Civil Beat Poll 1 Representative Sample: Likely Republican Primary Voters (219, Margin of Error +/- 6.6 percentage points) Fielded, July 19 – 21, 2018 And how about in the Republican primary for Lieutenant Governor? Would you most likely vote for Marissa Kerns of Jeremy Low? [Name order rotated] Kerns 26% Low 20% Unsure 54% Total 100% CROSSTABS Male Kerns Low Unsure Total 27% 18% 55% 100% Female 25% 24% 51% 100% 50 or older Kerns Low Unsure Total Under 50 29% 25% 46% 100% Kerns Low Unsure Total Caucasian Japanese Filipino Hawaiian Chinese Hispanic Other/Mixed 31% 22% 37% 26% 9% 52% 17% 24% 25% 15% 16% 8% 8% 29% 45% 53% 49% 58% 83% 40% 54% 100% 100% 100% 100% 100% 100% 100% 26% 19% 55% 100% Civil Beat Poll Representative Sample: Likely Republican Primary Voters (219, Margin of Error +/- 6.6 percentage points) Fielded, July 19 – 21, 2018 Kerns Low Unsure Total Liberal/progressive Moderate Conservative 39% 42% 25% 30% 8% 25% 32% 50% 50% 39% 42% 25% Kerns Low Unsure Total Democrat Republican Independent 40% 25% 32% 44% 17% 21% 16% 58% 47% 100% 100% 100% Kerns Low Unsure Total No College college Degree Degree 27% 24% 18% 27% 55% 49% 100% 100% Kerns Low Unsure Total Less than $50,000 - More than $50,000 $100,000 $100,000 34% 27% 32% 18% 26% 20% 48% 48% 48% 100% 100% 100% Kerns Low Unsure Total Oahu_1 Oahu_2 Hawaii Maui Kauai 27% 18% 16% 77% 23% 20% 14% 35% 15% 16% 54% 68% 50% 8% 60% 100% 100% 100% 100% 2 Civil Beat Poll 3 Representative Sample: Likely Republican Primary Voters (219, Margin of Error +/- 6.6 percentage points) Fielded, July 19 – 21, 2018 About the Poll Civil Beat surveyed a random sample of 1,090 likely primary voters statewide, including 871 likely Democratic Party primary voters (403 in the First Congressional District and 468 in the Second Congressional District), and 219 likely Republican Party primary voters. Calls were made from July 19 – 21, 2018, using interactive voice response technology (touch-tone polling). Cell phones were contacted by live call-agents who obtained consent and then patched consenting participants through to the same recorded survey as landline respondents. The sample consisted of 70% landlines and 30% cellphones. Results were balanced for gender, age, ancestry/ethnicity/race, county and Congressional District of residence, and education level. The margin of error for the full statewide sample of likely primary voters is +/- 3.0 percentage points, for likely Democratic primary voters it is +/- 3.3 percentage points (First Congressional District +/- 4.9 percentage points and Second Congressional District +/- 4.5 percentage points), and for likely Republican primary voters it is +/- 6.6 percentage points. Other sources of error, such as imperfect response rates and design effect (the effect of weighting results based on demographics) are also common in public opinion research and may affect the results. The adjusted margins of error, accounting for the design effects are: full statewide sample of likely primary voters +/- 4.4 percentage points, likely Democratic primary voters +/- 4.8 percentage points (First Congressional District +/- 7.3 percentage points and Second Congressional District +/- 6.2 percentage points), and likely Republican primary voters +/- 10.7 percentage points. The margin of error for any crosstabulated result is larger than the margin of error of the corresponding topline result. Some columns may sum to totals slightly higher or lower than 100% due to rounding. The poll was conducted by Civil Beat working with Merriman River Group, a full-service consulting organization specializing in opinion research, election management, and communications.