Inspire Democracy Logo
Site Menu

Voter Turnout among Younger Canadians and Visible Minority Canadians: Evidence from the Provincial Diversity Project

Share this report   Share on Facebook Share on Twitter Share on LinkedIn

3. Characteristics of Younger Voters

The next step in identifying why younger Canadians are less likely to vote than older Canadians is to identify the characteristics of voters. Analyses are performed separately for younger and older Canadians because some characteristics could be related to voting for one group, but not for the other one. The analyses rely on binomial logistic regressions to identify the characteristics of Canadian voters by age cohort, dividing respondents into those aged 18 to 24, 25 to 34, and 35 and older. The analyses were performed separately for federal and provincial elections. Table 4 reports the results of these analyses for voting in federal elections and Table 5 reports the results for voting in provincial elections.Footnote 4

Table 4. Correlates of Voting in Federal Elections among Younger Canadians
Voting in Federal Elections (1-0)
  18-24 25-34 35+
Socio-Demographic Factors B SE B SE B SE
Age .04 .13 -.02 .04 .04 .01 a
Woman -.34 .36 -.25 .23 .01 .21
Education (ref. no high school)
Post-secondary
University
Post-graduate
 
1.14
1.28
.62
 
.52
.45
.91
c
b
 
.99
1.05
1.32
 
.32
.32
.48
 
b
a
b
 
.46
.83
1.01
 
.25
.31
.52
b
Household income .08 .05 .11 .04 b .01 .04
Unemployed 1.73 .66 b -.65 .45 .34 .47
Student -.32 .37 -1.41 .41 a .77 .98
Married .36 .45 .46 .24 .58 .25 c
Time spent at religious inst. -.75 .53 .07 .34 -.22 .36
Visible minority .32 .45 -.40 .35 -.24 .40
Immigrant (ref. Can. Born)
0-10 years
11-20 years
More than 20 years
 
-2.18
-.39
-2.0
 
.72
.62
1.01
 
b
 
c
 
-2.08
-1.29
-.08
 
.57
.54
.50
a
c
 
-.51
.73
-.36
 
.79
.88
.41
Regions (ref. Ontario)
Quebec
BC
Prairies
Atlantic
 
.83
-.92
-.52
.55
 
.47
.51
.50
.74
 
.61
.17
-.60
.25
 
.39
.32
.30
.36
 
 
c  
 
1.05
-.22
-.12
-.08
 
.35
.27
.25
.27
 
b
 
 
 
Political Orientations and Attitudes
Interest in federal politics .07 .08 .11 .05 c .04 .04
Close to federal party 1.21 .41 b 1.14 .26 a .88 .21 a
Federal politics is too complicated .22 .67 -.25 .43 -.79 .39 c
People like me have no say in federal politics .39 .68 .09 .44 .65 .42
I would feel guilty if I did not vote in federal election 2.40 .55 a 3.17 .40 a 2.94 .32 a
I feel my vote can make a difference .91 .58 .93 .41 c 1.75 .39 a
Discussed politics during youth at home .96 .56 .54 .35 .21 .29
Confidence in HofC. -.16 .11 -.20 .08 c -.13 .05 b
Confidence in Elections Canada .21 .09 c .19 .07 b .14 .05 b
Attachment to Canada .09 .08 -.09 .05 .05 .04
Impact of federal govt. -.14 .09 .08 .06 .07 .04
Social trust .82 .39 c .00 .25 .03 .22
Constant -5.78 3.19 -2.85 1.40 c -5.68 .80 a
Number of observations 378 1358 3417

Source: Provincial Diversity Project. Entries report B coefficients (Binomial logistic regressions)
a: p<.001; b: p<.01; c: p<.05

Table 5. Correlates of Voting in Provincial Elections among Younger Canadians
Voting in Provincial Elections (1-0)
  18-24 25-34 35+
Socio-Demographic Factors B SE B SE B SE
Age .08 .13 .00 .03 05 .01 a
Woman .06 .34 .05 .24 .17 .21
Education (ref. no high school)
Post-secondary
University
Post-graduate
 
1.31
1.29
1.51
 
.44
.42
.89
 
b
b
 
 
.96
1.03
1.09
 
.31
.34
.46
 
b
b
c
 
.28
.89
.13
 
.26
.29
.39
 
 
b  
 
Household income .04 .05 .09 .04 c .02 .04
Unemployed .95 .66 -.83 .43 .33 .48
Student -.18 .35 -1.01 .40 c -.20 1.0
Married -.33 .46 .06 .24 .62 .23 b
Time spent at religious inst. -.09 .53 -.26 .30 -.39 .33
Visible minority -.20 .46 -.24 .41 -.48 .42
Immigrant (ref. Can. Born)
0-10 years
11-20 years
More than 20 years
 
-3.03
.33
-6.37
 
.94
.64
1.50
 
a
 
a
 
-2.18
-1.47
.42
 
.59
.68
.56
 
a
c
 
 
-1.73
.55
-.43
 
.58
.79
.39
 
b
 
 
Regions (ref. Ontario)
Quebec
BC
Prairies
Atlantic
 
1.59
.33
.50
.45
 
.42
.47
.53
.56
 
a
 
 
 
 
.91
.24
-.37
.23
 
.33
.31
.29
.32
 
b
 
 
 
 
1.40
-.18
.01
.32
 
.34
.26
.23
.26
 
a
 
 
 
Political Orientations and Attitudes
Interest in provincial politics .08 .08 .08 .04 .16 .04 a
Close to provincial party .66 .38 .70 .24 b .88 .21 a
Provincial politics is too complicated .36 .60 -1.21 .44 b -.68 .39
People like me have no say in provincial politics -.41 .63 .78 .43 .21 .39
I would feel guilty if I did not vote in provincial election 2.43 .50 a 3.52 .40 a 3.02 .27 a
I feel my vote can make a difference .19 .58 .49 .40 1.14 .36 b
Discussed politics during youth at home .75 .51 .41 .35 .14 .29
Confidence in prov. legislature -.06 .09 -.06 .09 -.19 .07 b
Confidence in Elections Canada .07 .09 .09 .06 .17 .08 c
Attachment to province .17 .07 c .10 .05 c .02 .04
Impact of provincial govt. .00 .09 -.04 .05 .00 .05
Social trust .20 .37 -.08 .23 .11 .21
Constant -6.94 3.15 c -3.85 1.19 a -5.34 .74 a
Number ofobservations 407 1385 3434

Source: Provincial Diversity Project. Entries report B coefficients (Binomial logistic regressions)
a: p<.001; b: p<.01; c: p<.05.

3.1 Socio-Demographic Characteristics of Younger Voters

Some socio-demographic characteristics systematically correlate with voting among all three age groups and for both provincial and federal elections. Education is one such characteristic. Canadians with higher education are more likely to report having voted. Being a recent immigrant is associated with a lower likelihood of reporting having voted in federal and provincial elections. Predicted probabilities were calculated based on the multivariate analyses by keeping all variables at the sample means and by varying whether or not respondents were recent immigrants or not. According to these predicted probabilities, the gaps in voter turnout between recent immigrants and the Canadian-born population are substantial. By keeping all other variables in the model at the sample mean (i.e. controlling for them), the predicted probability of having voted in the preceding federal election was 64% for Canadian-born respondents aged 18 to 24, compared to 17% for recent immigrants (those in Canada for up to 10 years) in that age group, a 47-point gap.Footnote 5 The observed gap is even larger for the provincial elections, at 58 points (67% vs. 9%). Among voters aged 25 to 34, the gap is 39 points with regards to federal elections (89% vs. 50%) and 45 points with regards to provincial elections (87% vs. 42%). Recent immigrants aged 35 and older also appear somewhat less likely to vote than the Canadian-born population, but the gap holds only for voting in provincial elections and is much smaller than among younger Canadians, namely 12 points (97% vs. 85%).Footnote 6

Another systematic correlate of voting among socio-demographic characteristics is region of residence, or more precisely living in Quebec. Quebecers are more likely than other Canadians to report having voted in the previous provincial elections, and this holds for all three age groups. The results are also consistent with recent voter turnout in provincial elections. For example, in the 2014 provincial elections in Quebec and Ontario, voter turnouts were respectively 71.4% and 52.1%. Quebecers' greater propensity to vote also holds to some extent for the federal elections, but only among those aged 35 and older.

Other socio-demographic characteristics are not systematic for all three age groups. Some are statistically significant among Canadians aged 35 and older only. As researchers studying other countries, including the United States (Wolfinger and Wolfinger, 2008), have found, being married is associated with a greater likelihood to vote. It holds for both federal and provincial elections, but only among older Canadians aged 35 and older.

Age is also such a characteristic. Of course, the age-range is much wider among our group of older Canadians (35+) than among the 18 to 24 and among the 25 to 34, but it is nevertheless worth emphasizing that the age variable is statistically significant among the 35 and older only. In other words, while we observe that someone who is 50, for instance, is more likely to vote than someone who is 35, we observe no difference in turnout between an 18-year-old and a 24-year old, or between a 25-year-old and a 34-year-old, after controlling for all other variables in the analyses.

Other correlates are statistically significant only among Canadians aged 25 to 34. Income falls in this category. The greater the household income, the more likely Canadians aged 25 to 34 are to report having voted in either federal or provincial elections. The predicted probabilities indicate a 14-point gap in voting in federal elections between the poorest and wealthiest Canadians aged 25 to 34. The gap is 15 points for voting in provincial elections. No such relationship is observed among Canadians aged 18 to 24 or 35 and older.

Being a student also falls into that category. Canadians aged 25 to 34 who are students are less likely to report having voted in either the last federal or provincial elections. The predicted probabilities indicate that the gap is 23 points for voting in federal elections (66% for students vs. 89% for non-students) and 10 points for voting in provincial elections. This finding is in sharp contrast with Gélineau's (2013) who observed that students were more likely – not less – to vote. We will come back to these divergent findings in the conclusion. Intriguingly, no relationship is observed among Canadians aged 18-24, and this holds both for federal and provincial elections. It is difficult to explain why being a student appears to decrease voting among the 25 to 34 but not among the 18-24. Maybe this could be explained by the fact that during the ages of 18-24, age trumps the student effect. In other words, voter turnout would be so low within that age group that it would not really matter whether one is a student or not. This reasoning, however, is speculative and before we conclude that being a student does not decrease voter turnout among the 18 to 24, we should wait until further analysis is conducted in the section on habitual non-voting. No relationship between student status and voting is observed among Canadians aged 35 and older, but it should be noted that very few people of this age group report still being in school (less than one per cent).

There is one socio-demographic characteristic that appears as a correlate of voting only for the 18 to 24, namely being unemployed. Its direction, however, is counter-intuitive, with unemployed people more likely to report having voted, and its effect is observed only for voting in federal elections, not for voting in provincial elections. It is difficult to make sense of this finding.

Other socio-demographic characteristics do not appear as correlates of voting for any group or at any level of government. Gender does not appear to systematically correlate with voting, nor does spending time with people at a religious institution. Interestingly, being of a visible minority background does not appear as a correlate of voting in any of the three age groups. While descriptive data in Table 1 indicate that visible minorities are less likely to vote in either federal or provincial elections, the analyses in Tables 2 and 3 indicate that such differences are not statistically significant once we control for socio-demographic characteristics and political attitudes.Footnote 7

3.2 Political Attitudes of Younger Voters

A certain number of political attitudes emerge as statistically significant correlates of voting among all three age groups and for both federal and provincial elections. First, Canadians who feel close to a party are more likely to report having voted than those who do not feel close to a party. Among Canadians aged 18 to 24, predicted probabilities indicate that 68% of those who feel close to a federal party report having voted, compared with 39% of those who do not feel close to a federal party; a 29-point gap. Similarly, the gap is 16 points for feeling close to a provincial party and voting in provincial elections; the relationship is statistically significant only at the .10-level however. Among the 25 to 34 year olds, those who feel close to a party are 16 points more likely to vote in federal elections and 10 points more likely to vote in provincial elections. Finally, there are gaps of four points and three points respectively for voting in federal and provincial elections among the 35 and older age group. The gaps are substantially smaller for this latter group, but nevertheless statistically significant. The sample size is much larger for this age group.

Second, Canadians of all three age groups who expressed that they would feel guilty if they did not vote were also more likely to report having voted than other Canadians. The gaps in voting between those who strongly agree that they would feel guilty and those who strongly disagree are substantial. This is in fact the strongest correlate of voting both for federal and provincial elections and among younger and older Canadians. Canadians aged 18 to 24 who strongly agree that they would feel guilty if they did not vote are 52 points more likely to report having voted than those strongly disagree with the statement (83% vs. 31%). A similar pattern is observed for provincial elections, with a 53-point gap. The strong correlation between the feeling of guilt and voting is also observed among the 25 to 34 and the 35 and older age groups. Among those aged 25 to 34, the gaps are 43 points and 54 points respectively for federal and provincial elections. Among those aged 35 years and older, the gaps are 18 and 21 points. These findings are consistent with previous research showing that the opinion that voting is a duty for citizens is one of the strongest predictors of whether or not someone votes (Blais, 2000: 100).

Other correlates are statistically significant, but not for all age groups. The perception that one can make a difference by voting is one such characteristic. Among Canadians aged 35 years and older, those who perceive that their vote can make a difference are more likely to vote in both federal and provincial elections. Those who strongly agree with the statement that they feel that by voting they can make a difference are seven points more likely to vote in federal elections than those who strongly disagree with the statement; the gap is five points for provincial elections. Among Canadians aged 25 to 34, the relationship holds only for federal elections, and it does not hold among Canadians aged 18 to 24 in either federal or provincial elections. It is difficult to make sense of these contrasting findings across age groups.

We also observe that confidence in the House of Commons or in the provincial legislature is negatively correlated with voting but not in all cases. Canadians aged 25 to 34 or 35 years and older who hold greater confidence in the House of Commons appear less likely to report having voted in the federal elections. The same holds for confidence in the provincial legislature and vote in the provincial elections, but only among those aged 35 years and older. Two observations are worthy here. First, even if we accept that greater confidence is associated with lower voter turnout, it does not mean that people abstain from voting because they have great confidence in their legislative bodies. Arguably, these findings make more sense if we look at them from the opposite perspective. These findings might instead signal that people who have lower confidence in the legislative bodies are more likely to vote either because they wish to replace the government or as a way of ensuring that governments are more accountable to the population. Second, even though the findings report a relationship between confidence in legislative bodies and voter turnout, the relationship is somewhat inconsistent, with greater confidence related to a lower propensity to voting among certain age groups only and not always for both orders of government. This is consistent with existing research indicating only a weak or even inexistent relationship between political discontent and political engagement (see Howe 2010: 38-40). Whether a relationship exists between political engagement and political discontent, what the direction of the relationship is, and why there is such a relationship remains a matter open for further clarification and discussion.

Confidence in Elections Canada is also a correlate of voting. For federal elections, we observe that Canadians of all three age groups who express greater confidence in Elections Canada are more likely to report voting. The gaps between those who express no confidence at all (0) and great confidence (10) are 44, 24, and 4 points respectively for Canadians aged 18 to 24, 25 to 34 and 35 and older.Footnote 8We also observe a relationship between confidence in Elections Canada and voting in provincial elections, a surprising finding considering that Elections Canada does not administer provincial elections. Presumably though, respondents are expressing a general confidence in our electoral system. For provincial elections, however, the relationship is observed only among the 35 years and older.

Attachment to one's province appears as a correlate of voting. Those expressing a stronger attachment are more likely to report having voted in the provincial elections. The relationship is observed among the 18 to 24 and the 24 to 35 age groups. The gaps in voting between those very strongly attached to their province (10) and those not attached at all (0) are 40 and 14 points respectively among the 18 to 24 and the 25 to 34 respondents. This is an interesting finding as it echoes Howe's findings (2010). It is also one of the only two political attitudes that emerge as a correlate of voting only among younger Canadians. No equivalent relationship is observed between attachment to Canada and voting in federal elections. The only other correlate of voting that is significant for younger Canadians only (18-24) is social trust (positively correlated with voting), but the impact is only for federal elections.

A number of political attitudes do not appear as systematic correlates of voting. First, surprisingly, interest in politics does not turn out to be systematic correlate of voting. It is a significant correlate of voting only among Canadians aged 24 to 35 for federal elections and among the 35 years and older for provincial elections. This does not mean that interest in politics is not related to voting. Additional analyses indicate that Canadians more interested in politics are likely to vote in all three age groups. The results do not remain significant, however, once the analyses control for other political attitudes. Results not presented.

Second, the perception that politics is too complicated (internal efficacy) sometimes emerges as a correlate of voting but its effect is not systematic. It correlates negatively with voting in provincial elections among the 24 to 35 years old and with voting in federal elections for the 35 years and older. The same holds for external efficacy. Canadians who believe that they have no say in political affairs are not more or less likely to vote in federal or provincial elections than other Canadians.

Third, the perception that the federal or provincial government exerts a significant impact on one's life does not relate to the propensity to vote. Finally, there is no evidence presented in Table 4 and 5 supporting a relationship between the frequency of political discussions at home during one's youth and the propensity to vote.

The variable “political discussion at home during youth” deserves a more in-depth discussion, however. The variable is different from all the other ones from a temporal perspective. Respondents are not indicating how frequently they discuss politics now, but rather how frequently they did that at home during their youth. This difference is potentially important because these political discussions likely temporally precede many or all of the political attitudes variables included in the model. One's interest in politics, closeness to a party or feeling guilt when not voting could have emerged in part because of the political discussions during one's youth. Accordingly, comparing their effect in a multivariate model like the ones in Tables 4 and 5 is unlikely to specify the model properly. To verify this possibility, additional analyses were performed assessing the relationship between frequency of political discussions and voting in federal and provincial elections, but this time controlling for only the socio-demographic variables. The results are unequivocal; when controlling for only the socio-demographic variables, the frequency of political discussions at home during one's youth is positively and significantly related to voting in both federal and provincial elections and for all three age groups (results not presented). Canadians aged 18 to 24, 25 to 34, and 35 and older who reported frequent political discussions at home were 25, 24, and 8 points more likely to have reported having voted in the federal election, respectively. Equivalent gaps are observed for voting in provincial elections.

Thus, there are a certain number of correlates of voting that are common to both younger and older Canadians. While some correlates apply only to younger Canadians (18 to 24 or 25 to 34), they are the exception (income, being a student, and attachment to the province). Moreover, it is important to emphasize that while there are a few correlates that are age-specific, they do not appear to be correlates specific to either provincial or federal elections. Overall, the dynamics of voter turnout in Canada appear to relate more to general political engagement than anything unique to either provincial or federal politics. These analyses are important as they allow us to further identify those characteristics that are likely to explain the gap in voting between younger and older Canadians. Political attitudes such as feeling close to party, feeling guilty when not voting, or confidence in Elections Canada are all significant correlates of voting and are attributes not equally shared by younger and older Canadians (see Tables 2 and 3). As to whether these political attitudes can really explain the gap in voting between younger and older Canadians and to what extent, we will come back to this in a later section.








Footnote 4 Multivariate analyses presented in Tables 4 and 5 (and all other following tables) are based on weighted data to respect the socio-economic composition in each province as well as the relative weight of provinces in Canada. Because the data are weighted, no pseudo R-squared is reported.

Footnote 5 Respondents were explicitly offered the choice to answer that they were not eligible to vote. Recent immigrants (and other respondents) who answered that they were not eligible to vote in the previous elections were not included in the analyses. Recent immigrants' lower propensity to vote is therefore not a reflection of their ineligibility to vote.

Footnote 6 Given the lack of variance in reported voter turnout among the 35 years and older (88% and 86% report having voted in the previous federal and provincial elections), the effect reported with the predicted probabilities are necessarily smaller than among Canadians. This holds for this variable and all others in the analyses.

Footnote7 It is important to note that the oversample of visible minority Canadians is not included in the analyses for Tables 2 and 3 while this oversample is used when reporting findings specifically for visible minorities (see Table 5).

Footnote8 Once again, the relationship for those aged 35 years and older is significant because of the larger sample size.