Secondary menu

Aboriginal Electoral Participation in Canada

5. The Determinants of Aboriginal Voter Registration

In the previous section, our analyses showed that being registered for the election is a significant driver of turnout among Aboriginals. In this section, we examine the determinants of voter registration. It should be noted that this study does not measure the presence of the respondent's name on the register of electors. Instead, it relies on answers to a question asking respondents whether they recalled receiving a voter information card addressed to them personally.16 The results of the pooled logistic regressions are listed in Table 2. Once again, we present our results for Aboriginal and non-Aboriginal electors, and specify separate models for 20042008 and 20042011.

We focus first on the results for 20042008, as these include measures of civic duty to vote. To a substantial degree, Aboriginal registration is driven by the same factors as Aboriginal turnout. Five key variables of the preceding section are again relevant here. The older the respondent, the higher the likelihood of registration. Aboriginal electors who are interested in, who pay attention to and are knowledgeable about politics have a greater tendency to be registered (a gap of 25 percentage points). Civic duty and income also have a small positive impact, and Aboriginals living on reserves exhibit slightly lower levels of registration. The second column exhibits comparable findings among non-Aboriginal electors, though the effect of political resources is substantially larger for Aboriginals than for non-Aboriginals.

One factor does not follow the same pattern: education predicts turnout, but it does not affect registration among Aboriginals, while it does matter among non-Aboriginals. Another factor shows the reverse: gender has no bearing on turnout, but women more so than men tend to remember receiving a voter information card with their name on it though only among Aboriginal electors.

These results are essentially unchanged when we include the 2011 election (columns three and four).

The determinants of voter registration are rather stable across the four separate elections (see Table A4 in Appendix 2). The main ones political resources and age have positive and significant effects in all four cases. Civic duty possesses a positive coefficient that reaches significance in two of the three elections. Gender and income have impacts in the same direction throughout; however, they often fail to turn up as significant in these 500-respondent models. As for the effect of living on a First Nations reserve, shown in the pooled setup, it appears to have been driven by the 2008 data.

Table 2 Pooled Regression Models, Voter Registration
Aboriginals (20042008) Non-Aboriginals (20042008) Aboriginals (20042011) Non-Aboriginals (20042011)
Living on reserve -.05*
(.03)
-.04*
(.02)
Women .07**
(.02)
.01
(.01)
.06**
(.02)
.01*
(.01)
Age .31**
(.04)
.22**
(.01)
.31**
(.03)
.23**
(.01)
Education -.01
(.03)
.06**
(.01)
.01
(.03)
.06**
(.01)
Income .10**
(.04)
.05**
(.01)
.09**
(.03)
.05**
(.01)
Rural residence .04
(.03)
.01
(.01)
.02
(.02)
.02*
(.01)
Region: British Columbia .01
(.03)
.02
(.01)
.03
(.03)
.02
(.01)
Region: Ontario .04
(.03)
.00
(.01)
.06**
(.03)
.00
(.01)
Region: Quebec .05
(.04)
.08**
(.01)
.05*
(.03)
.08**
(.01)
Region: Atlantic .07
(.05)
.03*
(.02)
.08**
(.04)
.02
(.01)
Region: North .06
(.04)
-.07**
(.02)
.07*
(.04)
-.07**
(.02)
Political resources .25**
(.04)
.05**
(.02)
.29**
(.03)
.09**
(.01)
Civic duty to vote .08**
(.04)
.10**
(.01)
Parties talk about impor. issues .01
(.04)
.03**
(.01)
Saw general EC ad .03
(.02)
.03**
(.01)
.05**
(.02)
.03**
(.01)
Saw Aboriginal EC ad .01
(.03)
.00
(.02)
Number of cases 1477 6046 1922 8599
Pseudo R-squared .10 .13 .11 .13

Cells contain marginal effects of logistic regression coefficients with standard errors in parentheses. All variables range from 0 to 1. **significant at .05; *significant at .10

Tables A5a and A5b in Appendix 2 present the results for 20042008 and 20042011, respectively, according to residence on or off reserves and in urban or rural areas. Political resources exercise a consistently positive effect. Similarly, age is regularly correlated with registration. The impacts of income are less consistent, and likewise the effects of gender. Unlike in the explanation of turnout, sense of civic duty exercises an inconsistent effect on registration.

Finally, recall of Elections Canada's general advertisement is found to be positively and sometimes significantly related to registration, suggesting that these advertisements are effective at reminding electors of the registration process.


16 With perfect recall, this would be a perfect measure of registration, since voter information cards are sent to every eligible voter whose name appears on the preliminary lists of electors. However, survey recall is never perfect. Both false positives and false negatives are possible in the existing data.