Secondary menu

Explaining the Turnout Decline in Canadian Federal Elections: A New Survey of Non-voters


12. Correlates of Personal/Administrative Factors in Turnout

A major correlate of voter turnout is age. We will use once again the division of the sample into age cohorts, outlined in Table 13 and used in Table 14. Table 57 presents the reasons for not voting presented in tables 211, rank-ordered by perceived importance for the sample as a whole, for the age cohort groups.8

Lack of interest stands as the most important reason for the entire sample, followed closely by lack of attraction to any of the parties or candidates. Both of these reasons are ranked high among all of the age groups, but lack of interest is mentioned more prominently in the younger cohorts. Another interest-related item, "didn't care about the issues", also ranks high among all age groups, with 36 percent of the all non-voters sampled rating this item "very important" or "fairly important" as a reason for not voting.

The personal and administrative reasons did not rank high in comparison to others. As we have already seen, not knowing where or when to vote and not being on the list of electors were rated as important factors by 18.7 percent and 17.3 percent of non-voters respectively. This cohort analysis reveals, however, that for the youngest cohort, (i.e. electors newly eligible in 2000), all of the personal/administrative factors with the exception of illness (which affects the oldest groups most) were rated somewhat higher as reasons for not voting. Thus, the cohort entering in 2000 is more likely to have reported being "busy at work" (as is the 1997 cohort) at a higher than average rate. This is true as well for being "out of town", for not knowing where or when to vote, and not being on the list. It is interesting to note as well that the oldest cohort was also likely to report that "not being on the list" was important to them. Although the oldest age group reported relatively few instances of being left off the list, those who did experience a problem in this area felt this was extremely important.


8  In Table 16, the "very important" and "fairly important" responses are combined.



Table 57 Importance of Reasons for Not Voting in the 2000 Election, by Age Cohorts (percentages)


Importance of reason...
(percentage very or fairly important)
(68+) (58 – 67) (48 – 57) (38 – 47) (30 – 37) (25 – 29) (21 – 24) (18 – 20) Total V
Just not interested
31.4
34.0
46.4
50.6
51.8
59.3
57.0
59.1
52.9
0.11
Didn't like parties/candidates
41.7
40.8
56.0
50.9
46.9
43.2
50.7
45.3
47.6
0.08
Vote wouldn't matter
30.6
37.5
47.1
37.9
41.1
36.7
34.3
30.4
37.1
0.14
Didn't care about issues
42.9
28.0
35.7
37.3
36.6
32.8
37.7
36.5
36.0
0.07
Busy at work
16.7
14.3
16.5
24.8
36.9
33.9
38.6
40.9
32.2
0.15
Out of town
19.4
34.7
16.7
19.3
18.3
21.5
25.1
24.8
21.8
0.09
Didn't know where or when
28.6
12.2
12.9
9.4
19.2
24.4
28.5
28.4
21.1
0.15
Not on the list
25.7
16.3
15.5
16.8
16.0
20.3
18.4
24.2
18.7
0.08
Too many elections
26.2
24.5
20.0
18.5
21.4
16.5
13.0
9.5
17.3
0.10
Illness
41.7
20.4
11.9
11.8
8.5
10.7
9.2
10.8
11.7
0.16
N =
35
49
85
161
224
177
207
148
1 086
 



As we saw earlier, only a relatively small number of respondents overall (7.5 percent in total) reported that they experienced a problem with the list of electors, or (9.5 percent) that their names were not on the list. Predictably, such a problem was more likely to be reported by non-voters, suggesting that the list may have had a small effect on the decision to vote for at least some respondents. However, Table 58 shows that problems are more clearly evident among the very youngest cohort, i.e. those who first became eligible to vote in the 2000 election. Nearly a third (31.1 percent) of the non-voters in this group reported that their names were not on the list. The number is also much higher among voters in this age group (14.3 percent) than in any other age category. While there are undoubtedly always going to be problems associated with registering newly eligible voters, the data clearly suggest that this is an area where improvements might be realized.

The data for the penetration of the pre-election TV commercials show that it was very high among all of the age groups. The coverage is as high among younger groups as in the older cohorts, suggesting that the public relations campaign reached the targets it was aiming at.

There is not as distinctive an age pattern with regard to finding out where to vote, although the numbers reporting such a difficulty are slightly higher in the four youngest cohorts than among older groups. The incidence is higher among non-voters, suggesting that the decision to vote or not may have been slightly influenced by the ready availability of such information.

In one section of the survey, respondents (both voters and non-voters) were asked to speculate on the possible effect that the Internet might have on their future behaviour. Specifically, respondents were asked whether they might use the Internet to check their registration information, or perhaps whether they might use the Internet to vote, were such an alternative available.9 Not surprisingly, younger respondents were much more likely to respond positively to these ideas. It is also encouraging to note that the percentages responding positively were high among both voters and non-voters, suggesting that the development of new voting options along these lines might have at least the potential to boost turnout. However, the results should be treated cautiously. It is easy for respondents to give positive answers to these types of questions, particularly if they have a high degree of familiarity with the Internet. We cannot be certain that actual behaviour in an election would follow accordingly, since some of the attitudinal factors that influence not voting would not be ameliorated by Internet voting options.


9 In this analysis, the response categories "very likely" and "fairly likely" were combined into a single category to indicate an overall positive response to the item.



Table 58 Responses to Questions Dealing with Administrative Issues, by Age Cohorts (percentages)


  (68+) (58 – 67) (48 – 57) (38 – 47) (30 – 37) (25 – 29) (21 – 24) (18 – 20) Total V
Experienced a problem with list
V
4.8
4.4
3.0
4.8
7.9
7.6
10.2
9.1
5.6
 
NV
11.4
4.0
8.2
5.6
8.9
8.5
10.6
14.8
9.3
 
All
5.8
4.3
4.8
5.2
8.5
8.2
10.2
14.0
7.5
0.11
Experienced a problem finding out where to vote
V
2.9
3.7
5.4
4.3
6.8
4.4
4.1
4.5
4.8
 
NV
4.8
6.1
8.2
8.0
10.7
15.8
17.4
15.4
12.5
 
All
3.6
3.8
6.0
6.0
9.0
12.3
14.9
14.0
8.9
0.11
Reported name NOT on list
V
0.0
0.7
0.6
1.1
6.8
7.6
2.0
14.3
2.9
 
NV
11.4
2.0
9.5
7.5
12.5
17.5
16.4
31.1
15.1
 
All
2.9
1.1
3.6
4.0
10.2
13.8
13.7
29.4
9.5
0.25
Saw TV commercial
V
62.1
74.8
73.5
76.5
77.4
75.8
83.3
77.3
74.6
 
NV
47.2
59.2
69.0
68.9
69.2
62.5
59.4
63.5
64.2
 
All
58.7
70.7
71.6
73.0
72.8
66.8
63.7
65.3
68.9
0.12
Might use Internet to check registration
V
22.3
42.5
57.2
61.5
65.7
78.0
75.5
50.0
56.6
 
NV
27.8
38.8
35.7
53.1
60.9
55.7
65.2
70.9
57.1
 
All
23.9
41.5
49.8
57.8
63.1
63.4
67.2
68.2
56.9
0.21
Might vote on Internet
V
16.5
35.6
45.5
50.3
54.8
63.7
73.5
50.0
47.0
 
NV
31.4
44.9
41.2
55.3
65.2
59.9
65.7
70.9
59.8
 
All
20.3
38.0
44.2
52.6
60.6
61.2
67.5
68.2
53.9
0.22


There are a number of other socio-demographic correlates with the various items measuring administrative problems with voting. A summary of these is presented here, mentioning only those correlations with a statistical significance of p < .05. In most cases, the correlations reported here are rather weak (below .2 and often below .1), but they are worth considering because of the patterns they represent. In the summaries below, the statistics reported are those appropriate to the measurement level of the variables. In general, Cramer's V is used when one of the variables is nominal in nature, and Tau is used when they are ordinal.

Education: Those of lower education are more likely to report not being on the list (V = .12, p < .001). Subsequently, there is a correlation of V = .08 (p < .01) with encountering a problem making sure the name was on the list. There is no significant correlation with seeing the TV ad, but there is one of V = .10 (p < .01) with finding the ad confusing. There is a substantial correlation TauC = .245 (p < .000) with the likelihood of using the Internet for list modification or registration, and TauC = .19 (p < .000) with the likelihood of using the Internet for voting.

Income: The income correlations are similar to those with education reported above. People of lower income are more likely to report not being on the list (V = .1, p < .01); problems making sure their names were on the list (V = .08, p < .01), and problems finding out where to vote (V = .09, p < .01). As well, lower-income respondents were more likely to find the TV ad confusing (V = .10, p < .01). They would be less likely to use the Internet to modify their information (TauC = .16, p < .001) and vote (TauC = .15, p < .001).

Mother tongue: Those with mother tongues other than French or English are more likely to report their names were not on the list (V = .10, p < .001). In comparison to the overall rate reported in Table 9, where about 10 percent said they were not on the list, and 8.5 percent did not know, among those with "other" languages, 14 percent said they were not on the list, and another 14 percent said they did not know if they were. People of "other" mother tongues also reported more problems in making sure their names were on the list (V = .10, p < .001) and also somewhat more problems finding out where to vote. This group was also less likely to have seen the TV ad (V = .11, p < .001) – 62 percent saw it, as opposed to 71 percent of English speakers and 79 percent of French speakers. They were also more likely to find it confusing when they did see it (V = .08, p < .01). The only statistically significant French-English difference is that French speakers are slightly less likely to report that they would use the Internet for modifying their information (V = .08, p < .01) or voting (V = .08, p < .01).

Born in Canada: Those not born in Canada are more likely to report not knowing where or when to vote (V = .15, p < .000). For 18 percent of non-voters born outside Canada, this was a "very important" factor, and a "fairly important" factor for 19 percent, in comparison to 8 percent and 12 percent, respectively, for those born in Canada. In addition, non-voters born outside Canada were more likely to say that not being on the list (V = .08, p < .01) was important. Of non-voters born outside Canada, 18 percent said not being on the list was "very important" and 4 percent said it was "fairly important"; the comparable figures for those born in Canada are 12 percent and 8 percent. In addition, those born outside Canada were more likely to report a problem finding out where to vote (V = .06, p < .05), and also were less likely to have seen the TV ad (V = .06, p < .05).

Community size: People living in large cities are more likely to report not being on the list (V = .08, p < .001). Table 9 shows that 9.7 percent overall said they were not on the list, whereas the figure for the "large city" residents is 13.8 percent. The only other correlate with community size is related to the Internet. Residents of larger cities say they would be more likely to use the Internet to modify their information or register (TauC = .14, p < .000) and to vote (TauC = .13, p < .000).

Length of residence: There are a number of correlations between mobility, as measured by the length of time people have resided in their current community and administrative variables. Non-voters who have resided for a shorter time in their current community are more likely to say that not knowing where or when to vote was important in 2000 (TauC = .09, p < .001). They are more likely to say they were not on the list of electors (V = .19, p < .000), and also that they experienced problems in making sure that their names were on the list (V = .13, p < .000) and finding out where to vote (V = .13, p < .000). The more mobile are more likely to say they would use the Internet to register or modify the list (TauC = .09, p < .001) and to vote (TauC = .10, p < .001).

Gender: There are few gender relationships. Women are more likely to report not voting because they were ill (V = .14, p < .001), a fact perhaps explained by age. They were slightly less likely to have seen the TV ad (V = .05, p < .05), but also less likely to have been confused by it if they did see it (V = .11, p < .001).