The Connecticut Senate race between Congressman Chris Murphy and former wrestling executive Linda McMahon is one of the most closely watched in the nation. There are many interesting dynamics in this race – President Obama’s coattails, McMahon’s television ads, her professional wrestling background, and both candidates’ handling of their personal finances. How much if at all do these and other issues determine vote preferences in the race? To answer this question Quinnipiac asked Langer Research Associates of New York, N.Y., to collaborate on a special data analysis project using a statistical technique called regression to seek to identify key factors in the contest. Their report follows.
–
Competitiveness on economic empathy and help from Barack Obama’s coattails are keeping Chris Murphy alive in the U.S. Senate race in Connecticut, while his Republican opponent, Linda McMahon, benefits from higher personal favorability overall, positive regard for her business experience and the impact of her saturation television advertising.
Each of these elements independently predicts vote preferences among likely voters in the closely divided race, holding constant a broad range of other demographic and attitudinal variables, according to a statistical analysis of an Oct. 2 statewide Quinnipiac University poll on the contest.
The survey, conducted before the current round of candidate debates, evaluated a wide range of issues in the closely followed contest. Subsequent research will evaluate how well each candidate has been able to capitalize on opportunities – and avoid the pitfalls – identified in this analysis.
The survey overall found several deficits for Murphy, in areas including voter enthusiasm, personal favorability and perceived qualifications for the job. But regression modeling indicates that he remains competitive chiefly on the basis of views of which candidate “better understands the economic problems people in Connecticut are having.” This perception is the single strongest predictor of vote choice by a wide margin; the fact that Murphy runs evenly with McMahon on empathy (45-46 percent, Murphy-McMahon) keeps him in the race.
On the other hand, as a Republican candidate in a predominately blue state, McMahon has to overcome partisan predispositions and presidential preferences – both also significant predictors of vote – to stay in the race. Modeling suggests that one of the ways she’s doing so is by presenting a more favorable persona to the public than her challenger. Views of McMahon personally predict candidate preferences to a greater degree than do views of Murphy; therefore maintaining or expanding her very slightly positive assessment (45-41, favorable-unfavorable) may be critical in the weeks ahead. That could complicate her task in the debates, as well as in her heavy campaign advertising – scoring political points against her opponent without letting her positive personal ratings slip.
This analysis is based on a statistical technique called regression, which measures the independent strength of the relationship between each potential predictor and vote preferences, by holding all other predictors constant. As noted, views of economic empathy are the single strongest predictor of vote preference, followed by favorable views of McMahon and by intent to support Obama in the presidential race.
Additional, weaker, predictors include:
• Exposure to McMahon’s ads and (to a lesser extent) views of their effectiveness. Likely voters who report seeing McMahon’s ads frequently and who perceive them as especially effective are more likely to support her. Whether this is because the ads themselves are persuading voters to support McMahon, or because McMahon supporters are more likely to attend to her ads and find them convincing, is an open question. In either case, though, McMahon’s ads show a significant independent impact on vote preferences – while Murphy’s do not.
• Whether likely voters view political or business experience as the better qualification for Senate. Viewing business as better experience is a significant independent predictor of voting for McMahon; seeing political experience as more important independently predicts voting for Murphy. It’s a net advantage for McMahon because more likely voters see business experience than political experience as important.
• Openness to switching candidates. With all else held constant, voters who indicate they might change their mind before the election are more likely to support McMahon than Murphy. This suggests that McMahon is currently wining a greater share of persuadable voters than Murphy. At the same time, this may provide an opening for Murphy, to the extent he can effectively persuade these voters to move to his side
• Perceiving McMahon as intentionally misleading voters. While McMahon has the clear advantage in ad exposure and effectiveness, modeling suggests there is a potential downside to this advantage. Views that she is intentionally misleading viewers are a significant independent predictor of voting for Murphy, whereas views that Murphy is lying in his campaign ads do not significantly impact vote choice – perhaps a sign of some pushback against McMahon’s ad campaign.
• Demographics. After controlling for partisanship and other factors, sex emerges as an independent predictor of vote choice (albeit a weak one), with men more likely to support Murphy than McMahon. (This is so only when controlling for the fact that women are more likely than men to be Democrats, and thus, absent controls, to vote for Murphy.)
Notably, perceptions of professional wrestling and of Congress, and ratings of McMahon and Murphy’s work performance in these fields, specifically, are not significant drivers of vote preferences. Also, despite the amount of attention paid to it, views of how the candidates have handled their finances also are not direct predictors of current candidate preference.
However, these items do significantly impact the favorability ratings of each candidate, and therefore have an indirect influence on the contest more broadly.
Specifically, likely voters who think McMahon has handled her finances appropriately, who view professional wrestling favorably and who approve of McMahon’s experience as a wrestling executive all perceive her more favorably than do others. On the other hand, Murphy is viewed more favorably by those who approve of his work in the U.S. House of Representatives and who think he’s handled his finances appropriately, compared with those who do not.
Since favorability ratings are a significant predictor of vote preferences, especially for McMahon, this indicates that each of these campaign themes has the potential to shift the race, albeit indirectly.
Unlike at the national level, where the economy has been the focus of the presidential race, ratings of the national and state economy, and expectations of where each is headed, do not appear to be benefitting one Senate candidate over the other. However, economic empathy is a significant vote predictor both at the national level and, as noted, in the Connecticut Senate race – both contests in which wealthy Republican candidates are working mightily to persuade likely voters that they feel their pain.
Models follow.
Model 1:
Significant predictors of CT senate vote (R2 = .87)
Variable Beta p-value
Economic empathy: McMahon vs. Murphy .42 .000
McMahon favorability -.24 .000
Presidential vote .14 .000
Murphy favorability .08 .000
McMahon advertisement exposure -.06 .000
Better experience: Business vs. politics .05 .007
Party ID: Democrat .04 .004
McMahon intentionally misleading voters .04 .008
Moveable CT voter -.04 .010
Sex: Male .03 .007
Note: Vote is coded 0=McMahon, 1=Murphy. Positive betas indicate a preference for Murphy, negative betas indicate a preference for McMahon. Economic empathy likewise is coded 0=McMahon and 1=Murphy. Better experience: Business vs. politics is coded 0=business, 1=politics; therefore the positive beta means that those who think political experience is better are more likely to vote for Murphy (and vice versa).
Model 2:
Significant predictors of McMahon favorability (R2 = .80)
Variable Beta p-value
Economic empathy: McMahon vs. Murphy -.32 .000
Ratings of McMahon’s handling of finances .13 .000
Presidential vote -.12 .001
McMahon intentionally misleading voters -.11 .000
Better experience: politics vs. business -.10 .000
McMahon advertisement effectiveness .08 .000
Ratings of pro-wrestling .08 .000
Think will win debate: McMahon vs. Murphy -.07 .000
Education -.06 .000
Approval of McMahon’s wrestling exp. .06 .004
Murphy intentionally misleading voters .05 .005
Note: Favorability is coded continuously from 1=very unfavorable to 4=very favorable. Economic empathy is coded 0=McMahon and 1=Murphy.
Model 3:
Significant predictors of Murphy favorability (R2 = .71)
Variable Beta p-value
Economic empathy: McMahon vs. Murphy .26 .000
Approval of Murphy’s Congress exp. .18 .000
Approval of Murphy’s handling of finances .14 .000
Approval of McMahon’s wrestling exp. -.08 .004
Murphy intentionally misleading voters -.08 .000
Party ID: Democrat .06 .005
Note: Favorability is coded continuously from 1=very unfavorable to 4=very favorable. Economic empathy is coded 0=McMahon and 1=Murphy.