In today's democracies, the battle for formal, legal political equality might appear to be won: women have the same right to vote, run for office, and participate in other political practices as men. But women do not always exercise their rights at the same rate as men, resulting in a gender gap in participation in many political practices. Gender gaps in political practices are problematic because women's participation makes a difference to policy. For instance, women activists tend to prioritize different issues (Schlozman et al. Reference Schlozman, Burns, Verba and Donahue1995), and women representatives tend to introduce different legislation (Paxton, Kunovich, and Hughes Reference Paxton, Kunovich and Hughes2007). Gendered asymmetries in democratic participation and influence entail collective outcomes that are less attentive to women's voices.
The nature of the participatory gender gap varies across different forms of participation; both more traditional forms of participation, such as voting and contacting elected officials, and newer forms of activism, such as “buycotting” (Coffé and Bolzendahl Reference Coffé and Bolzendahl2010; Dalton Reference Dalton2008; Stolle and Micheletti Reference Stolle and Micheletti2013). However, gendered variation in one important form of participation—attending face-to-face political discussion groups—remains largely unexamined. Experimental studies of small-group discussions show that in certain contexts—often, when women are the numerical minority—women speak less and exert less influence than men in group discussions (Karpowitz and Mendelberg Reference Karpowitz and Mendelberg2014; Mendelberg, Karpowitz, and Oliphant Reference Mendelberg, Karpowitz and Oliphant2014). Given the centrality of women's presence for their discursive influence, it is important to identify gendered asymmetries in participation in political discussion groups.
Unfortunately, little research considers whether women and men show up to discuss politics in face-to-face groups at similar rates. To date, Jacobs, Cook, and Delli Carpini's (Reference Jacobs, Cook and Delli Carpini2009) study of discursive participation in the United States is the only nationally representative survey study identifying variation—including gendered variation—in small-group discussions. However, Jacobs, Cook, and Delli Carpini make the classic “gender gap pitfall” (cf. Gidengil Reference Gidengil2007) of treating women as a homogeneous group and fail to consider how gender and other social attributes such as ethnicity intersect with gender to structure discursive participation.
My article offers the first comprehensive analysis of the gender gap in political discussion group attendance. I consider not only whether women and men are as likely to attend political discussion groups—meetings organized to talk about matters of local, national, or international concern—but also whether gender intersects with other identities to impact this form of discursive engagement. I first discuss the importance of face-to-face political discussion groups for democratic governance and review the literature on the gender gap in political participation. I then describe the data and cross-sectional research design. My results reveal that women report attending political discussion groups at significantly lower rates than men and that ethnicity intersects with gender in some important ways. I also find that poverty suppresses participation in small-group discussions, especially among men. Finally, I find that the presence of young children in the home functions to boost participation in small-group discussions. I conclude by discussing the implications of my findings for democracy and propose suggestions for boosting women's participation in face-to-face political discussion groups.
PUBLIC DISCOURSE AND THE PROBLEM OF GENDERED ASYMMETRIES OF PARTICIPATION
Discursive political practices—talking politics and deliberating—play essential roles in democratic systems. Researchers have paid particular attention to deliberative politics: not only is the study of deliberation now the “most active” area of political theory (Dryzek Reference Dryzek and Rosenberg2007), but also empirical studies of deliberation are “now more numerous than the normative works that prompted them” (Thompson Reference Thompson2008, 498). In line with Jürgen Habermas's original definition of deliberation as rational discourses about collective issues (Chambers Reference Chambers2009; Habermas Reference Habermas1984, Reference Habermas1990, Reference Habermas1998), I define deliberation as discursive practices that “generate influence through the offering and receiving of cognitively compelling reasons about matters of collective concern” (Warren Reference Warren2017, 47).
Theorists and researchers are interested in deliberation because deliberation achieves important outcomes that are essential for democratic governance (Beauvais and Bächtiger Reference Beauvais and Bächtiger2016; Dryzek Reference Dryzek and Rosenberg2007; Warren Reference Warren2017). For instance, there is considerable evidence that deliberation promotes citizen learning and more reasoned public opinion (Fournier et al. Reference Fournier, van der Kolk, Carty, Blais and Rose2011; Gastil and Dillard Reference Gastil and Dillard1999a, Reference Gastil and Dillard1999b; Knobloch and Gastil Reference Knobloch and Gastil2015; Landemore Reference Landemore, Landemore and Elster2012; Warren and Gastil Reference Warren and Gastil2015) and that deliberation strengthens the consistency between citizens’ principles and preferences (Fournier et al. Reference Fournier, van der Kolk, Carty, Blais and Rose2011). Deliberation also helps citizens reach mutual understandings (Fournier et al. Reference Fournier, van der Kolk, Carty, Blais and Rose2011), is expected to promote empathy (Mendelberg Reference Mendelberg2002), and deepens public faith in the legitimacy of democratic processes and outcomes (Benhabib Reference Benhabib1994; Gutmann and Thompson Reference Gutmann and Thompson1996; Karpowitz and Mendelberg Reference Karpowitz and Mendelberg2014; Parkinson Reference Parkinson2003; Warren and Gastil Reference Warren and Gastil2015). Deliberation can also catalyze other forms of democratic participation (Barber Reference Barber2003; Gastil, Deess, and Weiser Reference Gastil, Deess and Weiser2002; Knobloch and Gastil Reference Knobloch and Gastil2015).
Of course, deliberation is not the only form of public discourse that serves democratic aims. There are many instances in which public discourses about issues of collective concern do not satisfy deliberative standards of reason-giving. Citizens attending public rallies might shout slogans as assertions, and neighbors attending a local public meeting might reduce manifestos to declarations without accompanying justifications. There may be good underlying reasons for assertions and declarations, but these utterances typically take the form of concise claims without explanation (Walzer Reference Walzer and Macedo1999). Because nondeliberative political communication is less “rational” than deliberation—in the sense that nondeliberative political communication does not involve swaying others with cognitively compelling reasons—political communication might not be as effective as deliberation for producing informed public opinion (Beauvais Reference Beauvais2018b). However, nondeliberative political communication still serves important democratic aims. Political communication is an effective tool for capturing partisans’ or activists’ attention, and so it links with other political practices in important ways. Nondeliberative political communication—issuing slogans, declarations, polemics, accusations, and other assertions without justification—helps political parties mobilize partisans during elections and helps activists mobilize sympathizers at rallies and public meetings.
Many important deliberative interactions also take place in state institutions such as legislatures (Bächtiger et al. Reference Bächtiger, Spörndli, Steenbergen, Steiner and Rosenberg2007; Steiner et al. Reference Steiner, Bächtiger, Spörndli and Steenbergen2004) or the courts (Chambers Reference Chambers1998) and through the dialogue between state institutions and civil society. However, most public discourses—including both deliberation and nondeliberative political communication—take place in the public sphere. Public discourses occur through mass media's highly unidirectional communication flows or when neighbors, friends, and family talk about matters of collective concern (Beauvais Reference Beauvais, Bächtiger, Dryzek, Mansbridge and Warren2018a; Beauvais and Bächtiger Reference Beauvais and Bächtiger2016; Chambers Reference Chambers2009). Public discourse also occurs in small, face-to-face groups organized specifically to discuss matters of collective concern. This includes both high-quality (but rarer) designed forums such as deliberative mini-publics (Ryan and Smith Reference Ryan, Smith, Grönlund, Bächtiger and Setälä2014; Smith Reference Smith2009) as well as more informal (but more common) discussion groups organized to discuss collective issues in the public sphere (Cook, Delli Carpini, and Jacobs Reference Cook, Delli Carpini, Jacobs and Rosenberg2007; Walsh Reference Walsh2004, Reference Walsh2008).
While public discourses occur in many forums, face-to-face deliberating groups—particularly designed, deliberative forums such as mini-publics — are often considered the “gold standard” of citizen deliberation (Cook, Delli Carpini, and Jacobs Reference Cook, Delli Carpini, Jacobs and Rosenberg2007; Mansbridge Reference Mansbridge2010). This is because even though deliberating groups sacrifice widespread participation, they create “quality space for reflection” (Mansbridge Reference Mansbridge2010, 55). Small groups are important contexts for promoting democratic deliberation, which refers to “deliberation between equals orientated to a shared practical judgement” (Owen and Smith Reference Owen and Smith2015, 228). Self-selection into activist groups organized to discuss and act on matters of collective concern is also important. In such voluntary groups, self-selection “produces a greater than average homogeneity of interests and values” where “identity of interest may be the rule rather than the exception” (Mansbridge Reference Mansbridge1977, 336). Both deliberation and nondeliberative political communication within activist meetings can be highly motivating, encouraging political education or other forms of political engagement (Beauvais Reference Beauvais2018b).
In terms of empirically studying political discourse, researchers have increasingly focused on “small-scale deliberative experiments” such citizens’ assemblies, consensus conferences, and deliberative polls (Chambers Reference Chambers2009, 331). Chambers (Reference Chambers2009) criticizes this trend, noting that small-scale deliberating groups cannot replace the institutions and organizations of mass democracy. One method that offers insight into face-to-face public discourse while remaining attentive to mass democracy is the use of nationally representative surveys to study citizens’ participation in political discussion groups. Participating in political discussion groups includes attending not only small-scale deliberative experiments to talk about issues of collective concern (such as mini-publics) but also more informal meetings organized to talk about issues of collective concern within civic associations, community and church groups, or schools and workplaces.
Thus far, Jacobs, Cook, and Delli Carpini (Reference Jacobs, Cook and Delli Carpini2009) are the only researchers who have conducted a nationally representative survey study of “discursive politics”—including attending a meeting to talk about local, national, or international issues—in the United States. These authors use a variable that asks respondents whether they attended a meeting to talk about local, national, or collective issues as a measure of what they call “face-to-face deliberation.” However, this variable does not strictly measure deliberation insofar as deliberation is defined as reason-giving about matters of collective concern. This variable is a more general measure of participation in face-to-face public discourse that does not distinguish between deliberative and nondeliberative political group discussions.
Even though this measure does not distinguish deliberation as reciprocal reason-giving, it is still useful for identifying group-based variation in who is attending face-to-face political discussion groups. Understanding who is showing to talk about matters of collective concern in small groups is important, as both deliberative and nondeliberative political discussions play essential roles in democratic systems. If certain social group members such as women or women of color are excluded from political discussion groups, their voices will go unheard in important processes of opinion formation, and they may not develop the same motivation to participate in other forms of political engagement as those who are present in group discussions.
Jacobs, Cook, and Delli Carpini (Reference Jacobs, Cook and Delli Carpini2009) find some evidence that women report participating in political discussion groups at lower rates than men. While this study offers an important first look at variation in participating in political discussion group, it makes the “gender gap pitfall” that Gidengil (Reference Gidengil2007) calls “categorical thinking.” Categorical thinking refers to when gender gap research only compares women and men and ignores differences among women or among men. According to Gidengil (Reference Gidengil2007), gender gap research must attend not only to both women's and men's contributions to gendered patterns of political behavior but also to the ways in which gender intersects with other identities and social factors, such as ethnicity, poverty, and parenthood.
Research on gendered political participation must be attentive to intersectionality, which refers to the way gender intersects with other social identities and material conditions (Crenshaw Reference Crenshaw1991; Davis Reference Davis2008; Shields Reference Shields2008). The political attitudes of women of color are developed through intersecting experiences of racism, colonialism, patriarchy, and labor exploitation and may manifest in forms of political engagement that differ from White women's political engagement. Research shows that gender intersects with race or ethnicity (Gay and Tate Reference Gay and Tate1998) and Indigeneity in colonial/postcolonial contexts (Harell and Panagos Reference Harell and Panagos2013; Herrick Reference Herrick2018) to impact the attitudes and democratic participation of visible minority and Indigenous women and men in distinct ways. Empirical intersectionality research shows that the effects of mutually constitutive categories of (dis)empowerment vary across groups, attitudes, and practices (Weldon Reference Weldon2006). For instance, with respect to political attitudes, there are significant differences between American Latinos’ and Latinas’ opinions related to women's social roles and compassion issues, and these gaps vary across American respondents of Cuban, Mexican, and Puerto Rican origin (Bedolla, Monforti, and Pantoja Reference Bedolla, Lavariega Monforti and Pantoja2007).
With respect to political behavior, there is evidence that gender has a larger effect on vote choice among White voters than among people of color (among Black, Latinx, and Asian voters) in the United States (Bedolla and Scola Reference Bedolla and Scola2006). With respect to a gender gap among Indigenous peoples in Canada, Harell and Panagos (Reference Harell and Panagos2013) find that although there are gender gaps in vote choice among Indigenous peoples, there is a much larger gap between Indigenous peoples and settler Canadians—suggesting that while gender is important, the experience of colonialism has a bigger impact than gender in shaping Indigenous peoples’ political attitudes and practices. However, Harell and Panagos (Reference Harell and Panagos2013) find that Indigenous women are less likely than Indigenous men to be involved in less formal forms of political practices (such as being a member of recreational or explicitly political organizations). This finding contributes to the expectation that Indigenous women might be less likely to attend political discussion groups.
With respect to material conditions, the “feminization of poverty” thesis maintains that poorer women will be less able to donate money or engage in resource-intensive activities such as volunteering for campaigns (Brady, Verba, and Schlozman Reference Brady, Verba and Schlozman1995). Workplaces—particularly high-status workplaces—are important for socializing civic skills. Therefore, women living in poverty who occupy a marginal position in the labor market may be particularly unlikely to develop the kinds of civic skills requisite for democratic engagement.
Finally, researchers should consider whether gender influences the effect of parenthood on political engagement. Parents—particularly the parents of young children—are expected to have less free time for politics. But because child-rearing and housework duties often disproportionately fall on women's shoulders, parenthood is assumed to suppress mothers’ political participation more than fathers’ participation (Coffé and Bolzendahl Reference Coffé and Bolzendahl2010). However, empirical research offers a mixed picture: there is evidence that parenthood (including motherhood) can motivate political confidence or efficacy (Gidengil, Giles, and Thomas Reference Gidengil, Giles and Thomas2008; Thomas Reference Thomas2012) and can boost certain forms of political engagement (Micheletti and Stolle Reference Micheletti, Stolle, Thomas and Bittner2017; O'Neill and Gidengil Reference O'Neill, Gidengil, Thomas and Bittner2017).
The intersection of gender and other social and socio-structural attributes might lead to the expectation that women—particularly poor women, visible minority or Indigenous women, and women with children—are less likely to participate in political practices. However, the gender gap literature related to women's political engagement is mixed. Existing research on gendered variation formal political participation (such as working for a candidate, joining a political party, contacting political officials, or donating money to candidates or political parties), shows that—with the exception of voting—there is a small but persistent gender gap in most formal political practices across democracies worldwide (Coffé and Bolzendahl Reference Coffé and Bolzendahl2010; Dalton Reference Dalton2008; Inglehart and Norris Reference Inglehart and Norris2003; Paxton, Kunovich, and Hughes Reference Paxton, Kunovich and Hughes2007). However, analyses of less formal participatory practices (such as protesting, signing petitions, online activism, and boycotting or buycotting) are more mixed. For instance, men seem to be more willing than women to demonstrate or join a meeting or rally, while women seem more willing to sign petitions, boycott, or buycott (Coffé and Bolzendahl Reference Coffé and Bolzendahl2010; Stolle, Hooghe, and Micheletti Reference Stolle, Hooghe and Micheletti2005; Stolle and Micheletti Reference Stolle and Micheletti2013). Despite the rich literature on gender gaps in democratic participation, less work has been done on discursive democratic engagement leaving an important question unanswered: Is there a gender gap in attending political discussion groups?
METHODS
I use data from the Canadian Election Study (CES) 2015 Web Survey to identify whether there are gendered patterns of attendance in political discussion groups. The CES 2015 Web Survey was completed by a nationally representative panel of Canadian respondents recruited by Survey Sampling International.Footnote 1 Canada shares many attributes with other Western democracies, including high levels of government transparency, economic development, economic freedom, quality of life, and civic freedoms. The findings from this study likely generalize to other Western democracies. Of course, future researchers should endeavor to collect similar data in other countries to facilitate comparative analyses. The outcome variable of interest, attending a political discussion group, is an indicator variable measuring whether respondents have attended a meeting to discuss a local, national, or international issue in the past 12 months (have attended a political discussion group = 1) (see Table 1).
Table 1. Variable distributions
To offer an overview of who is participating in political discussion groups, I first estimate a main effects model to identify the main effects of gender, controlling for basic demographics (age, ethnicity, and region). Gender is a self-reported dummy variable measuring respondents’ gender (female-identifying respondents, or women+, whom I refer to as “women,” are coded as 1) (Table 1).Footnote 2 Age measures respondents’ age at the time of the survey (2015). Region is a dummy variable controlling for the unique effect of residing in Québec, the home province of a majority French-speaking population and Canada's largest national minority group, the Québécois (residing in Québec = 1).Footnote 3 It is important to control for living in Québec because the Québec/Canada Outside Québec (COQ) cleavage accounts for the greatest source of variation in political policy preferences (more so than language or region more generally) (Anderson Reference Anderson2010). Furthermore, according to Statistics Canada, residents of Québec report having fewer close friends, having less contact with friends and acquaintances, being less likely to participate in community organizations and associations, and having lower levels of generalized trust (Statistics Canada 2015). The controversial question of why social capital appears to be lower in Québec remains unanswered (Hamilton Reference Hamilton2017) and is beyond the scope of this article.
In Canada, the question of “race” is less well studied, as Canadian researchers often prefer to focus on linguistic or cultural differences (Nath Reference Nath2011). However, the effect of being racialized—being categorized as “non-White” in societies marred by the ongoing legacy of White racism (cf. Mills Reference Mills1997)—may have important consequences for democratic participation, including attending political discussion groups. Furthermore, in countries such as Canada, the United States, Australia, New Zealand, and other settler colonies, Indigeneity may intersect with gender in important ways to impact participation in political discussion groups and other political practices (Alfred, Pitawanakwat, and Price Reference Alfred, Pitawanakwat and Price2007; Harell and Panagos Reference Harell and Panagos2013; Herrick Reference Herrick2018). As such, my measure of “ethnicity” is a three-category variable indicating whether a respondent is White (coded White = 1), a “visible minority” (non-White, non-Indigenous, coded visible minority = 2), or Indigenous (self-identifies as Aboriginal, First Nations, Métis, or Inuit, coded Indigenous = 3).Footnote 4 This measure of ethnicity captures the effects of White privilege, racialized identities, as well as the potentially unique effects that the legacy of colonialism has on Indigenous peoples’ participation in political discussion groups. Although the variables are coded such that the baseline of gender is male, and the baseline of ethnicity is White, I estimate and report average marginal effects for these variables: comparing the predicted probabilities that White, visible minority, and Indigenous women and men attend political discussion groups. This helps avoid treating the baseline categories in the regression models as implicit normative standard against which other groups are compared.
I then consider interaction models to clarify whether and how gender intersects with other factors to impact attending political discussion groups. I estimate small, coherent logit regressions with minimal dummies and controls (Achen Reference Achen2005). To clarify whether the effect of gender on attending political discussion groups varies by ethnicity, I estimate an interaction between gender and ethnicity, controlling for gender and ethnicity and basic demographics (age and region) (Model 2). In Model 3, I estimate an interaction between gender and poverty, controlling for the main effects of gender and poverty and basic demographics (age, ethnicity, and region). Poverty is an indicator variable measuring whether respondents’ incomes fall below $39,999 per year. The government of Canada considers incomes at or below $31,344 per year to be “low income” for the average (three person) family (Statistics Canada 2017c), but below $39,999 is the lowest response option on the CES 2015. In Model 4, I estimate an interaction between gender and parenthood, controlling for the main effects of gender and parenthood and basic demographics. I measure the effect of parenthood with a dummy variable indicating whether there are any young children (six years or younger) living in the respondents’ home.
Because attending a political discussion group is a dichotomous variable comparing respondents who did and did not attend a meeting to discuss matters of collective concern at least once in the past year, I proceeded with logistic regression. I applied the CES 2015 survey weights (CombWgt), applying national poststratification weight to province, age, gender, and household size, and estimate robust standard errors.Footnote 5
RESULTS
The results of my ordered logistic regressions (Table 2, Models 1–4) reveal interesting findings. Model 1 (Table 2) shows that men are significantly more likely than women to attend a meeting to talk about issues of collective concern at least once in the past year. Estimating the average marginal effects, there is a 20% predicted probability a woman reported attending a political discussion group in the past year and a 25% predicted probability a man reported doing the same (see Table A1 in the online appendix).
Table 2. Logistic regression results
Note: Robust standard errors in parentheses
*** p < .01; ** p < .05; * p < .1.
Of course, women and men do not constitute homogeneous groups. Modeling the interaction between gender and ethnicity reveals a significant gender gap among White men and White women and among visible minorities (Table 2, Model 2). However, there is no evidence of a gender gap among Indigenous peoples. The predicted probabilities of attending group deliberation for men and women in each ethnic group are plotted in Figure 1 (and presented in Table A1 in the appendix). Although Indigenous peoples constitute a small proportion of the Canadian population and a small number of respondents in the sample, these groups should be sufficiently large to detect effects. After listwise deletions, there are 135 Indigenous respondents in Model 2 with a subtotal of 67 Indigenous men and 68 Indigenous women. As Figure 1 reveals, the confidence intervals for Indigenous men and women not only overlap they extend past the other point estimates—lending confidence to the conclusion that there is no Indigenous gender gap in attending political discussion groups.
Figure 1. Predictive margins for gender by ethnicity, 90% confidence intervals.
It seems that when it comes to participating in public discourse, the political practices of Indigenous peoples differ from their settler-Canadian counterparts. Comparing among women, the probability a visible minority woman attended political discussion groups was no different than the probability a White woman did (see Figure A1 in the appendix). However, Indigenous women were significantly more likely than either visible minority or White women to report participating in a political discussion group. Similarly, comparing among men shows Indigenous men were more likely to attend political discussion groups than White men (at marginal levels of significance, p < 0.1).
People living in poverty were less likely to report attending a political discussion group (Table 2, Model 3). Note that the effect of poverty is robust controlling for education (see Table A6 in the appendix). Interestingly, with respect to poverty's intersection with gender, poverty seems to have a leveling effect. The six percentage point gender gap in political discussion group attendance is only significant among those living above the poverty line; among the poor, there is no gender gap in attending political discussion groups (Figure 2).
Figure 2. Predictive margins for gender by poverty level, 90% confidence intervals.
Comparing results among women and among men, the predicted probability a woman who is living in poverty attended a meeting organized to talk politics is five percentage points lower than a woman who is not poor, whereas the predicted probability a man living in poverty attended a meeting organized to talk politics is a full 10 percentage points lower than a man who is not poor (see Figure A2 in the appendix).
The presence of young children in the home has a positive effect on attending political discussion groups to talk about collective issues (Table 2, Model 4). If anything, it appears the presence of young children disproportionately impacts the probability a man reported attending a political discussion group (Figure 3).
Figure 3. Predictive margins for gender by parenting of young children, 90% confidence intervals.
Comparing results among women and among men, the predicted probability a woman with young kids reported attending a political discussion group is six percentage points higher than a woman without young kids, whereas the predicted probability a man with young kids reported attending a political discussion group is nearly 10 percentage points higher than a man without young kids (see Figure A3 in the appendix).Footnote 6 The presence of young children in the home clearly boosts political discussion group attendance more generally, for both women and men.
DISCUSSION
On average, men are significantly more likely than women to attend meetings organized to talk about local, domestic, or international issues. What does this mean for public discourse? While a five percentage point difference in the predicted probability that a man or woman attended a meeting to discuss matters of collective concern may seem small, it is meaningful. Existing research shows that in certain contexts, when women are the numerical minority, they speak proportionately less in political discussion groups (Karpowitz and Mendelberg Reference Karpowitz and Mendelberg2014; Mendelberg, Karpowitz, and Oliphant Reference Mendelberg, Karpowitz and Oliphant2014). Specifically, in discussion groups in which decisions are made by majority rule, even when women make up half or even a majority (up to 60%) of discussion group members, there are large and significant gender gaps in discursive participation and influence over collective outcomes. In political discussion groups in which decisions are made by majority rule, women must make up a supermajority—80%—of political discussion group participants before women speakers contribute proportionately to total talk time.
The CES 2015 does not offer insight into how (or whether) participants attending political discussion groups reached decisions. However, majority rule is often used for decision-making at school board meetings, civic associations, and other neighborhood groups. A five percentage point gender gap in the predicted probability of attending a political discussion group means women are more likely to be underrepresented in small-group discussions and are very unlikely to make up a supermajority. Given the centrality of women's descriptive representation for their discursive participation and influence, a five percentage point gender gap in the predicted probability of attending a political discussion group threatens the ideal of discursive gender equality.
Comparing results among women of different ethnic backgrounds and among men of different ethnic backgrounds does not offer clear evidence that being an Indigenous or visible minority group member additionally suppresses participation in political discussion groups. In fact, while there are no significant differences between visible minority and White women (or between visible minority and White men), Indigenous women and Indigenous men are more likely to participate in political discussion groups than their settler-Canadian counterparts. Furthermore, among Indigenous peoples, there is no evidence of a gender gap in attending political discussion groups. Beginning in 2011 and continuing until the 2015 Canadian federal election, Canada's Indigenous peoples were mobilizing around a movement called Idle No More, which advocated for Indigenous land rights, environmental protections, addressing emergency situations in Indigenous communities, and committing to a nation-to-nation relationship between Canada and Indigenous peoples (Kino-nda-niimi Collective Reference Collective2014). Clearly, leading up to the 2015 election, Indigenous peoples in Canada—both men and women—were mobilized to attend meetings to discuss matters of collective concern. Future research might consider the long-term impact of Idle No More on Indigenous political attitudes and practices. Future quantitative studies should endeavor to oversample Indigenous respondents or employ more appropriate strategies for sampling low-incidence population—such as stratified designs, list-based selection, or density strategies (Berry, Chouhoud, and Junn Reference Berry, Chouhoud, Junn, Ray Atkeson and Alvarez2016)—so that Indigenous voices are fully included in survey research and researchers can endeavor to identify how Indigeneity intersects with other attributes such as region, living on or off reserve, or urban/ rural differences, in addition to gender.
While my finding that living close to or under the poverty line (earning less than $39,999 per year) suppresses participation in political discussion groups, what is interesting is that—with respect to the gender gap in attending political discussion groups—poverty has a leveling effect because poverty particularly suppresses men's attendance in political discussion groups. Another interesting finding is that, contrary to the expectation that children constitute a drain on time otherwise spent on democratic practices, my work reveals that young children are an important means for motivating attendance in civic discourse. The CES 2015 cannot tell us what kinds of meetings people are attending to talk about matters of collective concern, but it does show us that children are drawing the adults in their lives into civic discourse. Children may be drawing adults into civic discourse through parent associations, community debates over rezoning near schools, debates over school closures or busing, or discourses related to the public provision of childcare and parental leave.
CONCLUSION
My nationally representative analysis of gendered variation in political discussion group attendance reveals that women are significantly less likely to participate in political discussion groups than men. The five percentage point difference in the predicted probability a man or woman attends a political discussion group is meaningful. In certain contexts in which women make up the minority—or anything less than a supermajority—of a discussion group, men's aggressive conversational behaviors can silence the voices of women who are present, undermining women's discursive influence (Mendelberg, Karpowitz, and Oliphant Reference Mendelberg, Karpowitz and Oliphant2014).
The gender gap in attending political discussion groups varies by ethnicity. However, contrary to expectations, there is no evidence that identifying as either an Indigenous or visible minority group member suppresses attendance in political discussion groups. In fact, both Indigenous women and men are more likely than White Canadians to report attending political discussion groups. While much research focuses on the “double disadvantage” of gender and Indigeneity or other racialized identities, these findings suggest that Indigeneity can motivate participation in public discourse.
Indigenous extrastate political engagement is important for a number of reasons. For one thing, because Indigenous peoples represent only 4.9% of Canada's population (Statistics Canada 2017b), Indigenous peoples’ minority status and geographic dispersion mean that they are underrepresented in Parliament and have less electoral influence than settler Canadians.Footnote 7 Furthermore, many Indigenous activists reject the legitimacy of settler institutions and are averse to participating in settler state politics. As Michi Saagiig Nishnaabeg philosopher and artist Leanne Betasamosake Simpson (Reference Simpson2015) explains,
One of the biggest outcome of the Idle No More movement has been an increased interest in Canadian electoral politics. That's heartbreaking to me, as someone who dedicated a significant amount of time and energy to that movement… Attention is now focused on facilitating change through voting or running for office, which seems so ridiculous to me… Movement building is a productive or generative politics of refusal when we are building and reinvigorating and embodying and amplifying our instance of acting as peoples who belong to specific Indigenous nations.
My findings show that Indigenous people are turning out to discuss matters of collective concern at high rates outside the formal political arena. This might be taken as evidence that Idle No More not only increased interest in Canadian electoral politics but may also be motivating the kind of relational, discursive, generative politics that Simpson describes.
My work reinforces concerns that material deprivation suppresses poor people's ability to participate in meetings organized to talk about collective issues. While the existing data does not offer clear insight into why the poor—and particularly poor men—are less likely to participate in political discussion groups, economic marginalization often entails fewer resources and opportunities to talk politics. Future research should endeavor to clarify why poverty prevents people from attending political discussion groups, so that we can identify ways to boost engagement in public discourse among the poor.
Eliminating the gender gap in political discussion groups likely starts with stimulating political interest and efficacy at a young age among all children. Parents—particularly mothers—play an essential role in socializing girls’ political interest (Gidengil, O'Neill, and Young Reference Gidengil, O'Neill and Young2010). Organizations in civil society can also stimulate girls’ political interest and efficacy. For instance, Equal Voice, a nonpartisan organization dedicated to women's electoral equality, organizes a program called Daughters of the Vote (http://www.daughtersofthevote.ca/) aimed at interesting young women in electoral politics. With respect to engaging women, small-scale deliberative experiments such as mini-publics are an excellent way to achieve women's descriptive (or over-) representation in political discussion groups—even if they are a relatively rare form of deliberative engagement (Beauvais Reference Beauvais, Bächtiger, Dryzek, Mansbridge and Warren2018a; Beauvais and Bächtiger Reference Beauvais and Bächtiger2016; Karpowitz and Raphael Reference Karpowitz and Raphael2014). Finally, the array of conventions, lecture series, and local meeting places that constitute the feminist subaltern counterpublic are essential for creating woman-friendly spaces for public discourse and helping women “formulate oppositional interpretations of their identities, interests, and needs” (Fraser Reference Fraser1990, 67).
Identifying gendered patterns of participation is important because ongoing asymmetries in women's democratic participation are threats to political equality and democratic responsiveness. Political systems are democratic to the extent that people are included in political practices (such as voting or talking politics), can communicatively link personal preferences into collective opinions and agendas, and are empowered to turn collective agendas into collective decisions (Warren Reference Warren2017). It is important to note that democratic processes always begin with inclusion—the aims of forming collective opinions and agendas, and making collective decisions are undermined if those affected by collective endeavors are excluded from participating in them (Fung Reference Fung, Nagel and Smith2013; Goodin Reference Goodin2007; Young Reference Young2011). That women and men turn out to talk politics at public meetings at different rates is troubling both because women cannot influence the outcomes of small-group discussions if they are not present, and because—even when women are present—women's minority status in political discussion groups can further undermine their discursive influence in conversations. More closely approximating the ideal of discursive gender equality—the ideal of women's and men's equal participation and influence in public discourse—requires greater efforts motivate girls’ and women's democratic engagement and to create woman-friendly spaces for public discourses.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/S1743923X18000892
