Representative Sample of 975 Registered Voters Statewide
Margin of Error +/- 3.1 percentage points
Fielded July 27 - 30, 2020

Weighted Demographics

D1 County / Congressional District (from voter file)
Oahu_1
Oahu_2
Hawaii
Maui
Kauai

51%
17%
15%
11%
6%

D2 Gender
Male
Female

47%
53%

D3 Age
18-49
50 or older

33%
67%

D4 Ancestry/Ethnicity/Race
Caucasian
Japanese
Filipino
Hawaiian
Chinese
Hispanic
Other/Mixed

32%
22%
10%
15%
5%
2%
17%

1

 Representative Sample of 975 Registered Voters Statewide
Margin of Error +/- 3.1 percentage points
Fielded July 27 - 30, 2020

D5 Education (college degree)
College Degree
No College Degree

51%
49%

D6 Political affiliation
Democrat
Republican
Independent
Not sure

58%
18%
25%
5%

D7 Political ideology
Liberal/Progressive
Moderate
Conservative
Not sure

28%
32%
22%
18%

D8 Household income
Less than $50,000
$50,000 - $100,000
More than $100,000
No Response

20%
33%
31%
16%

2

 3

Representative Sample of 975 Registered Voters Statewide
Margin of Error +/- 3.1 percentage points
Fielded July 27 - 30, 2020

About the Poll
Civil Beat surveyed a random sample of 988 registered voters statewide. Surveys were
fielded from October 2 – 7, 2020, using a combination of interactive voice response
technology (touch-tone polling) and a survey administered online. Both the touch-tone
and online version of the poll were conducted using random, probability-based sampling
of registered voters in Hawaii. The touch-tone version was conducted by contacting
landline telephones. The online version was conducted by texting cellphones and linking
poll participants to an online survey optimized for smartphones. Questions in the
telephone and online versions of the survey were nearly identical, other than differences
necessitated by the different platforms. The text-to-online version of the poll replaces
our previous method of contacting cellphones via live call agents, but retains the same
method to identify a randomly-selected subset of registered voters to contact. The new
(text) method yields a much higher survey completion rate and a more demographically
diverse set of respondents than did the previous (call-agent) method. Approximately two
thirds of the total sample completed the poll via landline and the other one third
completed it via text-to-online.
Results were adjusted for voter likelihood and also balanced for gender, age,
ancestry/ethnicity/race, county and Congressional District of residence, and education
(college degree attainment) to match the demographic parameters of registered voters
statewide.
The overall margin of error is +/- 3.1 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 margin of error, accounting for the design effect is +/- 3.9 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 MRG Research, a public opinion
research organization specializing in issues and candidate polling, message testing,
and persuasive communications. Twitter: @MRG_Research