by Brandan McCammitt
April 2015
Negativity in politics has long been seen as the cause of everything wrong with politics, but what if it wasn’t? What if voters actually were motivated by negative campaigns that presented a contrast to them over the differences in the candidates so that it became clearer why each candidate was running for office? What if negative political campaigns actually increased voter turnout? It is my purpose in writing this paper to further the discussion on this topic and to increase the information on this topic within the political science community. It is my belief that if policy informative statements are made with a negative connotation, then they will cause an increase in voter turnout in the election between two or more candidates and their campaigns.
My hypothesis is that when negative campaign messages are based in policy and informative contexts then voter turnout will increase. My key independent variables are negative campaign messages and whether the messages or advertisements are policy informative. I intend to define political campaign messages as negative if they are negative in tone or connotation toward an opposing candidate or campaign. Also I will define messages or ads that are policy informative as a political campaign’s message which is specifically focused on a policy issue and that the majority of the message is informative in nature and not based on a character or personal attack on the opposing candidate or campaign. My hypothesis’ dependent variable is voter turnout. I will define voter turnout as the number of eligible voters who turn out to vote in an election. These three parts will make up my research characteristics and equation.
It is my belief that candidates and their campaigns will not have vastly divided messages in their efforts to win elections. I will be embracing the belief that there is a correlation between what a candidate and their campaign would put up on the air waves in campaign-funded television advertisements, what they would say in their speeches, what they would say on their website, and what they would say in their closing statements in a debate. This belief allows me to observe a different type of message than other researchers have observed while still using their understandings of results from negativity in political campaigns. It also allows a view of what candidates are saying to voters when it is actually them saying it and not a television advertisement, what part of their messages they are willing to say directly from their lips to voters.
Conventional Wisdom
Through reading multiple sources, I have found the conventional wisdom on negative campaign messages and their correlation to voter turnout to be somewhat of a mixed result. Many of the sources have stated that the general view of political scholars is that negative campaign ads and their messages hurt voter turnout, and that it results in a damaged political system (Krupnikov 796), even though nearly all of their results have said that when the negative campaign message is informative and shows how the candidates are truly different it results in an increase in voter turnout (Franz 108). So, for my paper, I am viewing the contemporary conventional wisdom on negative campaign messages and advertisements to be different from the previous trend. My understanding of the conventional wisdom is that negative and informative policy based campaign advertisements and messages increase voter turnout.
Building the belief that negative and informative policy campaign messages increase rather than decrease voter turnout is the book Campaign Advertising and American Democracy. In this book, the authors Franz, Freedman, Goldstein and Ridout come to the conclusion that campaign ads are helpful to increasing voter turnout (Franz 108), but also the fact that negative contrasting policy ads increase the knowledge of most voters about the election and candidates running for office (Franz 125). Then, Goldstein and Freedman’s article Campaign Advertising and Voter Turnout: New Evidence for a Stimulation Effect shows that, in fact, voter turnout is “stimulated” by negative advertisements and that these ads help bring people to the polls (Goldstein 722). The authors further found that positive ads had no real result, but that negative ads actually increased voter mobilization by “a significant and substantial” result (Goldstein 733).
Three other articles along with Franz’s had a major piece of research showing the importance of contrast in advertisements. Jackson, Mondak and Huckfeldt found that contrast ads were actually good for voters in helping them become more knowledgeable and more energized to vote (Jackson 61). Leighley and Nagler found that the contrast factor “perceived policy differences between the candidates lead to increases in turnout” (Leighley 131). Finally, Martin found that the contrast factor in negative ads was found to increase anxiety and fear of candidates, strengthening voter energy (Martin 554). All four groups found strong evidence supporting the belief that contrast is an important point in political advertisements.
Along with the need for negativity, contrast and policy information in campaign advertisements there also is the need to be aware of one’s own campaign: who they are, and who they are running against. Robert Hogan, in his article Campaign Spending and Voter Participation in State Legislative Elections, presents his research on how spending affects voter turnout. Hogan also revealed, along with a lot of information on campaign spending, that he found proof that voter turnout is positively affected by incumbents running for reelection (Hogan 858) and that when voters were contacted by state legislative campaigns it encouraged them to vote for the down ballot candidate (Hogan 859). These observations point to the fact that if voters have more knowledge about candidates and their campaigns it will help them to better make their choice, or even just a choice, on Election Day.
In reading an article by Yanna Krupnikov titled When Does Negativity Demobilize? Tracing the Conditional Effect of Negative Campaigning on Voter Turnout, we are able to learn that also the timing of campaign messages matters in that “late negativity about a person’s preferred candidate has a statistically significant, negative effect on turnout” (Krupnikov 804). The author believes this shows that voters can be persuaded to either change candidates late or convince them that both candidates are not worth their time (Krupnikov 809). The fact that these debates are late in the campaign season suggests that it may be more meaningful what a candidate will say because it could be the last time people hear them and their message.
In Jackson, Mondak and Huckfeldt’s article, Examining the Possible Corrosive Impact of Negative Advertising on Citizens’ Attitudes Toward Politics, they found “no support whatsoever” for the idea that exposure to negative campaign ads would decrease people’s beliefs in politics as a whole (Jackson 61). In Kahn and Kenney’s Do Negative Campaigns Mobilize or Suppress Turnout? Clarifying the Relationship between Negativity and Participation, they found that “negative information significantly enhances turnout” and that mudslinging hurts turnout (Kahn 883); negative information influences voters, but differently, based on their own personal political identity and knowledge (Kahn 885). They also found that voters were able to identify negative messages as appropriate or inappropriate (Kahn 887). In Lau and Pomper’s article Effects of Negative Campaigning on Turnout in U.S. Senate Elections, 1988-1998, they found evidence that says negative tone does affect voter turnout and it affects it in a positive way, causing an increase in voter turnout (Lau 811). They also found that issue-based negativism “does indeed tend to stimulate turnout” while person-based negativism “has a much smaller effect on turnout that is also positive” (Lau 812). All three of these articles found evidence validating each other’s arguments in favor of negative campaign advertisements being good for voter turnout.
In Leighley and Nagler’s book Who Votes Now? they found evidence that the more education someone has, the more likely they are to vote (Leighley 60). In Paul Martin’s article Inside The Black Box Of Negative Campaign Effects: Three Reasons Why Negative Campaigns Mobilize, he finds that negative ads were seen to make certain voters believe a race was closer and therefore more worthy of them spending their time to vote on it (Martin 556), so it increases the competitiveness factor in elections. In Raphaël Soubeyran’s the Contest with Attack and Defense: Does Negative Campaigning Increase or Decrease Voter Turnout?, evidence shows that “voters tend to vote more when candidates use both attack and defense” ads in a campaign (Soubeyran 348) and that it is easier for the opposition to attack than the incumbent because of a lack of record (Soubeyran 350).
All of these articles and books present one resounding answer to the main question of the effect of negative ads on voter turnout. It would seem that negativity in campaign messages and advertisements increases voter turnout. Further, a majority of the sources agree that there is a limit to that point, in that mudslinging will hurt voter turnout and that if the negativity presents a contrast it will help voters even more to make a choice in the election, which will increase voter turnout.
Research Design
My hypothesis is that when negative campaign messages are based in policy and informative contexts then voter turnout will increase. For my research, I will be reviewing 2010 toss-up U.S. Senate races by comparing debate closing statements to see how negative they are, and how informative they are to the voter turnout rate in the state on Election Day. This research was conducted through a comparison of The Real Clear Politics Election 2010 Battle for the Senate final state poll ratings, C-SPAN Debate footage from the C-SPAN Campaign 2010 Series, and Voter Turnout Results from the United States Election Project’s 2010 November General Election Turnout Rates page.
The reason I used 2010 Senate race data is because I was looking at non-national elections. At the time of my research, the 2014 midterm elections had just occurred so voter turnout numbers were either not present or were not fully complete, and therefore would not be fully reliable. The reason I did not use 2012 Senate races is because it is a more national election atmosphere and the national attention is being paid to the Presidential race. The issues and campaign tone are all being set by the national campaigns rather than the individual Senate campaigns, so there will be less variety. Also, national elections always have higher voter turnout rates than midterms, and the dynamics as a whole are different than in a midterm election cycle.
In order to keep my research testing within somewhat of a controlled setting, I will only use Senate Races that are ruled to be toss-ups from the 2010 Election Cycle, so it is clear that the competitiveness factor is already present in the races, and that there is some attention being paid to it already in the media and by the public at large. Despite having two independent variables, I only have a score for one, that is because when doing the research every candidates’ remarks when they were addressing their opponent were policy informative, so I did not include a score scale or a score in general for policy informativeness. As well, it was a 50/50 ratio for candidate remarks about themselves being positive personal and positive policy remarks.
Further, as a review of my subjects, I can see that two of the Senate contests I am using had Governor Races that were considered to be toss-ups (Colorado and Illinois), three others just had general non-toss-up Governor Races (California, Nevada and Pennsylvania), and the two others had no other major state-wide races (Washington and West Virginia). Every state had only one U.S. Senate election in 2010 in the November election, but two states (West Virginia and Illinois) had the stipulation that the winner of the election would take the seat immediately, so those races were both General Senate Elections and Special Senate Elections. Every single one of the U.S. House of Representatives seats within each of the seven states tested was up for election on the same day as the U.S. Senate Election was being held, also some of the states had other state-wide elections and referendums and local races.
I will be testing the negativity and informativeness factor of my question by using a five point scale graded by myself of the candidates’ comments during their closing statements at a late campaign season debate in their election cycle. The five point scale will be based on the number of references to the other candidate and their policy beliefs with negative policy connotations in their statements. The scale I will be following for my research will see a zero (0) as having no negativity at all in the statements of either candidate at the debate. A one (1) will symbolize one to three negative statements by either or both candidates combined during the debate, but not filling a majority of their time. A two (2) will be a 50/50 negative to positive statements by one or both of the candidates involved. A three (3) will be a mostly negative statement by one or both of the candidates at the debate. Finally, a four (4) will signify an entirely negative policy-based closing statement by one or both candidates involved in the debate. To signify that one or both of the candidates in their closing statements use mudslinging and personal attacks against their opponent at the debate, I will use a five (5) on my negativity scale.
What I am considering to be a policy informative statement that is negative is something like this statement from the California Senate Debate from Sen. Boxer: “My opponent is fighting for tax breaks for Billionaires and Millionaires and these companies who ship jobs overseas.” Something that would be uninformative or personal-based would be a statement like this from that same California Debate from Mrs. Fiorina: “She has become a multi-millionaire while she’s in Washington, DC.” Looking at positive examples, I would consider “I believe in you and I’m asking you to believe in me” from Gov. Manchin in the West Virginia Debate to be a positive personal statement. I will be considering this to be a positive policy informative statement: “We have the Bush tax cuts out there which I feel are very important, we need to make them permanent,” from Mr. Raese in the West Virginia Debate.
I am using a scale rather than the raw time of the statement that is negative or a timing percentage because the raw time or raw percentages could be skewed because not all of the closing statements were the same time. Some candidates went under or over the time they were allowed, and some spent time, rather than talking about issues, to thank the presenters and stations or to remind voters about when to vote. All of these, while important, can be taken into account better by a scale rather than just looking at the timing of the statement. There is no way in a timing percentage to show personal attacks apart from policy informative like I can through a specific mudslinging number on the scale.
I will also be testing the voter turnout factor of my question by using a 7 point scale graded by the amount of voters who turned out to vote in the election that year. The scale I will be using will see a zero (0) score as turnout in the range of 0 to 30%. A one (1) will symbolize voter turnout in the range of a 30.1 to 35%. A two (2) will symbolize voter turnout in the range of 35.1 to 40%. A three (3) will symbolize voter turnout in the range of 40.1 to 45%. A four (4) will symbolize voter turnout in the range of 45.1 to 50%. A five (5) will symbolize voter turnout in the range of 50.1 to 55%. A six (6) will symbolize voter turnout in the range of 55.1 to 60%. A seven (7) will symbolize voter turnout in the range of 60.1 to 100%. The voter turnout numbers I will be using are made up from the voting eligible population and all total ballots counted in the state. The reason I scale these is so that the percentages can be put better into a formula for a final result.
After conducting the testing of these two areas, I will be comparing the numbers together. If my hypothesis is correct, that negative and informative campaign messages increase voter turnout, then the numbers should be similar to each other. By similar, I mean the margin of error would be within one number length of each other on either side of the scale. I intend to test this by subtracting the negativity number by the voter turnout number for an equation of Negativity Score – Voter Turnout Score = Results, or in simplified form N-V=R. For instance if a senate race had a 5 on the negativity scale, but a turnout rate of a 3, then their score would be a 2, meaning that then the negativity did not help voter turnout. If my hypothesis is correct, the final numbers should be close to if not dead on 0. The more off they are, the closer they will edge to 5 or -5. If my hypothesis is wrong, that positivity helps voter turnout, then the average number score will be a more negative number. If my hypothesis is wrong, in that negativity did not help voter turnout, then the average number score should be a higher positive number.
After Tabulating the Final sums for all seven U.S. Senate races, I intend to separate them up into Senate Races with an Incumbent and Open Senate Races. I will do this to see if there is a significant difference with an Incumbent on the ballot or not. Then I intend to separate the Senate Races with a Governor Race on the ballot and those without. I will do this to see if there is a significant difference between states with or without a Governor race. Finally, I will separate the toss-up and non- toss-up Governor Race, to see if either of those groups have a significant difference. By separating these races, I intend to see if there are any other effects that could be playing into the toss-up U.S. Senate Races.
In my research there were other possible variables, such as: how different the candidates’ policy positions were (in that more centrist candidates or more moderate candidates could affect voter turnout), the race and gender of candidates, and the type of media through which a candidate transfers their message to the voters. My intention in this paper was to look at negativity and its effects on voter turnout. While there are many other possible variables that are important I will not be observing them or reviewing their effects on voter turnout in this paper.
Results
After conducting my research on negativity and informativeness, as well as its comparison to voter turnout in the seven toss-up U.S. Senate races held during the 2010 Midterm elections, I found that my research is inconclusive. According to my results, I found that my hypothesis was incorrect in that negativity did not help to increase voter turnout, but because none of the U.S. Senate races except for one which I tested had a highly negative closing argument section, I do not have enough evidence to prove or disprove my thesis. For all the races I tested, I did learn that there was not a lot of negativity in the debate closing statements. Voter turnout was above it, but on a relatively low scale. I learned that in the race characteristics portion there did seem to be some differences.
When looking at the U.S. Senate races divided by those with an incumbent seeking re-election and those races that were open, I learned that the races with an Incumbent running were more negative, and that turnout was in fact higher. The negativity scale averages showed that Incumbent involved races had a higher negativity by 1.25 pts on my negativity scale. The voter turnout rates were also higher in those races that saw incumbents seeking re-election; they were just under 10% higher in actual voter turnout and just under 2pts higher on my voter turnout scale. On the total score field when an Incumbent was seeking re-election, it was -0.64 pts higher in my scale. This area shows that, in comparison, more negativity did show an increase in voter turnout when an incumbent sought re-election compared with an open Senate race.
When looking at the U.S. Senate races separated by having or not having a Governor’s race on the ballot, I learned that when there was also a Governor’s race on the ballot, the U.S. Senate Debate closing statements were a point more negative according to my scale. There was less than a two point difference in actual voting percentage averages, and on my scale that equated to a difference of .6 higher voter turnout when a Governor’s race appeared on the ballot in the state. This, once again, would show that in comparison there is a slight difference in favor of negative closing statement in comparison of Governor’s race to Non-Governor’s race on the ballot.
When you look at the U.S. Senate races separated by having a Governor’s race and either being a toss-up or a non-toss-up, I learned that there were only minor differences in comparison to each other. If the Governor’s race was a toss-up, the U.S. Senate Debates tended to be less negative by just over a point and a half (1.667 pts) on my scaling mechanism for negativity over those races that did not have a toss-up Governor’s race. As for Voter Turnout rates, though, the Governor’s Races that were not toss-ups brought more people to the polls by around .667 on my scaling mechanism, but by just about 4.2% in the actual voter percentage average. On the final score count, non-toss-up Governor’s races showed to be just a few points closer to a perfect score than did a state with a toss-up Governor’s race.
After looking at the differences in each voting area, there is some evidence to show in favor of my hypothesis that negative informative policy messages do affect voter turnout in a positive manner. If there is a Governor’s race or an incumbent Senator seeking re-election, my hypothesis holds true. In comparison, between toss-up and non-toss-up governor’s races, my hypothesis does not appear to hold up, and in the full charting it appears my hypothesis does not hold up to the testing of my experiment on 2010 toss-up U.S. Senate Elections.
Due in part to the lack of substantial evidence on my overall question and the lack of multiple highly negative or positive races with higher levels of turnout, I am ruling my research to be inconclusive, with the understanding that on the characteristic surveys there is some evidence that could be taken from my research to prove my hypothesis correct and incorrect.
Conclusion
In doing this research, I made a correlation which I had not seen practiced in any research surveys conducted prior to my own: a correlation in which there is a connection made between a non-television advertisement type of campaign messaging and negativity in policy information as the means by which voter turnout increases. After conducting my research and discovering it to be inconclusive, I believe it is clear that more research should be done on this area of campaign politics in order to see if it was just a fluke year, or if it is in a similar connection to campaign advertisements and their negativity in policy information being a positive for voter turnout, or if it has no connection at all to voter turnout.
I believe that, in the future, it would be helpful to review not just one segment of the states up in an election, as I did with toss-up Senate races in the 2010 Election, but rather to test the entire field of states and seats up in a particular election cycle. This would make for a fuller study to take note of every debate in the general election campaign and its closing statements. This would show if there is a continued line of connection to negativity in policy-based contrast messaging, and if it is a cause of increasing voter turnout. Looking into all of a campaign’s messaging would be a good next step as well, if someone were to try and take the research I conducted here further by observing the negativity in campaign messaging and testing to see if ads, speeches, websites and debate remarks are similar in their amount of negativity or different.
One area of campaign politics where this type of research could present a lot of new data would be looking into whether primary campaigns or general election campaigns are more negative in their messaging. This would also be a means to see if the turn in polarization in national politics is also affecting party politics and local state politics.
It is also worth noting that this research was conducted in a midterm election cycle. This means that this information could change in a presidential election cycle with a more nationalized campaign atmosphere and the national candidates for the parties setting the debate on national issues, and setting the tone of the campaigns from their national headquarters. This research could change in off-year elections, or even in respect to Governor’s races. Having not looked into the possible side effects of negativity in the Governor’s races means there could be a major area of information that I did not tap into for my research paper. There are still many unanswered questions into how Governor’s races can interfere with Senate races and how Senate races can interfere with Governor’s races.
I would also like to state that all of the data I have, and all of the research on this project was conducted on the 2010 U.S. Mid-Term Elections. In those elections there was a larger than normal level of voters who chose not to turn out to vote, which means that the research I did was possibly on a fluke election cycle. The 2010 Election cycle saw many different events occurring in it compared to an average Mid-Term Election year, which I wanted to have noted.
I conducted all of this research in the interest of knowing more about how campaign messaging works. Specifically, I wanted to know more about negative and policy informative messaging and how they can affect voter turnout in elections. Elections are very important, in that they allow voters to choose who they want to represent them in the halls of government and who they want to lead them. Living in a democracy requires the participation of the citizenry, and those people who are working in politics and in campaigns are often seeking ways to energize voters and to get them to turn out and express their beliefs, so that the country can be led by the majority’s will. Only through research like this, the research cited within this article and other groups of research into the different aspects of campaign politics, will those people involved in elections be able to know the best ways to get more voters to the polls.
This Table contains the scores of each debate and each states voter turnout with a breakdown of each specific sub-grouping under research from the 2010 Senate Elections.
Appendix A
California Senate Debate: http://www.c-span.org/video/?295735-1/california-senate-debate
Colorado Senate Debate: http://www.c-span.org/video/?296189-1/colorado-senate-debate
Illinois Senate Debate: http://www.c-span.org/video/?296086-1/illinois-senate-debate
Nevada Senate Debate: http://www.c-span.org/video/?296018-1/nevada-senate-debate
Pennsylvania Senate Debate: http://www.c-span.org/video/?296168-1/pennsylvania-senate-debate
Washington Senate Debate: http://www.c-span.org/video/?296056-1/washington-senate-debate
West Virginia Senate Debate: http://www.c-span.org/video/?296062-1/west-virginia-senate-debate
Voter Turnout Data Source: http://www.electproject.org/2010g
Race Ratings Source: http://www.realclearpolitics.com/epolls/2010/senate/ 2010_elections_senate _map.html
Governor Race Rating Source:http://www.realclearpolitics.com/epolls/2010/governor/2010_ elections_governor_map.html
Works Cited
Franz, Michael M., Paul B. Freedman, Kenneth M. Goldstein, and Travis N. Ridout. Campaign Advertising and American Democracy. Philadelphia: Temple University Press, 2007. Print.
Goldstein, Ken, and Paul Freedman. “Campaign Advertising And Voter Turnout: New Evidence For A Stimulation Effect.” Journal Of Politics 64.3 (2002): 721-740. Academic Search Complete. Web. 2 Nov. 2014.
Hogan, Robert E. “Campaign Spending and Voter Participation in State Legislative Elections.” Social Science Quarterly (Wiley-Blackwell) 94.3 (2013): 840-864. Academic Search Complete. Web. 2 Nov. 2014.
Jackson, Robert A., Jeffery J. Mondak, and Robert Huckfeldt. “Examining the Possible Corrosive Impact of Negative Advertising on Citizens’ Attitudes Toward Politics.” Political Research Quarterly 62.1 (2009): 55-69. ProQuest. Web. 2 Nov. 2014.
Kahn, Kim Fridkin and Patrick J. Kenney. “Do Negative Campaigns Mobilize or Suppress Turnout? Clarifying the Relationship between Negativity and Participation.” The American Political Science Review 93.4 (1999): 877-889. JSTOR. Web. 17 Sept. 2014.
Krupnikov, Yanna. “When Does Negativity Demobilize? Tracing The Conditional Effect Of Negative Campaigning On Voter Turnout.” American Journal Of Political Science 55.4 (2011): 797-813. Academic Search Complete. Web. 17 Sept. 2014.
Lau R. Richard and Gerald M. Pomper “Effects of Negative Campaigning on Turnout in U.S. Senate Elections, 1988-1998.” The Journal of Politics 63.3 (2001): 804-819. JSTOR. Web. 2 Nov. 2014.
Leighley, Jan E., and Jonathan Nagler. Who Votes Now? Princeton: Princeton University Press, 2014. Print.
Martin, Paul S. “Inside The Black Box Of Negative Campaign Effects: Three Reasons Why Negative Campaigns Mobilize.” Political Psychology 25.4 (2004): 545-562. Academic Search Complete. Web. 2 Nov. 2014.
Soubeyran, Raphaël. “Contest With Attack And Defense: Does Negative Campaigning Increase Or Decrease Voter Turnout?” Social Choice & Welfare 32.3 (2009): 337-353. Academic Search Complete. Web. 2 Nov. 2014.
Leave a Reply