Abstract
To what extent is citizen political participation, such as electronic or personal contact with members of Congress, stimulated by membership in organized interest groups? I use data from a nationwide survey conducted by Zogby in 2007 to assess the extent to which Americans contact congressional offices, and whether membership in more activist-oriented groups, such as citizen groups, stimulates greater rates of contact than membership in professional associations or no group membership at all. I also examine whether this group ‘effect’ on participation breaks down by the method used, low-effort electronic contact (mail, email, web-based contact pages, on-line petitions) versus high-effort contact such as personal meetings with lawmakers. I find that the role played by interest groups in facilitating communication can be substantial. In the case of members from lower socio-economic backgrounds in particular, membership in a citizen group helps compensate for lack of knowledge and resources regarding how to contact Congress.
Introduction
Representation requires citizens to actively communicate their wishes and concerns to those elected to speak for them, to keep tabs on those officials to make sure they follow through, and to protest when they do not. Yet political participation in the United States has been found to be declining through much of the twentieth century, except when it comes to interest group membership. The size and ideological diversity of the American interest group system has grown dramatically over the past 40 years and brought about substantial changes in how political organizations seek to mobilize bias on behalf of their members (Baumgartner and Leech, 2001; Walker, 1991). Counts vary, but it is generally believed that groups exist today by the tens of thousands, articulating a wide range of citizen demands backed by an arsenal of pressure tactics.
But is greater group mobilization really resulting in more political participation by citizens who might not otherwise participate? More precisely, do people who join interest groups, and participate in politics as a result of this membership, differ from non-members by socio-economic status (SES), civic skills, or a simple inclination towards political action? If there is a group stimulating effect on participation, does it vary by the types of incentives groups use to entice individuals to join in the first place? And does familiarity with the Internet make members more or less likely to join and participate? I explore these questions in this article with data from a national survey. The results suggest that there actually is a group ‘effect’ on participation, at least when it comes to contacting members of the U.S. Congress, although factors such as SES and familiarity with communicating through the Internet do matter indirectly. Indeed, I find that lack of socio-economic status and, to a more qualified extent, Internet familiarity can be compensated for by joining a group, especially a citizen group. In other words, interest groups appear to be having something of a small levelling effect on citizen participation in the United States.
Political participation and interest groups
Participation is arguably the foundation of effective representation, ensuring that those who govern articulate the policy preferences of the governed. Presumably this requires citizens to follow events in capitols, weigh-in on which issues they feel government should debate, suggest and promote alternative agendas and policies, and, of course, communicate all of this information to their elected officials. The extent to which citizens of the United States actually do any of these things has consumed a significant amount of scholarly energy, and the conclusion reached regarding many basic forms of participation, such as voting, is pessimistic – too many citizens do very little of it (e.g., Bennett, 1986; Campbell et al., 1960; W. Miller, 1980). Individuals who are disproportionately better educated, wealthier, Caucasian and, very often, older, appear to participate more than others, though Rosenstone and Hansen (1993) still find that overall citizen participation, including contributing, contacting, voting, protesting and even feeling politically efficacious, has steadily declined.
Evaluations of participation through voluntary civic organizations have not turned out to be any brighter (Putnam, 2001; Skocpol, 2004), with one exception – membership in political interest groups appears to be growing. Census counts vary depending on each researcher’s definition of ‘interest group’ (see Baumgartner and Leech, 1998: 23; Knoke, 1986), but it is accepted that there are far more of them now than a century ago, sporting more organizational forms and articulating a broader range of political demands (Baumgartner and Leech, 2001). Some are supported more by wealthy philanthropists than individuals, but there is evidence that more people are joining interest groups than once thought (Baumgartner and Walker, 1988). But do more groups mean more participation? Are members more likely to participate in politics than non-members and, if so, is it because groups stimulate it, or were they likely to do so anyway?
Certainly many group leaders claim they are mobilizing more people for political action (Fowler and Shaiko, 1987; McFarland, 1984). Some types of groups, such as the more activist-oriented citizen groups, appear to be quite good at bringing individuals from latent social interests into the public arena by encouraging them to join protests, write letters and engage in other forms of grassroots advocacy (Berry, 1999; Walker, 1991: ch. 2). Yet there is little research on whether these members are really participating more than non-members and why (or why not), even in something as simple as contacting elected officials. Goldstein (1999) finds that groups can stimulate quick bursts of constituent contact in a few swing legislative districts on a few, high profile issues like healthcare reform, but not whether group membership generally leads to an increase in political contacting. Some scholars argue that membership often amounts to little more than just paying dues, and this ‘cheque-book participation’ may actually make members less inclined to participate in any other way (Godwin, 1988; Wilson, 1973).
One of the first to tackle these questions of interest group member political participation was Pollock (1982), who linked it to why individuals joined a group in the first place. Olson (1965), of course, argued that people join simply to get valuable private benefits and are content to leave politics to the group’s leaders, but research showed years ago that people often join because they see groups as ways to channel their passions into effective political action (Knoke, 1988; Moe, 1980). Some groups, often citizen groups, cater to these desires, offering activism opportunities to people otherwise unable to do so (Berry, 1999). Presumably they will participate more frequently than people in material benefit-emphasizing, work-related groups or groups promoting social networking opportunities, and certainly more than people in no group at all.
Pollock also tried connecting socio-economic status (SES) characteristics, such as education and income, to membership, and hence political participation. This is consistent with Verba and Nie’s (1972) finding that higher SES individuals were more likely to join local voluntary organizations and there develop the civic skills necessary for political participation, such as fundraising, public speaking and writing. Yet when he combined group membership incentives and member characteristics into his model, the results were surprising. Higher SES individuals were more likely to join social networking groups and vote, but while members of groups stressing activism were more likely to be involved in electoral and grassroots campaigns, neither membership nor participation were connected to SES. Pollock (1982) found that membership increases participation, but the murky role of SES leaves equally murky the question as to why.
Why purposive interest groups – the type that ought to exhibit a connection between SES, membership and participation – do not might be explained by looking more closely at the work on voluntary organizations. Brady et al. (1995), for instance, find that the connection between SES and participation is simply not that straightforward. Higher SES is positively associated with education and income, which, in turn, inclines individuals towards greater participation. Membership in voluntary organizations like churches and local non-profits, however, is largely unconnected to SES, suggesting that the civic skills learned there provide an alternative route towards participation independent of SES. Thus, it may be true that by joining such organizations lower SES citizens not only learn civic skills, they also enter the social networks of activists who recruit them for politically oriented groups, something Brady et al. suggest in a later article (1999). 1 They may be just as likely to join a group and thus participate in politics as high SES citizens. 2 Interest group membership may actually compensate for low SES.
Brady et al.’s findings may also provide insight into another odd finding in Pollock (1982) that can help us understand group participation. Membership in purposive groups, Pollock finds, does not increase a person’s interest in, or inclination towards, political action, but that membership in more social groups does. If higher SES leads to membership in voluntary non-political groups, as Brady et al. find, where civic skills are learned, then both SES and voluntary group membership might incline people to develop a greater inclination towards politics which then, in turn, might encourage them to join a purposive group and participate. Purposive citizen groups thus focus and direct political interest rather than create it, and SES has an indirect effect on joining while having a direct effect on participation independent of group membership.
As for work-based groups, such as trade and professional associations, Ayala (2000) finds that membership here does not build civic skills, and Olson argues that members seeking material benefits (typical of this group type he argues) are not politically inclined, so we should see no connection between SES, inclination and membership in these groups or any group-inspired or directed participation by their members. Leighley (1996) finds something like this, although she does not look at specific types of political participation such as contacting Congress.
To learn whether and how interest groups stimulate member participation I need to develop a model that tests the connection of SES and other resources, such as membership in non-political voluntary organizations, to individual political inclination that is independent of, and prior to, interest group membership. I then must test for a relationship between inclination and membership by the types of incentives used to entice individuals to join. Finally, I must use these relationships to predict the participation rates of purposive group members as compared to the members of other group types, as well as individuals not in any group at all.
Modelling political contacting through interest groups
Even this brief review shows how complex modelling citizen participation is. Focusing only on contacting elected officials in the U.S. Congress should make this research more tractable, as should drawing a distinction between an individual’s desire to make contact and his or her ability to do so. Desire essentially is political inclination – an individual wanting to express his or her views to their elected officials in Washington, D.C. and to join purposive interest groups that will make contacting easier. Based on the above research, I lay out in Figure 1 a system of hypothesized relationships showing how political inclination may be influenced by SES variables such as education and income (giving one a greater sense of a stake in a society shaped by public policy) as well as, and independent of, other resource variables such as membership in non-political organizations like churches and social clubs that teach civic skills.
Hypothesized relationships among variables influencing group membership and political contact.
Ability to contact Congress and use of the Internet
Ability is the individual ‘cost’ of participation, or the effort required to contact U.S. senators and representatives, which, in turn, influences the method of contact employed. While high-effort methods such as travelling to the Capitol to meet legislators is time-consuming and expensive, some may even find low-effort methods such as emailing or using Internet contact pages a little challenging. Even knowing who to contact can be a problem. More and more political contacting is now done through the Internet (Fitch and Goldschmidt, 2005), whether commenting on agency rules, requesting documents or simply communicating with Congress. The Internet may be a revolution in communication between constituents and representatives, but it requires the former to have convenient access to terminals and to know a fair amount about how to find and use the latters’ webpages (in turn shaped by website quality; see Burden, 2007). It requires money and education, which is why SES connects to ‘Internet familiarity’ in Figure 1. The Pew Internet and American Life Project also found that those who regularly use the Internet are more likely to vote than those who use it less (Rainie, 2007), so it makes sense to see if Internet familiarity influences other forms of participation such as contacting Congress.
Interest groups may compensate for members’ lack of SES resources, including the education necessary to effectively use the Internet. For decades advocacy groups have tried to make it as easy as possible for their members to contact elected officials to promote group interests. In the 1980s, groups began utilizing direct mail services, providing their members with form-letters or identical postcards they could sign and send to legislators. This resulted in a dramatic increase in the sheer volume of mail flooding congressional offices before key votes (Godwin, 1988; Sabato, 1984), what some called ‘astro-turf advocacy’. With the emergence of the Internet, paper letters have given way to electronic ‘letters’, with consultants creating websites to allow lobbyists to quickly send ‘alerts’ to members and then make it easy for those members to send email, generic or personalized, to Capitol Hill (Fitch and Goldschmidt, 2005). For those who want to travel to Washington, groups arrange places to stay and make appointments.
Of course group leaders do not want to level the playing field for everyone; they are not patriotically trying to strengthen ties generally between government and the public. They only seek to facilitate a connection between their members, the individuals who (usually) pay to belong and the lawmakers they believe are likely to advance their members’ interests. Since they can easily connect members to congressional contact sites, it is little wonder why even low SES individuals who become interested in participation would flock to interest groups as the political arena becomes increasingly congested with new groups and issues, webpages and blogs. It also means there should be an observable group ‘effect’, members contacting lawmakers at higher rates than non-members, perhaps even more than politically inclined non-members who have the desire but may lack the ability.
Political inclination and SES
As Pollock (1982) found, this effect may vary from group to group by the incentives they use to entice members to join. 3 If citizen groups tend to attract individuals who are already politically inclined by offering opportunities for activism (Berry, 1999), and if this inclination increases when civic skills are acquired in non-political organizations as well as when SES is higher (Brady et al., 1995), then there should be no direct connection between SES, voluntary group membership and citizen group membership. Instead, greater inclination should directly increase the likelihood of membership in political groups, and inclination should then increase as SES and involvement in non-political groups increases.
Furthermore, when it comes to contacting Congress, Berry (1999) finds that citizen groups often simply direct their already eager members to engage in widespread, high volume grassroots advocacy. Yet while they might prefer to stage protests on Capitol Hill, many citizen groups lack the financial resources necessary for sustained protests (Schlozman and Tierney (1986: 115), although Cigler and Nownes (1995) find that some are quite wealthy). Some individual members may be financially well off, but only the most committed are likely to want to pay the cost of taking weeks off from work for a long protest. Instead, what these groups provide, along with direction, are form letters, petitions and easy-to-use websites for sending email to online congressional accounts and webpages.
In contrast, professional and trade associations primarily recruit members by offering non-political material incentives, so here we would not expect to see a political inclination variable positively influencing membership. Many of these groups represent professions requiring at least a junior college education, so this SES variable should be positively associated with membership, as shown in Figure 1. Since many, perhaps most, of their members are not overly inclined towards politics (not even after they join one, Ayala (2000) finds), these groups should be less likely to push their members towards low-effort tactics, or any kind of contact at all. Yet these groups tend to be wealthier (Berry, 1999), or at least have wealthier members, so members who happen to be highly politically inclined might be inclined enough to travel to Capitol Hill to meet their elected officials, even if just on an annual ‘lobby day’.
Research design
The data I use to explore the influence of interest group membership on participation laid out in Figure 1 come from a survey of the population of the United States conducted in the fall of 2007 by Zogby International for the Congressional Management Foundation’s study of citizen communication with Congress. 4 A random sample of 1,071 voting-age individuals was successfully contacted by telephone from national directories for an approximately 3.1 percent margin of sampling error. The results are weighted by a variety of characteristics to better reflect the distribution of the population and account for non-responses. 5 Although some of the variables used in my analyses are discussed in the next section, the main ones are described here.
The first is a scale measuring a person’s inclination towards, or desire for, political action that, following Figure 1, will be statistically estimated using a set of explanatory variables capturing socio-economic status characteristics along with other factors. As the components of this scale must be independent of SES factors, and Brady et al. (1995) and Pollock (1982) argue that psychological factors are nearly impossible to measure, I focus only on reported political behaviour unconnected to (and presumably independent of) group membership. The survey asked five questions capturing this behaviour, and for every positive response I added 1 to a respondent’s scale score, scoring them 0 if they did not respond positively to any question. The first asked how likely they were to vote in the 2008 presidential election, with a 1 added if the response was ‘somewhat likely’ or ‘very likely’. The second regarded their party affiliation, a 1 added if they indicated that they were members of any political party. The third was whether they had ‘attended a political protest, speech, or rally’ in the last five years. The fourth was whether they had ‘volunteered for or given money to a political campaign’ in the last five years, and finally whether they had ‘volunteered for or given money to an advocacy campaign’ in the last five years. 6 Only 60 respondents, or 6 percent, scored 0, but only 64, or another 6 percent, scored a maximum of 5, while the mean was 2.43 and standard deviation was 1.2. 7
I used a similar approach to construct an index of Internet familiarity, with questions capturing how frequently individuals used the Internet for a variety of purposes. For this ‘ability’ measure, seven questions were asked (the wording and answers are presented in Appendix A) and the scale construction was done in the same way as the political inclination index. The resulting index ranges from 0 to only 6 because no respondent’s score reached the maximum possible of 7, and 14 percent of respondents have 0 scores. The mean was 2.13 and standard deviation was 1.3. As Figure 1 predicts, Internet familiarity correlates with SES measures of education (r = 0.41) and income (r = 0.40) (see below for how these were measured).
Another important variable, of course, is whether the observed citizen actually paid money to join either a citizen group or professional/trade association. A ‘citizen group’ in this case is defined as an organization that is first and foremost formed to lobby for change in public policy, or maintaining an existing status quo, and has an open membership allowing anybody to join (Berry, 1999). Although I assume that individuals who join these types of groups are primarily attracted by purposive incentives, and are more or less aware and supportive of the group’s policy goals, I cannot measure these incentives per se. Instead, because its incentive structure is designed to appeal to politically passionate people, I expect the inclination index to have a greater effect on joining this type of group than occupation-based, closed-membership, material-incentive offering professional and trade groups. Many of the latter type may have their headquarters in Washington, D.C. (Schlozman and Tierney, 1986: 67), but their purpose is to represent their members on a wide range of matters, not just in the political area (King and Walker, 1991).
Survey respondents were asked: ‘In the past five years have you joined or renewed your membership in a professional association?’ and a dummy variable was coded 1 if they had. Similarly, they were asked: ‘In the past five years have you joined or renewed your membership in an interest group?’ and a dummy coded 1 if they had. Peaking ahead a bit, the mean of the inclination index for citizen group members is 2.56, only a little higher than the overall mean of 2.43. It is 2.33 for trade group members, suggesting that they are only a little less interested in politics. More interesting is that it is 3.22 for those who happen to be in both groups and 1.92 for those in no group at all.
Granted, this second variable is more ambiguous given the vague nature of the term ‘interest group,’ even with the next question asking about membership in a professional association clarifying that the first primarily meant an advocacy group (most citizens are probably not familiar with the term ‘citizen group’). To help confirm the variable’s validity, I correlated it with another question asking: ‘In the past five years have you volunteered for or given money to an advocacy campaign?’ The coefficient was a strong 0.62, and only 0.16 with the professional group dummy. As it turns out, 43 percent of respondents belonged to a citizen group, 30 percent to a professional or trade association, and 18 percent to both. Also, while individuals responding to the ‘interest group’ question may have distinguished it from trade and professional associations, and were not members of intergovernmental groups (for which they were not eligible) or public interest groups (which do not have members), it is possible they confused the category with labor membership. Unions do engage in political advocacy, and while membership is often compulsory rather than optional, individuals may still have had them in mind when they answered the question. Respondents were asked in a separate part of the survey: ‘Are you a member of a union?’ but out of the 188 who were, only 96 also answered the ‘interest group’ question positively. Since I do not know whether union members confused their answers, and this is a small number of respondents anyway, I reduce the effect of any ambiguity in the analysis by entering the union membership dummy as a control variable.
My key dependent variable is individual contact with senators and representatives. ‘Contact’ encompasses a variety of methods ranging from low-effort options, such as signing petitions, clicking on websites and sending form-letters, to high-effort options, such as travelling to Washington, D.C. to personally meet with elected officials and/or their staff. In order to differentiate these, I created two binary variables. A survey question asked: ‘What method did you most recently use to contact a U.S. senator or representative?’ and a menu of choices given, though each respondent was only allowed one response (see Appendix B for a complete list of questions and response rates). All individuals responding with any method except ‘no contact’, ‘in person meeting’, ‘public meeting’, ‘other’, and ‘not sure’ were coded 1 for a low-effort contact variable with 412 or 38 percent of respondents so coded. 8 The 38 (or 4 percent) indicating they met privately or publicly with legislators were coded 1 for the high-effort dummy. I find that 201 contacts were made by email, websites or online petitions (19 percent of all respondents and 49 percent of those using low-effort contacts), while 198 used more ‘traditional’ methods.
Data analysis and discussion
Estimating political inclination as a partial function of SES
Again, searching for a group effect on contacting is the primary research question, but it must be done in the context of who joins these groups and why. As outlined in Figure 1, an individual’s attributes, including SES, should lead to a greater interest in politics that, in turn, should make him or her more likely to join a citizen group emphasizing opportunities for political action rather than (or in addition to) joining a trade or professional association or not joining any group at all. Yet even apart from fostering an interest in politics, some SES factors, like education, should also make it more likely that individuals will join a professional association as part of their job. To handle this endogeneity and capture the direct and indirect effects of these variables on joining, I create an instrumental variable standing in for political inclination in the joining models that is a function of SES and other variables. 9 Political inclination is therefore estimated twice, once in a system of equations using this instrumental variable to then estimate membership in a citizen group as a function of inclination and other explanatory variables, and then again to estimate membership in a professional or trade group.
The first socio-economic variable I regress on the inclination index is a respondent’s education level. Each was asked whether he or she had some high school education but no degree (coded 1), a high school but no baccalaureate degree (2), completed a four-year college degree (3) or holds a graduate level degree (4). Whether an individual is in a group historically discriminated against is a second measure of SES, so respondents were asked to identify their ethnic groups and a dummy variable was coded 1 for an ethnicity other than Caucasian, as well as a dummy coded 1 if the respondent was female. Income is another SES component, but a question asked regarding various brackets of income yielded data correlating with education at r = 0.46 and is unusable. I instead used other financial data reflecting assets upper income individuals are more likely to possess, such as 401k accounts, captured by a dummy coded 1, and a second retirement fund, also coded as a dummy.
Beyond these SES measures, I used other variables that may also influence an individual’s desire to participate. Citizens with more extreme ideological views might be more politically active (just as they are often more likely to vote). A question asked respondents to indicate whether they were ‘progressive’, ‘liberal’, ‘moderate’, ‘conservative’ or ‘very conservative’. A dummy was coded 1 if they indicated that they were progressive or very conservative. Based on Brady et al. (1995), Figure 1 shows that participation in non-political organizations, such as social clubs and churches, instill in individuals a greater interest in politics as they acquire civic skills, so a dummy variable is coded 1 if the respondent indicated that he or she has joined or renewed membership in a social club in the last five years. Another dummy was coded 1 if they indicated that they regularly attended a religious institution. Individuals with a family member serving in the military might also be more inclined towards politics since political leaders dictate the broad directions of military policy, so a dummy is coded 1 if the respondent indicated he or she is in such a family. Finally, to make sure that the variables hypothesized to only influence joining a group are not also influencing inclination (and biasing standard errors), I include them in the estimation of the index as well. However, with one exception they do not significantly influence the index, so I hold off on discussing them for the moment although the results for all of these variables on inclination are presented in Table 1.
Estimates of political inclination. Maximum likelihood estimates (robust standard errors).
*p < 0.05; **p < 0.01; ***p < 0.005.
In both estimations the SES variables of education, gender (in one model) and additional income all contribute to a greater interest in politics, which I anticipate will lead to a greater desire to join interest groups and contact elected officials. Hardly meant to be the last word regarding social impetuses towards politics, the results suggest some social stratification of political interest, although like Brady et al. (1995) I find that membership in non-political organizations like civic clubs and religious institutions also instills an interest in politics that may compensate for low SES. Internet familiarity also contributes to political inclination, which is consistent with Rainie’s (2007) finding that Internet users vote more frequently.
Citizen and trade and professional group membership
The next stage of the analysis uses this inclination instrumental variable capturing a respondent’s desire to participate in politics to estimate the likelihood of joining the two different types of interest groups. The other variables potentially influencing this choice are assumed to have effects independent of inclination and are grouped by qualification, opportunity/location and available time. This does not mean variables that were components of the index should not be included here as well; Figure 1 suggests that education is often a qualification for jobs likely leading to professional and trade group membership and is entered here (again, this is why it is necessary to have an instrumental variable for political inclination).
Citizen groups offer opportunities for individuals to express their passions for political causes, but other emotion-driven organizations may also provide outlets for these passions, giving individuals less need to join political groups. Schlozman et al. (1995), for instance, argue that churches provide opportunities for individuals to pursue activities leading to social gratification that they might otherwise find in political groups. Religious attendance was a positive influence on political inclination and it may replace citizen groups as opportunities for engaging in activism; it is therefore included here. Living in a metropolitan area may also provide social and purposive opportunities depressing the attractiveness of the purposive incentives offered by citizen groups, so respondents were asked whether they lived in a city, town or rural area and I coded a dummy as 1 for living in a city and another dummy 1 for a town, while leaving rural areas as the excluded category. 1
Lack of time due to personal circumstances may also make individuals less likely to join, even if they are politically inclined. In this case I measure whether respondents were single parents with a binary variable coded 1 if a respondent indicated in questions that they had children (age 17 or less) but considered themselves ‘single’. Although I do not predict whether it influences interest group membership, because it may reduce one’s need for a group to help contact Congress, I include it here to see if it has a depressing effect on membership. Finally, respondents in the survey could be members of both types of groups at the same time, and being a member of one might influence their decision also to be in the other type, so a dummy was coded 1 if the respondent was in both and this was used as a control.
The results are given in Table 2. As predicted in Figure 1, political inclination is positively associated with joining a citizen group. Of course, since the citizen group variable was also found to increase inclination (see Table 1), I cannot claim that the entire value-added of group membership is in terms of making contact easier for members, the ability side of encouraging contact with Congress. Exposure to more political information and the intense passions of others from being in the group has a feedback effect leading to an even greater desire to participate. Surprisingly, inclination is also associated with joining trade and professional groups. I do find this latter influence to be weaker when I calculate the likelihood of joining each group type with all other variables held at mean or modal values. Increasing inclination by one standard deviation increases the likelihood of joining a citizen group by 18 percentage points more than for joining a trade group. The politically inclined are more attracted to activist citizen groups than material-incentive offering work-related groups.
Estimates of choosing to join professional and citizen groups. Maximum likelihood estimates (robust standard errors).
*p < 0.05; **p < 0.01; ***p < 0.005.
As for the other explanatory variables, involvement in a religious institution does appear to lessen the need for involvement in citizen groups, as Schlozman et al. found. The same appears true for respondents living in urban areas where there are many opportunities for fulfilling needs for social and political involvement. This is not the case for professional and trade groups. Indeed, apart from political inclination, only education and citizen group membership are significant factors when it comes to joining work-related interest groups, and the latter is especially interesting because the coefficient is negative. Perhaps because citizen groups often oppose business groups in political fights, or because many career-driven individuals see them as a frivolous waste of time and money, individuals participating in professional groups are less likely to also be in overtly activist groups, even if they are politically inclined. As for Internet familiarity, the results reported here suggest that the lack of a connection in Figure 1 is justified because the Internet familiarity index fails to exhibit a significant effect in terms of joining citizen or trade groups. Internet-savvy people may not need a group’s help to be active, but membership may still compensate for a lack of Internet familiarity.
The interest group effect on contacting members of Congress
Now I tackle the primary research question: Does membership in either of these two types of organized political interests increase the likelihood that members will contact congressional offices on the group’s behalf? As described above, contact is divided into low-effort, or writing letters, sending email, signing petitions and so forth, and high-effort, which includes travelling to Washington, D.C. to meet in person with representatives or senators. Binary variables indicating whether the respondent was a member of a citizen group and/or a professional and trade group are used as regressors on both the low and high-effort contact dependent variables. Political inclination is again included as an instrumental variable estimated in the first step of both of these two-step models (all exogenous variables from both stages are included in both models).
In addition to inclination, another desire-to-contact variable I use is education. However, the maximum likelihood algorithm cannot converge with the original education variable included for both contacting and inclination, so for the contacting Congress stage I created a new binary variable coded 1 if the respondent has a four-year college degree. Individuals might also be inclined to make contact if they had already received a letter or email from their member of Congress, perhaps a newsletter or blanket form letter that senators and representatives often send. Respondents were asked whether they had ever received such un-solicited communications, and a dummy variable coded 1 for the 67 percent who reported that they had (only 26 percent reporting that they had received an un-solicited email). 2
Other explanatory variables attempt to capture respondents’ ability to contact Congress with low and/or high-effort tactics. The Internet familiarity index is a crucial ability variable here, but I include others as well. Internet access may still be difficult in some rural areas and political information possibly harder to come by, so I code a dummy 1 if the respondent indicated that they lived in a rural region (although this does not turn out to be supported). Although very little money is required to write a letter or send an email to Congress, it takes time and financial resources to meet in person with a legislator, or even their Washington staff (a possible indirect SES effect). Since the education variable is still included, even in a modified form, I continue using the two substitute income measures used in the other models to avoid multi-collinearity. Finally, 18 percent of respondents were in both citizen and work-related interest groups, which may confound the results and should be controlled for. I do this by simply coding a dummy 1 if the respondent indicated that he or she was a member of both group types.
Only the results of the estimates of the low-effort tactics contacting stage are reported in the second column of Table 3; the estimates for inclination are the same as in Table 1 in terms of statistical significance (available on request). They largely conform to the expectations in Figure 1. Specifically, there is a clear citizen group membership effect on contacting members of Congress while the trade and professional association variable is not significant. Holding all explanatory variables at their mean, and setting the value of membership in both groups at 0, the baseline likelihood of contacting Congress with low-effort tactics is 29 percent. Flipping the citizen group membership variable from 0 to 1 raises the likelihood of contact to 49 percent, an increase of 20 percentage points. This is the interest group ‘effect’, at least for citizen groups. The political inclination variable is also significant for low-effort tactics, suggesting that the value-added of citizen group membership is in ability to make contact with Congress. It has already been shown that politically inclined citizens are attracted to citizen groups, so while membership may stimulate a greater interest in participation, the real benefit for members is that group lobbyists can make contacting easy, directing their efforts where they will have (from the group’s point of view) the greatest impact. Of course, much of this is ‘thin’, astro-turf participation, sending identical form letters, postcards and email on issues given priority by group leaders.
Estimates of contacting Congress by all and Internet-based low-effort methods. Maximum likelihood estimates (robust standard errors).
*p < 0.05; **p < 0.01; ***p < 0.005.
Further evidence that the citizen group membership effect is largely in terms of ability to connect members to Washington, D.C. comes from exploring the extent to which it may be compensating for a member’s lack of familiarity with the Internet. Internet-savvy individuals are found to be more likely than others to contact Congress, presumably through online petitions and email, as predicted in Figure 1. Citizen groups, however, may be helping their less online-oriented members use this increasingly common vehicle for contact, thus compensating for their lack of knowledge and raising the likelihood of contact from what it would have been outside the group. Calculating the first difference effect for flipping the citizen group dummy from 0 to 1 when the Internet familiarity index score is 0 reveals that a member with no real online experience is just as likely (likelihood of 23 percent) to contact a congressional office as is a non-member whose familiarity with the Internet is almost exactly at the index’s mean.
Citizen group compensation for members’ lack of Internet familiarity may be even greater if only online methods of contacting are considered. I re-estimate the low-effort contact model with a new dependent variable indicating use of only Internet-based methods (email, online petitions and websites). The results are given in the last column of Table 3 and show that Internet familiarity is statistically significant and that the education variable is now significant when it was not in the general low-effort contacting model (although it is significant in both estimations of the political inclination index). To pursue this even further, I calculate the likelihood of contacting Congress electronically for all values of the Internet familiarity index for both citizen group and non-group members and graph the results in Figure 2. Unsurprisingly, when it comes to online contacting, it does not matter whether or not an individual is in a citizen group if they have no experience with the Internet. But the increase in the likelihood of electronic contact for citizen group members increases more quickly than for non-members with even small increases in Internet familiarity. The implication is that if members have just a little experience with the Internet, the organization compensates them for the rest, making it easier to fulfil the desire to make contact.
Likelihood of online contact by group membership and Internet familiarity.
The political inclination instrumental variable, however, is no longer significant. Perhaps because Internet use also contributes to inclination, or because the Internet has made contacting political officials easy for those who know how, that ability to participate in politics through the Internet trumps the desire to do so (though presumably there must be a little desire). If so, then while the Internet may be expanding communication between representatives and the represented, the quality of communication flowing from the districts to Capitol Hill may be lacking in informed content, personal political beliefs and even civility (as Goldschmidt and Ochreiter (2008) found is too often the case). Any wild thought, whether relevant to national politics or not, can instantly be sent off to lawmakers. It is beyond the scope of my work, but it is possible that citizen groups at least ensure some coherence to the communication they inspire; rambling and incendiary remarks are not likely to help them advance their policy goals.
As for high-effort, in-person contact with senators and representatives, I had suggested that wealthier professional groups might exhibit an effect here, but they do not and neither do citizen groups. In fact, none of the explanatory variables is statistically significant, not even political inclination, and I do not provide the results of the estimation here. It would appear that Olson (1965) is still right when it comes to work-related groups; few people join them out of a desire to be politically active. This may be especially true when participation means meeting personally with members of Congress rather than less personal and interactive contact like sending letters or email, although the statistical results provide no insight here.
This is still a little surprising because the political inclination variable did positively influence the likelihood of joining a trade and professional group here (recall Table 2). Cost may not be the entire explanation, because part of the high-effort contact variable included meeting the senator or representative in person at a local event, such as a town-hall meeting or at a civic organization (like a Rotary Club meeting). It may be that because these types of organizations typically do not rely as much on member involvement when it comes to political advocacy as citizen groups, largely preferring quiet personal contacts between legislators and lobbyists (Berry, 1999; McFarland, 1984), and still do not feel the need to encourage their members, even those who are politically inclined, to make either high-effort or low-effort contact with Congress.
Conclusion
What amendments should now be made to Figure 1 regarding the interest group effect and its causes? Most of the hypothesized relationships were empirically supported in the analyses. Changes include drawing a new connection from an individual’s level of political inclination to membership in a professional group, but then erasing the connection from it to in-person contacting. The first proved to be a significant relationship, just not as strong as the connection between inclination and citizen group membership. Otherwise, my claim that inclination precedes group membership, and is shaped by SES resources and membership in non-political organizations, is largely supported, even with some re-enforcing feedback on inclination from citizen group membership (causality is never quite so clear cut). It is, after all, being continually reinforced by the passions of group leaders and other members, even pushed in more ideologically extreme directions depending on how groups frame issues and policy solutions. There is also a connection from Internet familiarity to inclination to be added in Figure 1.
The findings reveal clear evidence of a stimulus due to group membership on contacting members of Congress, i.e. an interest group ‘effect’. This effect, however, is more on the ability to contact side rather than stimulating more interest in participation, and is largely confined to open-membership, purposive incentive offering, activist-oriented citizen groups. These groups make it easy for members to direct their passions at lawmakers whom group leaders believe are crucial to agenda-setting and winning roll-call votes. Of course, all of these data come from a single cross-sectional survey with all of its limitations. It also focuses only on a single type of target, members of Congress, not the president, implementing agencies or officials at other levels of government. Nor were characteristics of the targets included in the models, or finer distinctions made between types of groups beyond the simple, if theoretically grounded, typology. It also does not show whether citizen groups are improving the quality of communication, and increasing the ease of contact may actually degrade its quality and personal touch, even as it increases its quantity (possibly quantity of identical, astro-turfing messages).
Citizen inclination towards political action, the role of SES in shaping it and the role of organized interests has a long history in the literature, and rightfully so. If representative democracy is to function properly, with elected officials articulating the concerns and desires of the represented, then understanding who participates and why, and how to stimulate more of it, is a crucial job for scholars and practitioners. Indeed, the normative role of interest groups here is of concern. If they are pushing their members towards political action, then lawmakers are largely hearing the demands of narrow, if highly motivated, minorities whose success may come at the expense of unmobilized interests. On the other hand, the results also strongly suggest that interest groups, citizen groups at least, may be having a type of leveling effect on participation. Where citizens of high socio-economic status are, all things being equal more likely to develop sharp civic skills and, consequently, be more politically active, citizen groups compensate for this to some extent, helping lower SES citizens develop an interest in politics and giving them opportunities, directed opportunities to be sure, to express their interests. All of this should be the focus of future research.