Purpose:
Political participation is on the decline in Canada. For every election, fewer and fewer eligible voters are showing up to vote. Voter apathy is a very serious
issue, because as fewer people chose to have their say in politics, the less democratic our country becomes, as only a minority will actually be represented in parliament. The purpose of my research project is to determine what causes voter apathy and why political participation is decreasing.
Theory/Hypothesis:
I believe that feelings of apathy or disinterest in the election are not just a reflection of attitudes, such as politics being too complicated or ordinary people not feeling that they have a say, or thinking elected officials lose touch with the people anyway. I believe that there are actually certain types of people, or people with certain groups of characteristics, who do not care about political issues. Specifically, I think that age, education level, income level, and language, could all possibly explain why people do not vote. I believe that younger eligible voters will be less like to vote because of the logistics of finding out where they vote may be difficult as this age group is more transient. As well, senior citizen voters may be less likely to vote because it may difficult for some of them to physically get to a polling station, or other such disabilities. In terms of education level, I think that more educated people will be more likely to vote because they will feel less insecure in their knowledge of the issues because they were taught critical thinking skills. I also think that higher income earners will be more likely to vote because they probably have a higher education and because issues such as taxes could have a very big impact on their incomes. In the instance of language, I think that French speaking Canadians will be more likely to vote than English speaking Canadians because the issue of national sovereignty and protecting French language laws are so important to them, which is a major political issue and could have a very direct impact on their lives.
Literature Review:
Others who have researched the topic of interest in elections have come up with findings similar to, and in addition to, my own explanations for variation in this variable. During the 2000 Canadian elections campaign, Maclean Hunter Canadian Publishing Ltd (2000) conducted a study of how voters viewed the campaign. The researchers found that one third of the respondents claimed that nothing in the campaign grabbed them. Respondents attributed this to the fact that none of the leaders had any new ideas. Some respondents believe that only half of Canadians followed the election campaign. Because the government is so removed from the potential voters that there is no point in voting because people do not feel that they have a say in politics. Others said that they do not bother following because campaign promises are never fulfilled anyway. This means that, according to this study, it is attitudes towards politics that affect one's interest in the campaign.
An analysis study of the 1997 Canadian Election Campaign by Allan Fotheringham (1997) also found that lack of interest in the election could be due to a lack of interesting or important events in the campaign. The author claims that the highlights of the campaign were the candidates' hair dos, dental work, sex lives, and bad photo-ops. Fotheringham presents his observations in a way that make the election campaign's highlights worthy of interest purely for gossip, not for real political news that merits the voters' interest.
An article in America (1999) also emphasized people's lack of interest as being a result of simplistic media campaigns that focus on scandals rather than the facts. But the author claims that complex political issues seem to be over the heads of most voters. Both of these reasons could account for why people choose not to pay attention to media coverage of campaigns. The author also suggests, though, that the real cause of voter apathy could just be voter apathy, meaning that voters are sufficiently satisfied with the status quo, so there is no reason to get involved.
However, a study by Bruce E. Pinkleton, Erica Weintraub Austin, and Kristine K. Fortman ( 1998) found that people' s negative reactions to media campaign
Coverage do not affect one' s interest in the election or one' s likelihood of voting. Instead, voter turnout is associated with one's perception of efficacy, indicating that it is one's attitudes that affect his/her decision to vote or interest in the campaign. This study also shows that my indicators of voter apathy were good, as it found that most respondents base their voting decisions on what they see in the media. Therefore, lack of attention paid to different forms of media coverage of a campaign is a good indicator of voter apathy.
In another study conducted by C. Richard Hofsetter and Christopher L. Gianos (1997), the authors believe that variance in interest in the election is more due to
demographic gaps in certain social groupings. The study found that people who listened to political talk radio were more likely to follow politics in other media forms and have an interest in politics in general. This finding is consistent with my index on interest in the elections, which suggested that people who followed the election did so through different forms of media and there was a strong correlation between people who paid attention to the election media coverage and those who are interested in politics in general. The study found that more educated people paid more attention to talk radio, which is consistent with my belief that more educated people have a greater interest in politics because they have had the opportunity to develop their critical thinking skills. Very few respondents indicated that they agreed with the radio talk show host most of the time, which shows critical thinking skills. The study also found that listeners had higher incomes, but that age, race, and marital status caused no statistically significant differences in who listened and who did not.
Norbert Schwartz and Howard Shuman (1997) took another approach to explaining variance in interest in politics. They say that when collecting data on one's interest in politics, it is important to consider structure and the questions of the survey. They found that when people were first asked to rate their political interest and then given a question to test their knowledge of politics, more people rated themselves as more interested in politics. However, if the question to test the respondent's knowledge of politics came before the question of interest, and the respondent had trouble answering the question, he/she was then more likely to rate his/her interest in politics as low.
Mark Strama (1998), in his study of youth electoral participation, provided other possible explanations for why younger voters may be less likely to be interested in elections. Strama found that people between the ages of 18 to 24 chose to vote less than other age groups. The author does suggest that low interest in elections among youth may be attributed to the fact that they move residences and, therefore, will have less of a stake in the local politics where they vote. However, he also offers that low interest among youth could be because the main election issues, such as taxes, do not really apply to them and because politicians do not target their ad campaigns at younger voters.
Method:
In order to do my research, I will be using secondary data. This means that I will be using data that have already been collected, in my case the 1997 Canadian Elections Survey. This survey was conducted prior to the 1997 federal election, where 2000 Canadians were interviewed by telephone. Using this data set as my basis for my observations has its advantages and disadvantages. Firstly, by using an existing data set, I am saving an enormous amount of time and money because the work in collecting the data has already been done. As well, the fact that this particular data set is available on an online server, which is connected to SPSS means that I am able to quickly recode and analyze the data. However, the problem with using this method of research is that the questions may not always be worded exactly the way that will be necessary for my own interests when reanalyzing the data, and some questions may have been left out. As with surveys generally, I cannot go back to my respondents for clarification of their answer or to redirect the question, which means I have to work with what is there. Therefore, the analyses I do of the data set may not be completely valid because they may not measure exactly what I want.
I will have to take several steps using Webstats to test my hypotheses about the causes of voter apathy. To begin, I will choose questions that I see as indicators of voter apathy so that I can create a new variable with which to test what other ariables affect it. By choosing questions CPSB 1, CPSB2, CPSB3, CPSB4, and CPSB5, and declaring the 'Don't know' and 'refused' responses as missing, I created a new variable called "Interest", referring to how interested one is in the 1997 election. These question asked respondents how much attention they paid to information about the election on TV, the radio, the newspaper, and how interested they were in this specific election, as well as in politics in general. But before I could choose all of those questions as indicators, I had to create an index using all of those questions to ensure that they were all in fact related. According to my alpha, which was quite high, I found that all of these indicators were strongly related, so I did not have to take any of those questions out as indicators of my variable. I also chose to declare the 'don't know' and 'refused' responses as missing because there are so few of them, and they would wreck the inherent order of the variable, causing it to no longer be ordinal.
Once I have my new variable that ranks how interested people are in the 1997 election based on the indicators that I selected and found to be related, I can do cross-tabulations to see which independent variables are in fact related to differences in one's interest in the election. By looking at the measure of significance, I can determine which independent variables are related, and I can also rule out which ones are not. So first I, have to see if people's attitudes about politics affect their interest in the election. To do this, I have to run cross-tabulations with the three questions that I thought might have an effect on voter apathy and can perhaps be combined into an index. The three questions I chose were CPSB1OC, CPSB1OA, and CPSB 1OB, which ask respondents to relate how strongly they agree with the statements that government is too complicated, that political leaders lose touch with the people, and, that people like them do not have a say in politics. These questions are all ordinal in measurement level as they have an inherent order in the responses when I exclude those who responded 'don't know' and 'refused'. So since both variables being considered are ordinal and form rectangular tables, I can use tau-c as my measure of association. In order for me to claim that these independent variables do affect interest in the election, the measure of association will have to reach statistical significance, preferably with a magnitude above 0.2.
When I have completed my cross-tabulations to see if there is a correlation between attitudes towards politics and voter apathy and I can move on the second part of my hypothesis, which includes testing to see if there is a relationship between voter apathy and certain groups mentioned above. I will perform the same steps as when I checked for correlation between interest in the election and attitudes towards politics. Each of the groups mentioned above will be used as my independent variables. The first one, age, will have to be collapsed because there are a large number of options. I will break these up into three groups (18-30,31-64, and 65-105) because those groups represent younger voters, middle-aged voters, and senior citizen voters. To do this, I will have to compute a new variable with CPSAGE because the question asks what year the respondent was born in, and I want to know how old the respondent is, so I will have to compute 1997-CPSAGE to get the respondents' age instead of the year they were born. I will also declare the 'don't know' and 'refused' responses as missing so that this variable as recoded will be ordinal. It is not interval because there are groupings of ages instead of an equal distance between each possible response. When I cross tabulate this variable with my dependent variable of interest in the election, I will have a square ordinal table, so I can use tau-b as my measure of association. Once again, in order to claim a substantial relationship between the two variables, my measure of association will have to be above 0.2. In order to ensure that any observed variation in which age group is most interested in the election is not as a result of chance, I will have to calculate the statistical significance. To do this, I will tell the computer to calculate chi-square when I do my crosstabs of these two variables. If the probability associated with the chi-square is below .05, then the variation in responses is considered statistically significant because there is less than a 5% chance that it is due to chance because the chi-square score compares the result of my table to a table where there is no relationship between the different groups.
To test if one's education level is related to one's interest in the election, I will use the question about education level, CPSM3 as my independent variable. There are, once again, several possible responses, so I will collapse this variable into fewer categories. I will place all those who responded to less than high school education into the low category .I will label all respondents who completed high school as having a moderate education. All of the respondents who have some or completed college or university will be considered to have a high education level. In order to make this variable ordinal, I will declare the 'don't know' and 'refused' responses as missing. Once I cross-tabulate this with my dependent variable of interest in the election, I will have an ordinal by ordinal square table, so I will use tau-b to determine if there is a strong enough relationship between the two variables to consider them correlated. I will again have to test for statistical significance to deal with the possibility that my demographic gaps may have occurred due to chance. To see if the difference across my different education categories is due to chance, I will run a chi-square test or a one way anova test. If my statistical significance is less than .05, I can say that there is less than a 5% chance that the explained variation was due to chance.
Next I will have to test to see if income groups have an effect on voter apathy. There are two questions in the survey related to income that I will combine into a single measure. Since there are so many choices in this question, I will make it ordinal by placing respondents into groups of low, medium, and high income earners, and by eliminating the 'don't know' and 'refused 'responses. Those who earn less than $40000 a year will be considered as low income earners. The respondents who earn between $40001 and $80000 a year will be considered medium income earners. Those who earn more than $80 001 a year will be placed into the high-income earners category. I will declare the 'don't know' and 'refused' responses as missing. When I cross-tabulate this independent variable with interest in the election, I will have an ordinal by ordinal square table again, so tau-b will be my measure of association, and it should be above 0.2 for me to consider that the independent variable or income group has a substantial effect on voter apathy. I will use either chi-square or a one-way anova test again to ensure that the variation between my income groups' interest in the election is not due to chance.
My final test will be to determine if one's language usually spoken affects how interested one is in voting in the election. Using CPSM16, I will collapse the responses into three categories: English, French, and Other. I will declare the 'don't know' and 'refused' responses as missing because there are so few of them and they do not fit in with the new categories I have created. This variable is different from the previous independent variables because it is nominal, meaning that the categories are mutually exclusive and there is no inherent order in the possible responses. This means that when I do my cross-tabulation to determine whether my measure of association is enough to consider this variable as related to my dependent variable of interest in this election, I will have to use a different measure of association. Since my table will be nominal by ordinal, I will have to use Cramer's Vas my measure of association, but the number I use is still 0.2. There are, however, still three groups in my independent variable, so to measure for statistical significance, I will still use a chi-square or oneway anova test.
Analysis:
My analysis component will have two goals. The first will be to describe what I find after I have run my tests in SPSS. The idea is to point out statistical variation to determine which groups of people are more likely to be interested in the election. The next step of my analysis will be to explain why certain groups have different levels of interest and to relate my findings to the literature.
Fotheringham, Allan. "Notes from the Most Boring Election Ever". In: Maclean's,
no5. Canada: Maclean Hunter Canadian Publishing Ltd., June 2, 1997. p80.
C. Richard Hofsetter and Christopher L. Gianos. "Political Talk Radio: Action
Speak Louder than Words". In: Journal of Broadcasting and Electronic Media, Vo141,
no 4. California: Broadcasting Education Association, Fall 1997. p501-516.
Bruce E. Pinkleton; Erica Weintraub Austin; and Kristine K. J. Fortman.
"Relationships of Media Use and Political Disaffection to Political Efficacy and Voting
Behaviour". In: Journal of Broadcasting and Electronic Media, Vo142, no 1.
California: Broadcasting Education Association, Winter 1998. p34-50.
Nobert Schwarz and Howard Shuman. "Politcal Knowledge, Attribution, and
Inferred Interest in Politics: The operation of buffer items". In: International Journal
of Public Opinion Research, Vo119, no 2. UK: Oxford University Press, Summer
1997. p191-196.
Strama, Mark. "Overcoming Cynicism: Youth participation and electoral
politics". In: National Civic Review, Vo187, no 1. US: National Civic League Inc., Spring 1998.
p71-78. US: National Civic League Inc.,Spring
Babara Wickens; John Demott; Brenda Branswell; Susan McClelland; Brian Bergman;
and Ken MacQueen. "Looking On With Cynicism". In: Maclean's. Canada:
Maclean Hunter Canadian Publishing Ltd., December 4, 2000. p29.
"Of Many Things". In: America, Vo1180, issue 10. US: American Press Inc., 1999.
p2