A FOX 5/Opinion Savvy poll conducted July 17-18 shows Brian Kemp with a very substantial lead over Casey Cagle in the runoff for the Republican nomination for Governor of Georgia. The survey of 466 registered likely voters has a margin of error of 4.5%.

The results:          Kemp: 55%

                               Cagle: 37%

                               Undecided: 8%

Opinion Savvy Analysis: “At first glance these numbers might seem hard to believe given the relatively close nature of the race over the last month. However, Casey Cagle’s support has steadily eroded in recent weeks, based on the results of other polls and our survey. Cagle was in serious decline on Tuesday night prior to the endorsement of Brian Kemp by President Trump. Donald Trump’s endorsement clearly accelerated the decline and has placed Cagle’s campaign into what is best described, at least for now, as a free fall.

As of Wednesday night:

  • Kemp leads early voting by a 56%-41% margin
  • Cagle and Kemp are tied among the youngest of voters (18-29) but Kemp has huge leads (well over twenty points) among those age 30-44 and 45-64. Among voters of age 65 or over, the two candidates are statistically tied.
  • Kemp leads among male voters and has a substantial lead among female voters.

It appears that the various negative ads from the Cagle campaign (or independent groups who back Cagle) aimed at Kemp in the last few weeks of the campaign have failed to drive up Brian Kemp’s negative/unfavorable ratings among voters. That may have deprived Cagle of establishing a more positive image of himself. It is difficult to measure the damage done to Cagle by the various tape recorded conversations with Cagle revealed by former candidate Lindsey Tippins. But clearly the endorsement by President Trump lifted Kemp to a very dominant position in the race based on Wednesday night’s results (which were gathered just hours after Trump’s endorsement was tweeted out). President Trump is extraordinarily popular among Georgia Republicans and the ultimate impact of his endorsement on the race could conceivably keep Cagle in the low-to-mid forty-percentile range once all votes are counted.

There are some caveats that those interested in the race should keep in mind. First, since the results for early voting in Georgia are closer among the two candidates in the poll, a low voter turnout next Tuesday could make the race much closer. Second, for those watching returns on election night, early and absentee votes are usually reported first in many areas of the state. It is conceivable that the early returns will show the race much tighter. If turnout is moderate or above, and the trends hold, Kemp will most likely increase his lead as the night goes on. If voter turnout is weak next Tuesday, then the race could remain much closer to the very end.”

Polling Notes:


Fox 5 Atlanta served as the sponsoring organization. Opinion Savvy, LLC conducted the survey on behalf of the sponsor.

The sample frame includes registered voters with either a landline telephone or internet-enabled mobile smartphone. Registered voters without access to either a landline telephone or an internet-enabled smartphone are excluded from the frame.

Blended sample; mixed mode

Contacted voters of both modes were selected randomly from a list of registered Georgia voters purchased by Opinion Savvy from Aristotle, Inc. Only voters determined likely to vote in the 2018 Republican primary were selected to participate in the survey, based on past election participation. Voters without a valid landline telephone number were selected for the mobile sample.

Landline: selected voters were contacted on July 17th and 18th 2018 using an automated interactive voice response system. The survey was presented in English.

Mobile: selected voters not available by landline telephone were contacted on July 17th and 18th 2018 on their mobile devices. Using voter data, selected voters were first matched to online and mobile application profiles electronically, using IP verification to ensure accurate selection and to prevent duplicate results. Selected voters were shown a brief message, push notification, or mobile advertisement requesting their participation in the survey. Respondents were offered a small, non-monetary prize for survey participation. Respondents answered a web- based survey instrument identical to that of landline respondents in written form. The survey was presented in English.


Total Number of Respondents Margin of Error*

The blended sample was weighted for age, race/ethnicity, and gender using propensity scores. Weighting benchmarks were determined using the 2014 Georgia Republican primary turnout figures with estimates provided by the Georgia Secretary of State.


±4.5% (95% confidence; not adjusted for weighting)

*The reported margin of sampling error is calculated using multiple assumptions that modern surveys rarely fit. The estimate assumes that, insofar as probability-based sampling is possible under real-world conditions, this model conforms to these parameters. Disregarding response rates, each individual within the studied population is equally likely to be selected for survey participation.



Lost your password?