[…] When North Carolina House Speaker Thom Tillis was shown in a picture with Governor Pat
McCrory his favorability rating also declined, but this negative effect was much smaller than
Obama’s negative effect on Hagan. Even Republican’s viewed Tillis more negatively when seen
with the governor. Independents saw almost no change in their perception of Tillis when he is
depicted in an image with McCrory.[…]
How do we determine how helpful or harmful the governor is to Thom Tillis? One method is a to conduct a
randomized experiment whichpresents half of respondents with an image of Thom Tillis by himself and
the other half an image of Thom Tillis next to the governor […] The experiment found that when
respondents saw Thom Tillis with Governor Pat McCrory their perceptions of Tillis were negatively
affected. Although the negative effect that McCrory has on respondents was statistically significant for several
groups […], the effect was much smaller overall than Obama had on Hagan’s favorability.
Tillis should probably also distance himself from the governor but the ramifications of being tied to
McCrory is not as detrimental as being tied to President Barack Obama.
The Elon folks also suggest that Sean Haugh — contrary to conventional wisdom — hurts Hagan worse than Tillis:
The following table presents the results of the experiment to determine which candidate would
likely lose more votes to Haugh. The experiment uses survey data obtained from an opt-in online
panel and was not designed to measure how much of a lead a candidate has (See the September
2014 Elon University Poll which uses a RDD/Wireless telephone sample if interested in how likely
voters plan to vote in November). The experiment found that Hagan is hurt more than Tillis by the
presence of Sean Haugh on the ballot. Hagan loses 8 percentage points to Haugh, while Tillis only loses 4 percentage points.
This suggests that if the race is extremely close, Haugh’s presence on the ballot could influence the outcome
by benefiting one candidate, specifically Tillis.
They also found some interesting information regarding ballot order:
The order in which the candidates appear on the ballot is determined by a process designated by
state law (G.S. 163-165.6). Candidates are listed on the ballot “in alphabetical order by party
beginning with the party whose nominee for Governor received the most votes in the most recent
gubernatorial election.” Because Pat McCrory won the governor’s race in 2012 Tillis will be listed
first on the ballot. Could this ballot order influence the election outcome?
Prior research suggests that for low information races ballot order can matter.
Low information races are races where there is little media attention, such as local elections or nonpartisan
elections. The North Carolina U.S. Senate rate does not resemble a “low information” race and
therefore one might expect that ballot ordering will not influence the outcome of this election. We
test this hypothesis by conducting a randomized experiment within an online survey which presents different ballot options to different
respondents. For one ballot Tillis is listed first, for an alternative ballot Hagan is listed first.
Elon also ran some studies about voter attitudes when candidates are connected to specific issues:
Table 3 below presents the results of the randomized experiment. Because the data was not
obtained through a random sample the absolute numbers cannot be interpreted as representing
the size of Hagan’s advantage. Instead, the experiment is designed to see if there are any
differences in evaluations by policy issue. The overall results suggest that Hagan’s favorability
(relative to Tillis) is largest when respondents are primed with the education (.60) and the
women’s health postcards (.56). Hagan’s favorability (relative to Tillis) is not as large when respondents are shown
postcards dealing with families (.27) and national defense (.27) and immigration (.23). This suggests that respondents
are less likely to see Tillis in a favorable view when they are reminded of education and women’s health. Hagan’s
relative advantage drops when respondents are reminded of issue related to families,
national defense and immigration.The overall numbers are useful but
candidates may want to know how Independents see the two candidates when primed with
different policy issues. Hagan again has her strongest advantage when Independent registered
voters are shown the education postcard and the smallest relative advantage when respondents
are shown the economy and voter identification postcards.
Among men Hagan’s advantage disappears when male respondents are shown the national
defense postcard and nearly disappears when shown the family (gay marriage) and the economy
postcard. Hagan’s advantage also declines with women when female respondents were shown the
The results suggest that Tillis would benefit by attacking Hagan on issues such as immigration,
national defense, voter identification and the economy. It also suggests that Tillis needs to
improve his image regarding a number of policy issues, but especially with education, and