Site MenuThe Electoral Participation of Diverse Canadian Youth in the 2015 Federal Election
7. Explaining Electoral Participation of Diverse Youth:
A Multivariate Model
After reviewing the different individual characteristics and factors that are known to affect voter turnout, and looking at differences in these factors across the subgroups of Canadian youth, we now assess how these factors impact the electoral participation of the different youth groups. We thus include these factors in one multivariate model of voter turnout, and estimate the effect of socio-demographic factors, access barriers, political resources and attitudes, social and political mobilization, and social and political engagement on the probability of youth voting in the 2015 federal election, holding all other variables constant.Footnote 16 We estimate five separate models, differentiating respondents by: Aboriginal status (model 1, with non-Aboriginal youth as the reference category), visible minority status (model 2, with non-visible minority youth as the reference category), occupational status (model 3, with employed youth as the reference category), living environment (model 4, with urban youth as the reference category), and disability status (model 5, where youth with no disability is the reference category). As the dependent variable, voter turnout, is dichotomous, we use logistic regressions. The effect of each factor is presented as an odds ratio (O.R.), expressing the likelihood that a respondent will vote, compared to the reference group.Footnote 17
Table 5 presents the five multivariate models for the different youth groups. When explaining turnout of the different youth subgroups, we will first focus on the similarities between the different groups. In a second step we will discuss differential effects of various turnout predictors for each sub-group.
Our initial bivariate relationship could not reveal significant differences in turnout for Aboriginal youth versus non-Aboriginal youth, for visible minority youth versus non-visible minority youth, and for youth with disabilities versus youth with no disabilities. The results remain after controlling for socio-demographic characteristics, access barriers, motivation, mobilization and engagement. For occupational status, the differences in turnout remain the same after controlling for the different factors. Employed youth are 3.3 times more likely to vote, compared to unemployed youth.Footnote 18 Finally, after taking socio-demographics characteristics, access barriers, motivation, mobilization, and engagement into account, turnout differences between youth living in rural and more urban areas is no longer statistically significant. This means that turnout differences between youth living in rural and urban areas can be explained through differences in their socio-demographic profiles and differences in political resources, attitudes, and access barriers.
The results of models 1 to 5 suggest that, of all the socio-demographic variables, education is the only relevant factor. As expected, education is systematically and positively related to turnout for all youth groups. Unlike most socio-demographic characteristics, access barriers are all important factors in the explanation of youth's voter turnout. Knowledge about how to vote, and knowledge about the necessity to prove one's identity and address when voting, substantially boost turnout among all youth groups. Additionally, receiving a voter information card from Elections Canada makes youth on average 3.8 times more likely to vote compared to youth who did not receive a voter information card. Respondents who used Elections Canada's online registration platform were also twice as likely to vote, compared to those who did not use the online platform. The results also suggest a consistently strong positive relationship of respondents' perceived ease of voting and turnout. As for political attitudes, we find a consistently strong positive relationship between interest in the federal election and turnout for all youth subgroups. Moreover, general interest in politics does not seem to matter for turnout, when controlling for other access barriers and political attitudes. More telling is respondents' belief about whether voting is a civic duty. Youth who belief that voting is a civic duty are on average 5 times more likely to turn out, compared to youth who belief that voting is a choice. Another systematic positive predictor of turnout is youth's belief about the responsiveness of politicians and policy makers. Youth's self-reported political competency does not seem to have an effect on turnout.
When we look at the variables capturing political mobilization, we find that, contrary to general expectations, being mobilized to vote and being contacted by a political party exert no significant impact on turnout when controlling for other turnout predictors, such as access barriers, political resources, and attitudes.
Finally, the results suggest a positive relationship between political engagement and turnout. Being more politically engaged increases youth's likelihood to vote. Volunteering and contacts with governmental offices are not significant factors in the explanation of youth's turnout when controlling for other predictors.
In order to examine whether the predictors have a differential impact for the separate youth groups, we tested the interaction effects between the predictors and the different youth categorizations (i.e. Aboriginal status, visible minority status, occupational status, living environment, and disability status). These additional analyses revealed significant differences between the youth subgroups. The results of the interaction effects will be discussed in the text.
When examining differential effects for turnout between visible minorities and non-visible-minority youth, we find that knowledge about how to vote has a stronger positive effect (O.R.=2.4, p=0.002) on non-visible-minority youth compared to visible minorities. This does not mean that this factor is not important for visible minorities; the results indicate mainly that this is of greater importance for non-visible-minority youth when predicting turnout. Other interaction analyses also reveal several differential effects along the lines of occupational status. First, we find that marital status is an important explanatory variable for turnout among unemployed youth. Married unemployed youth are 9.5 times more likely to vote compared to employed youth. For employed youth, the belief that voting is a civic duty is of greater importance. Employed youth who believe that voting is a civic duty are 5.7 times more likely to vote, compared to unemployed youth. The interaction effects also reveal some differential effects for the student population. More specifically, students who moved during the past year are 2.3 more likely to vote, compared to employed youth who also moved in the past year. In other words, residential mobility seems to be more detrimental for voter turnout among employed youth, compared to students. When focusing on youth's living environment, the results suggest one striking difference. Female youth living in rural areas are 2.5 times more likely to vote than their male counterparts, compared to urban female youth vis-à-vis urban male youth. Gender seems to be a significant predictor only for youth living in rural areas.
Our results suggest that electoral participation of youth can be explained by the same factors that explain the participation of older Canadian citizens. Education, access barriers, political attitudes, and political engagement operate in the same way for youth as with older citizens. So the traditional resource model seems to be applicable to youth as well (Blais and Loewen 2011, Gélineau 2013). The results nonetheless suggest some differences with the traditional turnout models with regard to the socio-demographic characteristics. Only youth's education level was important when predicting turnout. Age, gender, immigration status, marital status, residential mobility, and having children proved to be of little importance when predicting youth's probability to turnout. This might be due to the fact that certain variables show less variation for youth, compared to the general population, such as marital status and the number of children. Alternatively, other socio-demographic characteristics, such as residential mobility, might have a different meaning for youth. Youth generally tend to be more mobile and are less likely to be homeowners compared to older citizens, which may account for the limited explanatory power of this factor.
| Voting in the federal election (0–1) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Youth subgroups | Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||||
| O.R. | S.E. | O.R. | S.E. | O.R. | S.E. | O.R. | S.E. | O.R. | S.E. | |
| Aboriginal youth | 1.9 | .40 | ||||||||
| Visible minority youth | 0.8 | .23 | ||||||||
| Unemployed youth | 0.3** | .39 | ||||||||
| Students | 0.9 | .22 | ||||||||
| Youth living in rural areas | 0.7 | .18 | ||||||||
| Youth with disabilities | 1.6 | .43 | ||||||||
| Socio-demographics | ||||||||||
| Education | 1.2** | .05 | 1.2** | .06 | 1.2** | .05 | 1.2** | .05 | 1.2** | .05 |
| Married | 1.1 | .20 | 1.0 | .21 | 1.0 | .21 | 1.1 | .21 | 1.0 | .20 |
| Children | 0.7 | .23 | 0.8 | .23 | 0.7 | .23 | 0.7 | .24 | 0.7 | .23 |
| Residential mobility | 0.9 | .16 | 0.9 | .16 | 0.9 | .16 | 0.9 | .16 | 0.9 | .16 |
| Immigration status | 1.0 | .26 | 0.9 | .28 | 1.0 | .26 | 1.0 | .26 | 1.0 | .26 |
| Gender (male) | 0.9 | .17 | 0.9 | .17 | 0.9 | .17 | 0.9 | .17 | 0.9 | .17 |
| Age | 1.0 | .02 | 1.0 | .02 | 1.0 | .02 | 1.0 | .02 | 1.0 | .02 |
| Access barriers | ||||||||||
| Knowledge about how to vote | 1.4** | .12 | 1.4* | .12 | 1.4** | .12 | 1.4** | .12 | 1.4** | .12 |
| Knowledge about proofs of address and identity | 1.5** | .13 | 1.5** | .13 | 1.5** | .13 | 1.5** | .13 | 1.5** | .13 |
| Used online registration | 2.1*** | .21 | 2.1*** | .21 | 2.2*** | .21 | 2.1** | .21 | 2.2*** | .21 |
| Received voter information card | 3.8*** | .18 | 3.8*** | .18 | 3.7*** | .18 | 3.8*** | .18 | 3.7*** | .18 |
| Ease of voting | 2.1*** | .07 | 2.0*** | .07 | 2.0*** | .07 | 2.0*** | .08 | 2.1*** | .07 |
| Political resources and attitudes | ||||||||||
| Interest in politics | 0.9 | .13 | 0.9 | .13 | 0.9 | .13 | 0.9 | .13 | 0.9 | .13 |
| Interest in election | 3.4*** | .14 | 3.5*** | .14 | 3.5*** | .14 | 3.3*** | .14 | 3.3*** | .14 |
| Voting is a civic duty | 5.0*** | .17 | 5.2*** | .17 | 5.1*** | .17 | 5.1*** | .17 | 5.1 | .17 |
| Satisfaction with democracy | 0.9 | .10 | 0.9 | .11 | 0.9 | .11 | 0.9 | .11 | 0.9 | .11 |
| Feeling of political competence | 0.9 | .08 | 0.9 | .09 | 0.9 | .08 | 0.9 | .08 | 0.9 | .08 |
| Perception of responsiveness | 1.5*** | .09 | 1.5*** | .09 | 1.5*** | .09 | 1.5*** | .09 | 1.5*** | .09 |
| Political knowledge | 1.0 | .07 | 1.0 | .07 | 1.0 | .07 | 1.0 | .07 | 1.0 | .07 |
| Mobilization | ||||||||||
| Mobilized to vote | 0.9 | .05 | 0.9 | .05 | 0.9 | .05 | 1.0 | .05 | 1.0 | .05 |
| Contacted by a party | 1.2 | .18 | 1.2 | .18 | 1.2 | .18 | 1.2 | .18 | 1.2 | .18 |
| Engagement | ||||||||||
| Volunteering | 1.1 | .18 | 1.1 | .18 | 1.2 | .18 | 1.2 | .18 | 1.1 | .18 |
| Political engagement | 1.1** | .04 | 1.1** | .04 | 1.1** | .04 | 1.1** | .04 | 1.1** | .04 |
| Contact with government | 0.9 | .16 | 0.8 | .16 | 0.9 | .16 | 0.8 | .16 | 0.9 | .16 |
| Constant | 0*** | .97 | 0*** | .96 | 0*** | .98 | 0*** | .97 | 0*** | 1.0 |
| Pseudo R-square | 0.572 | 0.570 | 0.578 | 0.571 | .18 | 0.570 | ||||
| Number of observations | 1,845 | 1,825 | 1,842 | 1,821 | 1,840 | |||||
Note: The dependent variable with regard to voting behaviour in the 2015 federal elections is measured as a dichotomy where 1=yes, and 0=no. Significance levels: *** p<.001 - ** p<.01 - * p<.05
Despite the generally limited predictive power of the socio-demographic variables for youth, our analyses suggest some differential effects for certain youth subgroups. Being married is an important factor to explain turnout among unemployed youth. Gender is important only when explaining voter turnout among youth living in rural areas, with women voting at a higher rate than men. Residential mobility, on the other hand, turns out to be more detrimental for turnout of employed youth compared to the student population. The results also suggest a differential effect of an access barrier for visible minorities and non-visible-minorities, where knowledge about how to vote is more important for non-visible-minorities when predicting turnout. Finally, for employed youth, the belief that voting is a civic duty is more important, compared to unemployed youth.