Estimation of Voter Turnout by Age Group at the 38th Federal General Election (June 28, 2004)

Final Report

December 2005

Executive Summary

Voter turnout at the 2004 federal general election continued the downward trend observed since 1993. Research conducted after the 2000 general election shows that this overall phenomenon was largely due to a significant disengagement of Canadian youth. In order to further refine its analyses in this area, Elections Canada opted for a new methodology to measure turnout at the 2004 general election by age group: it cross-referenced actual votes with data from the National Register of Electors to establish voter turnout by age group.

The results show that, among the 18–21½ year-olds who were eligible to vote federally for the first time in 2004, turnout was estimated at 39%. Interestingly, this was 4% higher than the estimated turnout rate for the next group (21½–24 year-olds), although the difference falls below the margin of error for this study. The turnout rate increased steadily with each age group, reaching its highest at 75% among 58–67 year-olds.

Our estimates are based on a representative sample of 553 polling divisions in 31 electoral districts across the country and all advance poll votes as well as the national subset of votes cast by special ballot. The results, based on the Canadian voting age population, have a statistical reliability of ±4.3%, 19 times out of 20. While the new approach does not allow for a direct comparison of our findings with previous figures, it does enable us to set new benchmarks against which to compare future trends.

Background

The past decade has seen a steady decline in voter turnout at Canadian federal general elections. Participation has dropped from a post-war average of 75% to 70% in 1993, 67% in 1997, 64.1% in 2000, and 60.9% in the June 28, 2004, election.¹ Official turnout in Canada is calculated by taking the number of voters registered on the National Register of Electors at the close of an electoral event.

In Canada, as in many other developed democracies, research on electoral participation has confirmed that young people vote less than older people. The Canadian Election Study (CES) – an academic endeavour that has been conducted for every federal general election but one since 1968² – brought this trend to light following the 2000 general election. The CES suggested that the decline in electoral participation in Canada over the past 15 years is chiefly due to an unprecedented drop in turnout among the youngest age groups, combined with generational replacement.³

To further investigate this trend, Elections Canada commissioned a major study in 2002 by professors Jon Pammett and Lawrence LeDuc.⁴ Their research, which drew from an original survey involving equal numbers of self-reported voters and non-voters, found that just over 25% of eligible 18–24 year-olds had voted in the 2000 federal election. It should be noted that this estimate was based on self-reported voting behaviours.

Elections Canada Initiatives

In light of the evidence produced by these and other studies, Elections Canada adopted a multilayered strategy that includes educating and informing young people about the electoral process and their right to vote, ensuring that the electoral process is as accessible as possible, and raising awareness about the decline of youth voting among opinion leaders and stakeholders.

The Chief Electoral Officer of Canada, Jean-Pierre Kingsley, described Elections Canada's initiatives on a number of occasions, including: the Symposium on Electoral Participation in Canada held in March 2003 at Carleton University; the Roundtable on Youth Voting in May 2003 at the University of British Columbia; the National Forum on Youth Voting in October 2003 in Calgary; and the Roundtable on Aboriginal Youth and the Federal Electoral Process in January 2004 in Ottawa.

Elections Canada's initiatives included a new Young Voters Web site designed to reflect its consultations with young people. In February 2004, the Chief Electoral Officer wrote to 1.1 million young Canadians who had turned 18 since the 2000 general election, reminding them of their democratic right to vote. A registration kit was included for those who were not registered. Once the 2004 election was called, Elections Canada wrote again to the remaining 250,000 unregistered youth, informing them about how they could register during the revision period or at the polls.

In preparation for the June 2004 election, 80 returning officers appointed community relations officers to identify neighbourhoods with large concentrations of students, carry out special registration drives, assist youth in finding accessible polling locations, and provide information to community youth leaders and youth media about registration and voting.

During the 2004 election, Elections Canada's mainstream advertising campaign targeted young people with the message that people should "speak up when everyone is listening." Elections Canada also sponsored a Canada Road Trip contest developed by an outside company and promoted on the Web and on the MuchMusic and MusiquePlus television channels.

Other initiatives relied on partnerships with various non-governmental organizations interested in civic education. Examples of such initiatives include:

• partnership with Cable in the Classroom on "Your Vote... Your Voice," a program that challenged students in grades 10–12 across Canada (secondary IV, secondary V and CEGEP in Quebec) to create public service announcements telling their peers why democracy is important and why it is important to vote
• partnership with Rush the Vote, an organization that aims to increase youth voter turnout and political awareness through musical events at which performers encourage voting and democratic involvement
• Student Vote 2004, an innovative educational program that provided students who had not reached voting age with the opportunity to experience the 2004 federal electoral process through a parallel election in their schools
• partnership with the Dominion Institute on Youth Vote 2004, for a youth voter education kit
• four post-secondary student associations, to develop a poster for the 2004 general election

As a result of these and other initiatives, youth participation became one of the most important issues during the 2004 election campaign. While it did not become politicized, youth participation was addressed by a number of political parties and was a major focus of the national media. To our knowledge, this is a first.

Assessing Youth Voter Turnout

A well-known problem encountered by researchers who use survey data to study voter participation is that self-reported turnout is consistently and significantly higher than the official turnout, generally by about 15 to 20%. While corrective measures may be applied, there is no complete solution to the problem. More importantly, it is impossible to know whether or how this tendency may vary with the age of the respondents.

Several factors can explain this over-reporting phenomenon. First, people who are less interested in politics, and who are more likely to be non-voters, are also more likely to refuse to participate in electoral surveys, leading to the under-representation of non-voters and the over-representation of voters in the survey samples.

Another factor identified by survey specialists is the natural human tendency to over-report behaviours that are considered to be positive socially accepted norms – voting appears to be such a behaviour. Although this problem is less likely to occur in telephone than in face-to-face interviews, it exists nonetheless. Researchers have tried different ways of asking survey respondents if they had voted or not, and the results have remained consistently higher than the official turnout rates. Since it is virtually impossible to measure and control the phenomenon, the most effective solutions have consisted of methodological data corrections, such as weighting procedures, in order for the survey results to correspond to official turnout figures.

A New Approach

Given these methodological difficulties, Elections Canada has adopted a new approach to measure turnout across age groups. Exercising an authority granted to the Chief Electoral Officer, Elections Canada used various data gathered for the purpose of administering the electoral process to create a large sample of electors who voted at the 2004 general election. This information was then cross-referenced with National Register of Electors data to obtain a breakdown of voters by age group. This study was designed to strengthen Elections Canada's ability to serve electors and provide public education and information in keeping with section 18 of the Canada Elections Act.

As part of the methodological approach, we selected a national sample of 553 polling divisions in 31 electoral districts across all provinces and territories. It provided estimates with a statistical reliability of ±4.3%, 19 times out of 20, when generalized to the entire Canadian voting age population.⁵ Because there was also an interest in estimates of turnout at the provincial level, the original sample was increased by another 139 polling divisions in order to obtain provincial/territorial estimates with a reasonable statistical reliability. In this report, estimates at the national level are based solely on the original sample, while those at the provincial level are based on the augmented sample. Further details about the methodology for this study are provided in the Appendix.

Observations

National Highlights

Figure 1 shows the estimates for turnout rates by age group at the national level. These estimates are calculated by comparing the number of voters to the estimated voting age population for each age group (citizens only).

The general trend consists of a linear relationship between age and participation, consistent with the patterns traditionally observed using other research methods. However, the results showed one interesting exception to this relationship: for the youngest group of electors (18–21½ year-olds) for whom June 28, 2004, was the first federal general election in which they were eligible to vote, the turnout rate was four points higher than for the next oldest group (21½–24 year-olds), that is, those who had been eligible to vote for the first time in 2000. Yet it is worth noting that this difference falls below the margin of error for the national sample, which means that the gap between the two estimates could be explained by the sampling error and does not necessarily indicate a true difference between the estimates.

Among the older age brackets, voter turnout increased steadily with each group, reaching a high of 75% among 58–67 year-olds and declining slightly to 71% in the oldest group. Again, this pattern is consistent with what is traditionally reflected in public opinion surveys.

In addition, at the national level, only the three oldest age groups (people 48 years of age and older) boasted a participation rate higher than the average.

Provincial Highlights

The following table presents the turnout rates by age group for each province and territory. Overall turnout based on the population of Canadian citizens 18 years of age and older for each province is featured in the last column on the far right.

In absolute numbers, among first-time eligible electors, those from Prince Edward Island had the highest participation rate (51%), followed by Quebec (47%), Ontario (42%) and New Brunswick (41%). Youth participation rates were generally consistent with the overall turnout for each province and territory.

Figure 2 below provides a graphic comparison of the two youngest age groups. The greatest deviation is found in New Brunswick, where a difference of 10% separates the youngest cohort from the second youngest. That province is followed by Prince Edward Island (8%), Quebec (7%), Alberta (6%) and Ontario (5%).

In the remaining provinces and territories, particularly British Columbia, Nunavut and Yukon, the trend between the two groups is reversed and falls back in line with the traditional linear relationship between age and turnout. It should be noted, however, that any comparison must take into account the margins of error reported for the geographical units of interest (see Table 5 of the Appendix).

(Based on citizens in the voting age population)

[ Previous ]    [ Next ]