in interest… Both hypotheses relate to ‘representation’ by interest groups. With this, we point at the descriptive composition of the members of interest associations rather than the substantive political positions taken by the leaders or paid staff of associational organizations. Of course, there are several arguments to assume that the substan- tive interests of descriptively different individuals differ and that particular empirical patterns are theoretically plausible. For instance, citizens with high education may attach relatively great value to particular ‘post-material’ causes, men may work in economic sectors with a traditional interest in pay rises, and immigrants may have particular positions on ‘cultural’ issues.3 These views may drive their member- ship choices and form the fertile grounds for the organizational survival of interest groups. We believe a precise assessment of such patterns has merit but think that it must come after an examination of descriptively varying membership composi- tions, as developed in the underlying article. Political preferences are not formed in a sociopolitical vacuum and joint experiences of (potential) members, as critical for long-term survival of collective organizations, plausibly strongly rely on demo- graphic similarities among members of a group. In other words, the demographic characteristics, as theorized in relation to inequalities of political participation, are important in themselves, rather than only in relation to the political views and inter- ests of members.4 Research design Our results are based on the Comparative Interest Group (CIG) survey among leaders of interest associations with citizen members5 in the Netherlands (for the questionnaire, see www.cigsu​rvey.eu). Although the survey has been conducted in multiple countries, questions specifically related to the demographic structure of membership were only asked in the Dutch case, which is therefore the focus of our study. We rely on two sources to identify the population of interest groups in the Netherlands. One of the sources oversamples organizations with high levels of polit- ical involvement whereas the other is likely to also include organizations with lower levels of political involvement (Berkhout et al. 2018; Hanegraaff and Pritoni 2019). This way, we aim to balance our sample (see Online Appendix 1 for a detailed overview of the project). The survey was send to 2316 organizations, we received 3 Our data give credence to some of these patterns. Appendix Table 9 shows the averages per group type (professional, union, identity, public and leisure associations) and, for instance, shows that professional associations tend to indicate to have relatively highly educated members and the public interest groups tend to be supported by a relatively female membership. 4 In the appendix and for illustrative purposes, we assess several broad differences among group types in terms of their membership. 5 The demographic characteristics identified in our RQ pertain to individuals (e.g., a company, as group member, does not have a ‘gender’). At higher levels of conceptual abstraction, one could similarly spec- ulate about relevant ‘demographic’ characteristics related to members of business interest associations (e.g., large / small companies, capital/labor intensive, and so on). This is not the focus of our RQ and cre- ates conceptual incoherence. We therefore excluded non-citizen groups from our analysis even though we did collect such data for business interest associations as well.  M. Hanegraaff et al. responses from 875 (37.8%) and 212 of those had citizens as members. Due to item non-response, our final analyses are based on 197 citizen groups. We believe that our focus on the Netherlands allows us to realize relatively broad external validity. With regard to our mobilization hypothesis, 80% of the Dutch pop- ulation is member of at least one association (Dekker and van den Broek 2005). This top international ranking increases the likelihood of interest representation that is more or less proportional to the distribution of interests in society. In research design terms, this makes the Dutch case, a least likely case to find support for our first hypothesis, and potential supporting evidence will have relatively strong exter- nal validity. We also believe that the Netherlands provides a least likely institutional context for policymakers to prioritize relatively privileged groups. Several bureau- cratic practices and electoral institutions such as formal advisory roles of con- sumer associations in committees, formal “tri-partite” socioeconomic bodies, and the strong proportionality of the party system favor the inclusion of, also, weakly organized but societally relevant groups (e.g., Visser and Hemerijck 1997) (but note recent studies indicating that the political preferences of high-income citizens are somewhat better represented in the Netherlands, e.g., Giger et al. 2012, p. 56; also see: Schakel 2019; Hakhverdian and Schakel 2017). If, as noted in H2, prioritization of privileged groups is found in the Dutch case, similar mechanisms are likely to exist in institutional contexts that are less conducive in this regard. Next, we present the operationalization of our variables (see Table 1). The analy- sis of proportionality in mobilization is based on the answers to the following ques- tion: “Looking at the following social characteristics, could you for each of these characteristics estimate the profile of your membership?” The response categories were provided on 10-point scales of which the endpoints ranged as follows: mostly women versus men, lower educated versus higher educated, and citizens with a migrant background versus citizens without a migrant background. This is an impre- cise measurement, not based on exact, externally validated actual membership lists. We see two potential problems with this: first, organizational leaders may be imper- fectly informed about their members, for instance, because they tend to only inter- act with the most active members only. On this, we do not think that this is plausi- ble given the critical importance of membership for the long-term survival of the organization, and for various day-to-day decisions on marketing, internal communi- cation, and so on. Our respondents are therefore likely to be relatively well informed (as also shown in the very limited item non-response on this question) or, at least, unlikely to systematically over- or underestimate the proportions in particular cat- egories.6 Yet, to be safe, we cross-validated our responses with another data source, 6 Note that we also checked the answers and they seemed mostly intuitively correct to us: for instance, respondents choosing the maximum category in ‘education’ of members include learned societies such as the Royal Netherlands Academy of Arts and Sciences (KNAW) and International Society for In Vitro Methods; in the ‘mostly men’ category, we find representatives of male dominated professions such as the Union of Dutch Veterans (UVV) and Dutch Shipmasters’ Association (NVKK); in the maximum cat- egory on ethnicity (i.e., ‘mostly native Dutch’ (allochtonen)), there are a number of local heritage asso- ciations such as the Committee on Skûtsjesilen (SKS) (Frisian Sailing competition) and the Foundation for Utrecht Castles. Table 1  Overview of variables Variable Survey question Mean SD Min Max H1: Proportional representation Gender Does the organization have mostly female members (1) or male members (10)? 5.75 2.11 1 10 Education Does the organization have mostly lowly educated members (1) or highly educated members (10)? 6.98 2.28 1 10 Migration background Does the organization have mostly members without a migration background (1) or members with a 4.01 2.91 1 10 migration background (10)? H2: Proportional access Access gained Over the last year how often have you been in contact with politicians at their initiative? 2.74 1.18 1 5 Exploring the proportionality of representation in interest… Control variables Political orientation Interest in politics (standardized scale constructed from 3 items: involvement in lobbying, interest repre- −.08 2.49 − 2.62 5.90 sentation and resources spent on advocacy) Specialization How many policy fields does the organizations incidicate to focus on 1.69 2.93 0 11 Professionalization Professionalization of decision-making, hiring policy, training of staff and career prospects of employ-.00 1.66 − 5.15 5.23 ees (standardized) Resources Budget of the organization measured on a 9-point scale 3.79 2.09 1 8  M. Hanegraaff et al. namely the World Value Survey. This survey is held once every few years and has a set of questions related to membership of citizens in interest groups. Moreover, the survey asks respondents about their level of education, their gender, and migration background. Based on this data, we checked whether they matched our results. More specifically, we analyzed whether some demographic groups were over- or under- represented as far as the membership in membership groups compared to their share in the population. The results are presented in Online Appendix 1. Importantly, the results are much in line with our findings. There is a strong underrepresentation of citizens with a lower education; women are slightly underrepresented as members, while there is no underrepresentation based on migration background (see empiri- cal section for more details). Based on these results, we are therefore confident our results are reliable and interest group leaders made, perhaps imprecise, but still valid estimates of their membership. Second, the scales do not refer to concrete proportions of empirical categories, as is the case in the population statistics we use as comparison. This is a source of error as respondents may have different understandings of the meaning of the 10 points on the scale. For instance, a representative of an association with 70% female members may have responded ‘1’, which indicates the endpoint and is labeled ‘mostly’ female members, or may also have used category ‘2’ or ‘3’. This type of error is common in survey research and reduces the reliability of the measure. We do not have theo- retical arguments to assume that this error is non-random and therefore affects the validity of the measure. To compare the survey data with the demographic data, we transformed the scales and assume that our respondents interpret the meaning of the endpoints (the ‘mostly’ categories) to more or less match the lowest/highest catego- ries used in the population statistics. We use the demographic data reported by the Dutch Central Bureau for Statistics (CBS 2016a, b). For each of the three scales, we had to make particular decisions on the cut-off of the categories (and we report the precise distribution of survey answers on the gen- der and migration questions in the Online Appendix 3 and 4). All recoding is based on the distribution of subgroups in society (see Online Appendix 5 for a more elabo- rate description of these choices). To start, with regard to education, we transformed our measures in a five-point scale to resemble the classification of the Dutch Central Bureau for Statistics (identifying citizens with primary, lower secondary, upper sec- ondary, post-secondary, and tertiary education). We group values 1 and 2 as ‘low,’ 3 and 4 as ‘middle-low,’ 5 and 6 as ‘middle,’ 7 and 8 as ‘middle-high,’ and 9 and 10 as ‘high.’ As regards gender and migration, we dichotomize the scale (1–5: more female, 6–10 more male); 1–6: non-migrant, 7–10: migrant). The second part of the analysis deals with proportional representation in access. We rely on the following question in the survey: “Could you indicate how often policymakers reached out to you to provide input?” Respondents could indicate never, once a year, once every three months, once every month, or once every week. Remember that the theoretical argument presented on our second hypothesis heavily relies on the assumed ‘needs’ on the part of policymakers rather than the incentives of interest groups to attempt to gain access (Crepaz et al. 2019). We therefore choose an indicator that emphasizes the access provision by policymakers rather than the attempts made by interest groups. The survey does not ask about ‘having contact.’ Exploring the proportionality of representation in interest… With this, we follow the emphasis on gatekeepers’ control over actual access provi- sion noted in the definition of access formulated by Binderkrantz et al. (2017, p. 307) that ‘access is present when a group has entered a political arena (parliament, administration, or media) passing a threshold controlled by relevant gatekeepers (politicians, civil servants, or journalists). This implies that for access to be present, interest groups need to seek it, and relevant gatekeepers need to allow it.’7 We control for several alternative explanations for policy access. First, we account for “resources,” measured as the annual budget of 2015. The scale is from one to eight, ranging from less than €10,000 to more than €10,000,000 (with increasing step sizes). Second, we control for the professionalization of an organiza- tion by combining items on internal decision-making, hiring policy, training of staff, and career prospects of employees. Third, we control for the political orientation of groups, i.e., whether groups indicate that lobbying is one of the two core tasks of the organization. Fourth, we control for the “policy breadth” (the inverse of specializa- tion) of an organization using the following question: “How involved is your organi- zation in the following policy areas?” For each organization, the number of fields on which it is “very active” is summed up (Berkhout et al. 2019, 2020; Hanegraaff et al. 2020). Finally, policy access is likely to systematically vary across policy areas. We control for such variation on the basis of the answers given to the question on which policy area are you most active? We account for systematic difference between pol- icy areas by using a multilevel model with random intercepts at the level of policy fields. This is a methodological control as we do not theorize about the manner by which particular policy area differences are related to particular membership profiles of groups. 7 This merits discussion of two important choices regarding our operationalization of access: (1) our focus on ‘being contacted’ rather than ‘seeking contact’ and (2) our focus on policymakers rather than other venues or means of access. As regards the first, both types of measures are available in the survey, and we think that it is more likely that the ‘being contacted’ indicator validly reflects the actual access received compared to questions about interest group frequency to seek contact, given that numerous con- tact attempts may not be reciprocated in actual policy access provided in terms of meetings, commit- tee participation, and so on. The survey does not ask about ‘having contact.’ As a robustness check, we examined the ‘seeking contact with policymakers’ indicator in the appendix (see appendix 8). The results are largely in line with the stricter access measure we use here. Second, we focus on (self-reported) direct contact with policymakers rather than participation in parliamentary or bureaucratic commissions, or presence in other arena’s such as the media, which may be measured through actual observation in rel- evant sources. To start, the assessment of interest group access on the basis of, for instance, agendas of parliamentary committees requires the identification and choice regarding the relevant venues to which interest groups may have access. Given the fragmented nature of the policy process, any choice in this regard has limitations related to the particular institutional locus of each venue. We therefore consider that a more general, not institution-specific formulation ‘policymaker,’ which can include politicians, civil servants, and other officials at any level / function of government (‘beleidsmaker’ in Dutch) to more validly reflect the broad policy access of groups than institution-specific indicators. The plausible supe- rior validity of this measure outweighs the potentially somewhat lower reliability of self-reported meas- ures compared to observational data.  M. Hanegraaff et al. 45% Sample 39% 40% Populaon 35% 32% 28% 29% 30% 25% 22% 20% 18% 15% 11% 10% 10% 5% 5% 5% 0% low middle-low middle middle-high high Level of educaon Fig. 2  Frequency distributions of educational attainment in the population and of the CIG survey answers 57.47% 51.43% 49.57% 42.53% Sample Populaon Women Men Fig. 3  Frequency distributions of gender in the population and of the CIG survey answers Results We analyze the survey data using two approaches. First, we evaluate the member- ship hypothesis by comparing the membership profile of the interest groups sur- veyed with the distributions of the categories studied in the actual population (H1). Second, we examine the access hypothesis (H2) through several OLS regression models. In the first part of our analysis, we focus on disproportionality in mobilization. Figures 2, 3, and 4 present the frequency distributions of the answers provided by groups with citizen members in our survey and the distributions of the categories studied in the actual population. Figure 2 compares the distributions among organi- zations and within the population in terms of “education.” The results suggest that the part of the public that is identified by group leaders as their members is higher educated than the population as a whole. To be more precise, 28% of the popu- lation belongs to the upper categories of attained education, whereas 61% of the respondents claim that their membership mostly includes the higher educated (see Exploring the proportionality of representation in interest… 77.90% 69.74% Sample 30.26% Populaon 22.10% Migrant Non-migrant Fig. 4  Frequency distributions of individuals with a migration background in the population and of the CIG survey answers the two upper categories). For the lowest educated citizens, the reverse is true. 33% of the population is considered as low educated, whereas only 10% of the interest group members are regarded as low educated. These findings are in line with ample research that focuses on the political capabilities of citizens with varying levels of education. Moreover, they fit with the proportional inequality observed in parlia- ments toward highly educated citizens (Bovens and Wille 2017; Hakhverdian and Schakel 2017). This result confirms our expectation that highly educated citizens are overrepresented as members of interest groups in the Netherlands. We depict the distributions of “gender” in Fig. 3. The figure clearly illustrates that men are overrepresented in interest group membership as well. In the population, 49.5% are men, whereas among interest group members, the proportion is 57.5%. As the difference is marginal at first sight, the precise distribution is presented in Online Appendix 3 and allows us to provide additional interpretation regarding the lower and upper categories. 65% of the respondents note that their membership is more or less balanced in terms of gender (categories 4 to 7). 33% of the organizations tend to mobilize predominantly men (> 8), whereas 12% have a largely female membership (< 3). In other words, organizations of men are more common than organizations of women. As noted in the theory section, this finding may be related to differences in SES, with men more frequently working in well-organized professions. Never- theless, how this observation can be squared with the finding that women slightly more than men tend to engage in non-conventional means of political participa- tion requires further study into membership motivations (e.g., Marien et al. 2010, p. 193). To conclude, our findings indicate that men are overrepresented in Dutch interest groups. Figure 4 presents the proportions of individuals with a migration background. The results depict a mismatch between the population and interest group member- ship, but in the opposite direction. A relatively large proportion of our respondents indicate that they have members with a migration background (> 7). To put this into more contexts, we therefore specify the upper and lower categories of (non-) migrant membership of interest groups in Online Appendix 3. In detail: 21% of our respondents claim that their membership largely consists of people with a migration  M. Hanegraaff et al. background (> 8), and 60% identify their members as largely non-migrants (< 3); a diversity of membership is noted (4–7) in only a minority of cases (20%). This particular distribution suggests that interest association representing people with a migration background seem to inhabit a separate island in the land of interest groups with separate organizations catering for particular migrant communities. We expected people with a migration background to have relatively few political capa- bilities and therefore somewhat low presence. However, this expectation is not sup- ported by the survey respondents. If anything, the data indicate a (very) slight over- representation of citizens with a migration background as members of Dutch interest groups. In sum, we find mixed support for our first hypothesis that interest groups tend to mobilize relatively privileged citizens in society. This is true for, first, highly educated citizens, and, second, for men. Yet, we do not observe that people with a migration background are underrepresented as members in the Dutch interest group system. Importantly, as highlighted in the RD, these findings are consistent when we use the World Value Studies as indicators for interest group membership. Here, we find a strong overrepresentation in membership for highly educated citizens, a slight overrepresentation of men, and no difference across migration groups as well (see Online Appendix 1). Overall, these results indicate that interest groups indeed ‘sing with an upper-class accent’ in the case of education and gender, yet not in the case of migration background. The question now is: do policymakers mimic this accent when they reach out to lobbyists? In the second part of our analyses, we examine whether these inequalities are exacerbated or reduced in the political access granted to interest groups. We pre- sent several multilevel OLS regressions. Our dependent variable is access granted by policymakers. Our independent variables are the levels to which interest groups represent certain segments in society (see the first part of the analysis). We control for resources, professionalization, lobby organization, and policy breadth. We pre- sent four models in Table 2: one for each demographic type, and a full model. Note, in Online Appendix 6, we provide a robustness check where we include random intercepts per policy area to see whether the effects are strongly driven by varia- tion among the policy areas. The results are entirely the same, however. Because we lose some observations in the multilevel model (almost 15%), we present the simple OLS in the main text. Moreover, in Online Appendix 7, we rely on a less stringent indicator of interest groups access. We hereby rely on a question in the survey where we asked how often groups had been in contact with policymakers at the initiative of the interest group. The results are the same, highlighting the robustness of our findings. Our results provide key insights. First, we assess variation in terms of the edu- cational level of interest group members. We observe no difference in gaining access related to the educational level of members, that is, groups that represent more highly educated constituencies are not invited more often by policymak- ers. The reason may be that the distinct motivations of politicians outlined in the theoretical section balance each other out. That is, interest groups representing citizens of low, high, or mixed levels of education all offer potential benefits to politicians. Nevertheless, perhaps different types of benefits cater to politicians of Exploring the proportionality of representation in interest… Table 2  OLS regression predicting access to policymakers Model 1: education Model 2: gender Model 3: migration Model 4: full Independent variables Education − 0.003 − 0.011 (0.028) (0.028) Gender 0.058** 0.063** (0.031) (0.031) Migration − 0.008 − 0.006 (0.024) (0.024) Controls Political orientation 0.504*** 0.502*** 0.490*** 0.512*** (0.088) (0.087) (0.090) (0.090) Policy breadth 0.015 0.013 0.016 0.015 (0.019) (0.019) (0.019) (0.019) Professionalization 0.075* 0.082** 0.072* 0.081** (0.040) (0.039) (0.040) (0.040) Resources 0.283*** 0.277*** 0.282*** 0.279*** (0.034) (0.034) (0.035) (0.035) Diagnostics Constant 1.116*** 0.797*** 1.074*** 0.792** (0.273) (0.241) (0.223) (0.330) R2 0.41 0.47 0.39 0.48 Observations 195 197 195 193 The model is OLS regression predicting the level of political access citizen groups gain at the initiative of policymakers. Regression coefficients, standard errors (in parentheses), and significance are presented, whereby:*P < 0.1; **P < 0.05; ***P < 0.01 distinct parts of the political spectrum, such as traditional left politicians reach- ing out to relatively low educated and more progressive politicians seeking input from interest groups with mostly higher-educated members. Future studies may delve deeper into this issue. In either case, we find no “bias” of politicians toward groups representing members of varying educational backgrounds. In the model 2, we estimate the effect of “gender” differences. We find a posi- tive and significant effect of male membership on access. That is, organizations with dominantly male membership are more frequently contacted by policymak- ers compared to organizations with a more mixed or a predominantly female membership. To reiterate, we control for policy area effects; hence, this effect is not caused by the particular distribution of groups over policy areas. This would, for instance, be the case if labor market policies disproportionately attract inter- est groups with dominantly male membership, and international development is the focus of groups with a relatively strong female support. Moreover, this effect retains its significance in the full model reported in the last column. Even as we control for socioeconomic background, we find that groups representing women are invited less frequently to the policy process. This suggests that one could  M. Hanegraaff et al. expect that, for instance, the Cabin Crew Association, with a dominantly female membership, compared to the Association of Aviation Technicians, with a domi- nantly male membership, is less frequently approached by policymakers (assum- ing similarity of these associations on all other significant explanatory factors). Finally, we do not detect any differences in access between groups that represent people with a migration background and others (model 3). Overall, we find support for both H2b and H2c, but not for H2a. More precise, we do not observe differences across groups representing higher educated members or individuals with a migration background (in line with H2b), but we do find that groups with a male constituency are more frequently contacted by policymakers than groups with female members (in line with H2c). Importantly, however, the null findings on education and migra- tion indicate a false “neutrality” of policymakers. That is, given the unequal nature of interest group mobilization, the neutrality on the part of policymakers effectively supports the relatively privileged groups in society, offering them the full benefit of their mobilization effort rather than attempting to prioritize the interest groups that represent less privileged subgroups in society. Furthermore, our findings on gender inequality in the policy process raise norma- tive concerns. In this case, policymakers further reinforce the pre-existing inequali- ties in representation in favor of men. This assertion indicates that the policymaking process is “a man’s world,” which is seemingly unwilling or unable to reach out to interest associations with a relatively large female membership. Moreover, we do not believe that this practice is a remnant of institutional traditions from the 1950s; over the past decades, government–group relations have encountered sufficient organiza- tional turnover and volatility to allow for path-dependent patterns to shift. As previ- ously noted, we argue that the Dutch case is to a substantial degree a least likely situation to find this pattern. More specifically, with regard to gender equality in politics and the bureaucracy, the Netherlands does not traditionally have low scores compared to other countries (e.g., having large proportions of female civil servants and legislators, and public opinion in favor of gender equality). These hospitable circumstances to gender equality suggest that this effect is likely to travel to other countries and may even be stronger elsewhere. Conclusion Schattschneider’s depiction of the interest group universe as “the heavenly chorus that sings with an upper-class accent” is arguably one of the most extensively used quotations in the literature on interest groups. Schattschneider (1960, pp. 32, 34) highlighted that disadvantaged individuals in society tend to be underrepresented as members of interest associations and possibly lose out in receiving voice oppor- tunities in the policy process. However, these particular mobilization and access inequalities have hardly been studied empirically. We have attempted to revive this debate in the European context by posing this question: Do interest groups dispro- portionally have members that are sociopolitically relatively privileged? Addition- ally, we assessed whether organizations that rely on relatively underprivileged con- stituents face relatively steep obstacles to have their voice heard politically. Exploring the proportionality of representation in interest… Our findings are mixed. We found that some groups are underrepresented as members of Dutch interest groups. Interest groups have disproportionally highly educated and male citizens as members. This finding signifies that interest organi- zations indeed tend to sing with an upper-class accent in terms of who supports such organizations. As noted earlier, we do not know how this exactly relates to the beneficiaries of the substantive political positions voiced. For instance, broad sections of society may eventually benefit from the advocacy on privacy and data protection by cause groups supported by relatively highly educated citizens. The results on the differences in access offered by policymakers do not exactly match our expectation. The membership profile in terms of education and migra- tion background does not affect the extent to which policymakers reach out to interest groups. This finding suggests that policymakers listen largely to anyone mobilized, whether with an upper-class accent or not. However, by not compen- sating for the disadvantages, policymakers maintain mobilization biases in their provision of policy access. Importantly and in addition, policymakers do prefer to listen to a male choir. Organizations with a dominantly male membership gain more access to policymakers than those organizations that represent more female constituencies. Thus, the bias in mobilization of men is strengthened in the rep- resentation stage, whereas other biases in mobilization are ‘merely’ maintained. We would like to identify a number of limitations of our study and sugges- tions for future research. First, we rely on self-reported membership descriptions. A clear benefit of these data is that they allow us to analyze a broad range of actors and organizations that are active in many different policy areas. However, a downside is that the self-reported membership descriptions on our 10-point scales are not very precise. Although we lack any indication that interest group lead- ers intentionally and systematically misrepresented their membership, potential deviation might still occur as interest group leaders might not be fully aware of the demographic background of their entire membership and, for instance, might report the characteristics of their more active subgroup of members. Despite the daunting task of acquiring such data, the descriptive data and exploratory analy- ses in this paper certainly warrant such exercises. That is, given the large differ- ence we observe, we are fully confident that findings are reflective of biases in the real world. Given the importance of such differences for democratic politics, we hope to set the agenda with this paper to focus more on representational biases among type of citizens in the interest group literature. An obvious next step is the evaluation of the actual membership of interest groups and how this relates to dif- ferent groups in society. Second, we believe that an assessment of cross-cutting membership profiles is likely to be fruitful. How do the dimensions of class, gender, and migration back- ground relate to one another in terms of organizing collectively? And what are the merits or challenges of membership diversity? This also relates to recent concerns about the decline of “cross-class” associations within which citizens from a diverse class-background congregate (Pierson and Skocpol 2007) and the ‘double barriers’ faced by individuals at intersections of class, gender, and migration. Other individ- ual sociopolitical drivers of collective interests merit attention in this regard such as wealth, income, religion, age, living environment (rural/urban), and plausibly  M. Hanegraaff et al. also relation to political identification (party identification or placement in a broader political space such as on a left–right dimension). A third limitation is that we were only able to test our argument in one country. We do expect the results to travel to other contexts because we found substantial “representational biases” across demographic segments in society, even in relatively inconducive circumstances. With regard to our null finding on the relative absence of an “access bias,” although surprising, we do expect the directionality of this phe- nomenon to travel to other contexts. That is, we have theoretical reasons to expect that Dutch associations, compared to those associations in other countries, are sub- ject to selection mechanisms that are favorable to ‘disadvantaged’ groups in soci- ety. This implies that our null finding may partially be ascribed to favorable insti- tutional Dutch circumstances and that in other countries, membership groups with privileged constituents may see their privilege strengthened by selective outreach of policymakers. Nevertheless, further comparative research is necessary to verify these assertions. Last, we see our study as a first step in linking descriptive membership profiles to political preferences of members, the groups’ perceptions of those, and the actual interests successfully or unsuccessfully represented before government. We see a number of theoretical and empirical challenges for such a research program. Con- ceptually, the concepts of descriptive and substantive representation cannot have the same meaning in the field of elected representation and in ‘self-appointed’ repre- sentation (e.g., Montanaro 2017; Rasmussen et al. 2020). Most notably, in group politics, one can’t meaningfully depart from the equality norm of ‘one (wo)man, one vote’ (e.g., Lowery et al. 2015) and therefore need to account for the intensity and proportionality of the interests in society. Furthermore, substantive representative acts by interest organizations such as statements in parliamentary hearings require passing relevant gatekeepers and probably only affect policy outcomes in conjunc- tion with various other political pressures, which are themselves potentially heavily shaped by the same inequalities noted here (e.g., Schakel 2019). Even if there is a sufficient conceptual base that links participation, collective action, and the policy process in a coherent manner, empirically, data are needed on demographic charac- teristics of individuals, their political preferences, their collective organization, and articulation of those substantive political priorities in public policy. This is a major challenge but it would certainly create a better understanding of the translation of societal inequalities into unequal political voice and outcomes. More broadly, the results contribute to the debate on whether some demographic groups (based on ethnicity, education, etc.) are represented proportionately in poli- tics. Interest groups are one channel through which citizens can enter the political arena. We demonstrate that the patterns of bias found in studies on political parties and social movements are similarly present in interest group politics. Thus, interest groups cannot be viewed as a cure for inequalities in political agenda setting and decision-making. Instead, they largely reflect the unequal distribution of political capabilities in society and reproduce these patterns in the political arena. We hope future studies also seek to build bridges between important subfields in the disci- pline to gain an enhanced understanding of why some citizens are better heard in the political process than others. Exploring the proportionality of representation in interest… Compliance with ethical standards Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest. References Berkhout, J. 2013. Why interest organizations do what they do: Assessing the explanatory potential of ‘exchange’ approaches. Interest Groups & Advocacy 2 (2): 227–250. Berkhout, J., J. Beyers, C. Braun, M. Hanegraaff, and D. Lowery. 2018. Making inference across mobi- lisation and influence research: Comparing top-down and bottom-up mapping of interest systems. Political Studies 66 (1): 43–62. Berkhout, J., M. Hanegraaff, and P. Statsch. 2019. Explaining the patterns of contacts between inter- est groups and political parties: Revising the standard model for populist times. Party Politics 1354068819856608. Berkhout, J., M. Hanegraaff, and P. Statsch. 2020. Interest groups in multi-level contexts: European inte- gration as cross-cutting issue in party-interest group contacts. Politics and Governance 8 (1): 61–71. Beyers, J. 2004. Voice and access: Political practices of European interest associations. European Union Politics 5 (2): 211–240. Binderkrantz, A.S., and H.H. Pedersen. 2017. What is access? A discussion of the definition and meas- urement of interest group access. European Political Science 16: 306–321. Bouwen, P. 2004. Exchanging access goods for access: A comparative study of business lobbying in the European Union institutions. European Journal of Political Research 43 (3): 337–369. Bovens, M., and A. Wille. 2017. Diploma democracy: The rise of political meritocracy. Oxford: Oxford University Press. Burns, N., K.L. Schlozman, and S. Verba. 2001. The private roots of public action. Cambridge: Harvard University Press. Central Bureau for Statistics. 2015. Statistics Netherlands: Werkzame beroepsbevolking. Central Bureau for Statistics. 2016a. Statistics Netherlands: Hoogst behaald opleidingsniveau. Central Bureau for Statistics. 2016b. Statistics Netherlands: Population; key figures. Cho, W.K.T. 1999. Naturalization, socialization, participation: Immigrants and (non-) voting. The Jour- nal of Politics 61 (4): 1140–1155. Cobb, R.W., and C.D. Elder. 1983. Participation in American politics: The dynamics of agenda-building. Baltimore: Johns Hopkins University Press. Crepaz, M., M. Hanegraaff, and R.S. Salgado. 2019. A golden key can open any door? Public funding and interest groups’ access. West European Politics 1–25. Dahl, R.A. 2005. Who governs?: Democracy and power in an American city. New Haven: Yale Univer- sity Press. Dalton, R.J. 1994. Communists and democrats: Democratic attitudes in the two Germanies. British Jour- nal of Political Science 24 (4): 469–493. Dalton, R.J. 2017. The participation gap: Social status and political inequality. Oxford: Oxford Univer- sity Press. Denzau, A.T., and M.C. Munger. 1986. Legislators and interest groups: How unorganized interests get represented. The American Political Science Review 89–106. De Bruycker, I. 2016. Pressure and expertise: Explaining the information supply of interest groups in EU legislative lobbying. JCMS: Journal of Common Market Studies 54 (3): 599–616. Dekker, P., and A. van den Broek. 2005. Involvement in voluntary associations in North America and Western Europe: Trends and correlates 1981–2000. Journal of Civil Society 1 (1): 45–59. Dür, A., and G. Mateo. 2013. Gaining access or going public? Interest group strategies in five European countries. European Journal of Political Research 52 (5): 660–686. Dür, A., and G. Mateo. 2016. Insiders versus outsiders: Interest group politics in multilevel Europe. Oxford: Oxford University Press.  M. Hanegraaff et al. Eising, R. 2007. The access of business interests to EU institutions: Towards elite pluralism? Journal of European Public Policy 14 (3): 384–403. Erzeel, S.M., and L.M. Mügge. (eds.) 2016. Special section: Intersectionality in European political sci- ence research. Politics 36 (4): 341–373. Giger, N., J. Rosset, and J. Bernauer. 2012. The poor political representation of the poor in a comparative perspective. Representation 48 (1): 47–61. Greenwood, J. 2007. Organized civil society and democratic legitimacy in the European Union. British Journal of Political Science 333–357. Hakhverdian, A. 2015. Does it matter that most representatives are higher educated? Swiss Political Sci- ence Review 21 (2): 237–245. Hakhverdian, A., and W. Schakel. 2017. Nepparlement?: een pleidooi voor politiek hokjesdenken. Amsterdam: Amsterdam University Press. Hall, R.L., and A.V. Deardorff. 2006. Lobbying as legislative subsidy. American Political Science Review 100 (1): 69–84. Hanegraaff, M., and A. Pritoni. 2019. United in fear: Interest group coalition formation as a weapon of the weak? European Union Politics 20 (2): 198–218. Hanegraaff, M., J. van der Ploeg, and J. Berkhout. 2020. Standing in a crowded room: Exploring the rela- tion between interest group system density and access to policymakers. Political Research Quarterly 73 (1): 51–64. Janssens, J., and P. Verweel. 2014. The significance of sports clubs within multicultural society. On the accumulation of social capital by migrants in culturally “mixed” and “separate” sports clubs. Euro- pean Journal for Sport and Society 11 (1): 35–58. Jordan, G., and W. Maloney. 2006. “Letting George do it”: Does Olson explain low levels of participa- tion? Journal of Elections, Public Opinion and Parties 16 (2): 115–139. Jordan, G., and W.A. Maloney. 1998. Manipulating membership: Supply-side influences on group size. British Journal of Political Science 28 (2): 389–409. Junk, W.M., J. Romeijn, and A. Rasmussen. 2020. Is this a men’s world? On the need to study descriptive representation of women in lobbying and policy advocacy. Journal of European Public Policy. https​ ://doi.org/10.1080/13501​763.2020.17671​79. Klüver, H. 2013. Lobbying in the European Union: Interest groups, lobbying coalitions, and policy change. Oxford: Oxford University Press. Knoke, D., F.U. Pappi, J. Broadbent, and Y. Tsujinaka. 1996. Comparing policy networks: Labor politics in the U.S., Germany, and Japan. Cambridge: Cambridge University Press. Leech, B.L., F.R. Baumgartner, T.M. La Pira, and N.A. Semanko. 2005. Drawing lobbyists to Washing- ton: Government activity and the demand for advocacy. Political Research Quarterly 58 (1): 19–30. Mahoney, C., and M.J. Beckstrand. 2011. Following the money: European Union funding of civil society organizations. JCMS: Journal of Common Market Studies 49 (6): 1339–1361. Marsh, D. 1978. More on joining interest groups. British Journal of Political Science 8 (3): 380–384. Marien, S., M. Hooghe, and E. Quintelier. 2010. Inequalities in non-institutionalised forms of political participation: A multi-level analysis of 25 countries. Political Studies 58 (1): 187–213. Morales, L., and M. Giugni. 2011. Social capital, political participation and migration in Europe: Mak- ing multicultural democracy work? Basingstoke: Palgrave. Pierson, P., and T. Skocpol. 2007. The transformation of american politics: Activist government and the rise of conservatism. Princeton: Princeton University Press. Putnam, R.D. 2000. Bowling alone: America’s declining social capital. New York: Palgrave Macmillan. Rasmussen, A., and S. Reher. 2019. Civil society engagement and policy representation in Europe. Com- parative Political Studies. https​://doi.org/10.1177/00104​14019​83072​4. Salgado, R.S. 2014. Europeanizing civil society: How the EU shapes civil society organizations. Hound- mills: Palgrave Macmillan. Salisbury, R.H. 1969. An exchange theory of interest groups. Midwest Journal of Political Science 13 (1): 1–32. Schakel, W. 2019. Unequal policy responsiveness in the Netherlands. Socio-Economic Review. https​:// doi.org/10.1093/ser/mwz01​8. Schakel, W., and A. Hakhverdian. 2018. Ideological congruence and socio-economic inequality. Euro- pean Political Science Review 10 (3): 441–465. Schattschneider, E. 1960. The semi-sovereign people: A realist’s view of American democracy. New York: Holt, Rinehart and Winston. Exploring the proportionality of representation in interest… Schlozman, K.L., S. Verba, and H.E. Brady. 2013. The unheavenly chorus: Unequal political voice and the broken promise of American democracy. Princeton: Princeton University Press. Skocpol, T. 2004. Voice and inequality: The transformation of American civic democracy. Perspectives on Politics 2 (1): 3–20. Smets, K., and C. van Ham. 2013. The embarrassment of riches? A meta-analysis of individual-level research on voter turnout. Electoral Studies 32 (2): 344–359. Smith, R.A. 1995. Interest group influence in the US Congress. Legislative Studies Quarterly 89–139. Schmitter, P.C., and W. Streeck. 1999. The organization of business interests: Studying the associative action of business in advanced industrial societies. MPIfG Discussion Chapter, 99(1). Strolovitch, D.Z. 2006. Do interest groups represent the disadvantaged? Advocacy at the intersections of race, class, and gender. The Journal of Politics 68 (4): 894–910. Truman, D.B. 1951. The governmental process. New York: Alfred A. Knopf. Theocharis, Y., and J.W. van Deth. 2018. The continuous expansion of citizen participation: A new tax- onomy. European Political Science Review 10 (1): 139–163. Verba, S., K.L. Schlozman, and H.E. Brady. 1995. Voice and equality: Civic voluntarism in American politics. Abridged. Cambridge: Harvard University Press. Visser, J., and A. Hemerijck. 1997. “A Dutch miracle”: Job growth, welfare reform and corporatism in the Netherlands. Amsterdam: Amsterdam University Press. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Marcel Hanegraaff is Associate Professor in Political Science at the University of Amsterdam. He researches the politics of interest representation in a transnational and EU context, as well as on the func- tioning of international organizations in the fields of climate change and global trade. Joost Berkhout is Associate Professor in the Department of Political Science at the University of Amster- dam. He is specialized in research on interest organizations. Jens van der Ploeg was working at the Departments of Political Science of the University of Copenha- gen and the University of Amsterdam. He is currently working as a researcher at the Netherlands Court of Audit, where he studies policies related to agriculture, environmental affairs, economic affairs and infrastructure.