Intersectionality & Gendered Political Behaviour in Canada – Study Notes

Overview & Context

• Article: “Intersectionality and Gendered Political Behaviour in a Multicultural Canada” (Harell, 2017) – published in the Canadian Journal of Political Science.

• Central Goal: To systematically integrate intersectional feminist insights into large-N quantitative studies of political behaviour, specifically demonstrating how gender, ethnoracial background, and immigration status collectively and interactively shape electoral turnout and broader forms of civic engagement in Canada.

• Data Source: The Ethnic Diversity Survey (EDS), conducted in 2002 by Statistics Canada and Citizenship and Immigration Canada. This comprehensive survey involved a significant sample size ($n = 41{,}695$), crucially including oversamples of ethnocultural minorities to ensure adequate representation for subgroup analysis.

Core Concepts & Definitions

• Gendered Political Behaviour: Refers to observable patterns in political action, such as voting in elections, participating in community volunteering, or engagement with associational life (e.g., clubs, organizations), which exhibit distinct differences based on gender.

• Intersectionality (Crenshaw 1989/1994): A theoretical framework asserting that multiple systems of social oppression and discrimination (e.g., racism, sexism, classism, xenophobia) do not operate independently but rather intersect and interact, creating unique and distinct experiences for individuals at their junctures that cannot be understood by analyzing identity categories in isolation. The widely cited "traffic-intersection metaphor" illustrates this, where different axes of discrimination (like roads) collide, resulting in complex and compound forms of disadvantage.

• Situated Comparisons (Dhamoon 2010): A methodological approach within intersectional analysis that involves:

  1. Comparing the interplay of domination processes within similar or comparable social, political, or historical contexts.

  2. Examining how these interactions manifest across different social levels, from individual micro-experiences to broader macro-societal structures.

  3. Analyzing relative distinctions and power dynamics across various social groups (e.g., comparing the experiences of white women versus Asian women).

• Patriarchal Structures: Societal arrangements and systems that traditionally privilege male power, authority, and social roles over those of women. These structures are often identified as fundamental sources of various barriers to women's participation in political, economic, and social spheres.

• Visible Minority (Canadian usage): A specific demographic category used by Statistics Canada to refer to persons, other than Aboriginal peoples, who are non-Caucasian in race or non-white in colour. This term approximates the concept of racialized status in Canadian academic and policy discourse.

Literature Foundations

• Second-wave feminist scholarship: This wave of feminism, particularly prominent in the 1970s, extensively highlighted the situational (e.g., domestic and family roles), structural (e.g., disparities in education and employment opportunities), and socialisation factors (e.g., gender-role conformity) that historically limited women’s access to and participation in positions of power (Welch 1974).

• Critiques of universal womanhood (Butler 1990; Collins 2000; Mohanty 2004): Emerging from post-structuralist and women of color feminist thought, these critiques challenged the idea of a singular, monolithic experience of "woman" or "womanhood." They underscored the crucial necessity of accounting for intersecting social categories such as class, race, ethnicity, and sexuality, arguing that these factors fundamentally reshape gendered experiences.

• U.S. evidence:
  • Studies have shown that Black women sometimes demonstrate higher rates of political participation than both white women and Black men (Cole & Stewart 1996; Tate 1991), challenging simplistic gender or racial analyses.
  • For African-American women, racial identity is often reported to be a more salient and influential factor in shaping political attitudes and behaviours than gender identity (Gay & Tate 1998).

• Canadian evidence:
  • Research indicates that visible-minority women tend to exhibit lower levels of political knowledge and engagement compared to white women (Gidengil 2007), suggesting compounded disadvantages.
  • Aboriginal women show notably larger gender gaps in vote choice, especially those residing on-reserve (Harell & Panagos 2013), highlighting unique challenges rooted in colonial histories and ongoing systemic issues.
  • Immigrant status has been identified as a significant moderator of political participation, with distinct patterns observed across different generations of immigrants (Chui et al. 1991; Couton & Gaudet 2008).

Methodological Challenges

• Dominance of survey research: Traditional survey designs often rely on a binary sex variable (male/female), which inherently masks the substantial intra-gender diversity and varied experiences that arise from intersecting identities.

• ar{x}= rac{\sum{i=1}^{n}x{i}}{n} Averages: The reliance on simple averages for overall populations or broad gender categories tends to statistically overweight the numerically dominant groups (e.g., white, middle-class, non-immigrant respondents), thereby obscuring the distinct patterns of minority subgroups.

• Small subgroup sizes: When analyzing highly specific intersectional subgroups (e.g., first-generation East Asian immigrant women), the sample sizes can become very small, leading to increased standard errors and making it statistically difficult to achieve significance for observed differences.

• Necessity of oversampling or alternative data: To overcome these limitations, studies on intersectionality often require deliberate oversampling of ethnocultural minorities or need to incorporate alternative data sources such as administrative records for broader coverage or qualitative methods (e.g., interviews, focus groups) for deeper mechanistic exploration.

Research Questions / Expectations

1. Do ethnoracial minorities in Canada exhibit lower rates of political participation compared to Canadians of European descent, and is this disparity partly attributable to the concentration of immigrant populations within these groups?

2. Are gender gaps in political participation more pronounced or manifest differently within specific minority communities compared to the general population, potentially reflecting unique challenges and barriers faced by immigrant women or women of color?

3. To what extent are any observed participation gaps mediated or explained by differential socioeconomic resources, such as variations in educational attainment, income levels, or employment status, across gender and ethnoracial groups?

Data & Operationalisation

• Dependent Variables
  • Turnout (self-reported) at three levels of government: federal elections, provincial elections, and municipal elections, offering a comprehensive view of electoral engagement.
  • Associational involvement: This composite measure includes self-reported membership in various types of voluntary groups such as Art and Culture groups, Community and Neighbourhood associations, Ethnic or Immigrant associations, Hobby or Interest clubs, Youth organizations, Professional associations, Religious groups, Service clubs, and Sports organizations. It also includes whether the respondent volunteered for any of these groups within the past 12 months.

• Key Independent Variables
  • Gender (male = 0, female = 1).
  • Ethnoracial background (categorized based on the respondent’s self-identified first ancestry only): European, Chinese, South Asian, African/Caribbean, East/Southeast Asian, Arab/West Asian, Latin-American, allowing for nuanced comparisons beyond broad categories.
  • Immigration Generation: Categorized as 1st generation (foreign-born individuals), 2nd generation (Canadian-born individuals with at least one foreign-born parent), and 3rd+ generation (Canadian-born individuals with both parents born in Canada).
  • Controls: Age cohort (18–29, 30–54, 55+), education level (categorized into 4 levels, from less than high school to university degree), full-time employment status, marital status, and presence of children in the household. These controls help account for known demographic predictors of participation.

Empirical Findings

1 Turnout Patterns

• Descriptive (Table 1):
  • Overall, there are minimal gender gaps in turnout across all levels of government. The only statistically significant gap observed is among European-descent women, who reported slightly higher municipal voting (67%) compared to European-descent men (65%).
  • A consistent pattern shows that European-descent voters overwhelmingly report higher turnout rates (approximately 80%) compared to all other ethnoracial groups, whose rates range from 65% to 76%, highlighting a significant ethnic-racial disparity.

• Immigration interaction (Table 2):
  • Both 1st and 2nd generation non-European voters consistently report lower turnout rates than their European counterparts across all levels of government (e.g., federal 1st generation: 73–70% for non-Europeans vs 85–84% for Europeans).
  • A traditional male-advantage gap in turnout (men participating more than women) is notably visible for 1st-generation non-Europeans at the federal level (− 3 percentage points; p < .05), suggesting the influence of origin-country gender norms or specific barriers related to recent immigration.
  • The reverse gender gap at the municipal level (women participating more than men) is exclusively present among European-descent individuals across all immigration generations.

• Multivariate (Figure 1; logistic regression):
  • Analysis using a three-way interaction term (gender × descent × generation) strongly suggests that ethnicity-immigration cleavages are the dominant factors influencing turnout, with gender showing statistically insignificant effects within most specific ethnoracial and generational categories.
  • The inclusion of socioeconomic resource controls (e.g., education, employment, income) significantly compresses observed turnout gaps, indicating that a substantial portion of these differences can be accounted for by resource disparities.
  • Turnout rates generally increase with generational status, particularly for non-Europeans, suggesting a process of integration or acculturation over time.
  • Socioeconomic variables emerge as powerful and consistent predictors of turnout: a university degree, for instance, significantly increases the log-odds of voting (β≈ $1.03$), and full-time employment also shows a positive effect (β≈ $0.19$).

2 Associational Involvement

• Raw membership rates (Table 3):
  • In sports groups, men consistently show higher participation rates than women across every ethnoracial group, with a male advantage ranging from 7 to 17 percentage points.
  • In contrast, women often report higher participation in Community and Youth groups (1–3 percentage points advantage).
  • For Religious groups, women show higher involvement rates than men among Europeans, African/Caribbean individuals, and East/Southeast Asian individuals, with differences ranging from 4 to 8 percentage points.

• Overall involvement (Table 4):
  • When considering membership in any type of group, there is an overall small male advantage (men 49% vs. women 46%, a 3 percentage point gap).
  • Crucially, when participation in sports groups (which are male-dominated) is excluded from the overall measure, the gender gap reverses, with women showing higher rates of engagement across most remaining group types.
  • Rates of volunteering within groups are generally high (approximately 70%) among members of European, South Asian, and African/Caribbean descent. Gender gaps in volunteering are mostly negligible, with the exception of Chinese and East/Southeast Asian groups, where women report significantly higher volunteering rates (+7 pp and +6 pp respectively).

• Multivariate (Figure 2):
  • When sports participation is excluded from the dependent variable and socioeconomic controls are added, women consistently surpass men in overall associational membership across both European-descent and non-European ethnoracial categories.
  • The positive gender gap favoring women is particularly strong among European-descent women (β≈ $0.24$ after controls), suggesting a robust propensity for civic engagement in non-sports activities.
  • Among non-Europeans, a significant male advantage is initially observed only in the 2nd generation before controlling for resources; however, this advantage disappears or even reverses after adjusting for socioeconomic factors, underscoring the influence of these resources on participation patterns.

Interpretations & Theoretical Implications

• Turnout: The findings suggest that for electoral participation in Canada, the intersection of ethnicity and migration generation is more salient and explanatory than gender. The traditional male-advantage gender gap primarily surfaces among 1st-generation non-Europeans, which may reflect the influence of political norms and gender roles prevalent in their countries of origin.

• Civic Life: A consistent and sometimes stronger reverse gender gap (favoring women) emerges in broader civic engagement once male-dominated sports participation is analytically removed. This finding strongly underscores the pervasive role of gendered socialisation, which shapes individuals' preferences and opportunities for choosing specific types of organizational involvement.

• Resources vs Socialisation:
  • Situational and structural controls, such as educational attainment, employment status, and family configurations, are powerful predictors and explain a significant portion of the variance in electoral turnout.
  • However, these socioeconomic resources are less explanatory for patterns observed in associational involvement, indicating that participation in civic groups is more deeply influenced by culturally embedded and gendered socialisation norms rather than solely by material resources.

• Intersectionality Value: The study clearly demonstrates the significant value of an intersectional analytical framework. It highlights crucial variations in political behaviour that are concealed or averaged out when using simplistic, binary sex comparisons. It emphatically underlines the necessity for simultaneously modelling and considering the interactive effects of gender, race/ethnicity, socioeconomic class, and immigration generation to fully grasp the complexities of political engagement.

Methodological & Practical Takeaways

• Oversampling minority women: This is deemed absolutely essential for conducting statistically reliable and robust intersectional analyses, as it ensures that subgroups are large enough to detect meaningful differences and relationships.

• Mixed methods recommended: Combining large-N quantitative surveys (like EDS) for broad generalizability and identification of patterns with qualitative work (e.g., in-depth interviews, ethnographic studies) is highly recommended for exploring the underlying mechanisms, lived experiences, and nuanced motivations behind observed political behaviours.

• Policy relevance: The findings have direct policy implications, suggesting the need for targeted civic outreach and integration strategies that explicitly recognize and address the differentiated gendered experiences and unique barriers faced by women within various immigrant and racialized communities, rather than employing one-size-fits-all approaches.

Limitations & Future Directions

• EDS lacks attitude measures: The current dataset does not include detailed measures of political attitudes, policy preferences, or perceptions of discrimination, which limits the ability to fully probe the psychological and ideological underpinnings of participation.

• Small Latin-American subsample: The relatively small sample size for the Latin-American community ($n \approx 383$) restricts the precision and generalizability of findings specific to this group, leading to less reliable statistical inferences.

• Aboriginal peoples excluded: The EDS specifically excluded Aboriginal peoples, indicating a critical gap in the study's scope. This necessitates separate and dedicated research initiatives focusing on the unique experiences and political behaviours of Indigenous women and men in Canada.

• Future work: Recommended future research directions include:
  • Utilizing longitudinal designs to track changes in participation over time and across generations.
  • Incorporating objective turnout records (e.g., from electoral registries) to mitigate self-report biases.
  • Conducting organizational case studies to analyze specific types of civic engagement at a micro-level.
  • Further exploration of additional linguistic and regional intersections (e.g., examining the distinct experiences of Québec anglophone versus francophone women).

Key Numerical Highlights (selected)

• Total EDS sample: $n = 41{,}695$ completed interviews, reflecting a robust dataset.

• Women’s municipal turnout (European): $67\%$ compared to $65\%$ for men among European-descent individuals (p< .01), indicating a statistically significant, albeit small, reverse gender gap.

• Male advantage in sports participation: A substantial male advantage, up to $17\%$ (observed particularly among the African/Caribbean group), highlights a distinct gendered pattern in this specific form of association.

• Logistic regression coefficient for University degree: A university degree increases the log-odds of voting by approximately $1.03$ (p<.01), signifying education as a strong positive predictor of turnout.

• Overall organizational volunteering among members: Approximately $70\%$ for members of European, South Asian, and African/Caribbean descent, demonstrating high rates of civic contribution within these groups.

Ethical & Philosophical Reflections

• Averaging obscures minority women’s realities: The practice of relying on broad averages in political behaviour research can inadvertently obscure the unique and often more challenging realities faced by women from minority backgrounds, potentially perpetuating epistemic injustice—the wrongful withholding of knowledge and understanding from marginalized groups.

• Intersectional lens resists essentializing “woman”: Adopting an intersectional lens actively resists the problematic tendency to essentialize or treat "woman" as a singular, undifferentiated category. This aligns with and supports decolonising feminist praxis, which advocates for recognizing diverse standpoints and de-centering dominant (often white, Western) perspectives within feminist thought.

• Recognizing differentiated barriers: Acknowledging and understanding the varied and compounded barriers (e.g., racialisation processes, complexities of immigration bureaucracy, issues related to credential recognition for immigrants) faced by diverse groups is crucial for developing genuinely equitable and inclusive democratic participation strategies.