October 2021 Seattle Survey Methodology

Change Research surveyed 617 likely voters online for the Northwest Progressive Institute (NPI) in Seattle from October 12th-15th, 2021.

Change Research used the following sources to recruit respondents:

  • targeted advertisements on Facebook
  • targeted advertisements on Instagram
  • text messages sent, via the echo19 and/or Scale To Win platforms, to cell phone numbers listed on the voter file for individuals who qualified for the survey’s sample universe, based on their voter file data

Regardless of which of these sources a respondent came from, they were directed to a survey hosted on SurveyMonkey’s website.

Ads placed on social media targeted all adults living in Seattle. Those who indicated that they were not registered to vote were terminated. As the survey fielded, Change Research used dynamic online sampling: adjusting ad budgets, lowering budgets for ads targeting groups that were overrepresented and raising budgets for ads targeting groups that were underrepresented, so that the final sample was roughly representative of the population across different groups.

The survey was conducted in English.

Post-stratification was performed on age, gender, race/ethnicity, education, and 2020 presidential vote. Weighting parameters were based on the demographic composition of the November 2017 electorate, obtained from the voter file. That is, if a given age bracket or gender group represented x% of the November 2017 electorate on the voter file, then that same group would be weighted to x% in this survey. Presidential results were obtained from precinct data.

The modeled margin of error* for this survey is 4.1%, which uses effective sample sizes** that adjust for the design effect of weighting.

Change Research has adopted The Pew Research Center’s convention for the term “modeled margin of error”(1) (mMOE) to indicate that our surveys are not simple random samples in the pure sense, similar to any survey that has either non-response bias or for which the general population was not invited at random. A common, if imperfect, convention for reporting survey results is to use a single, survey-level mMOE based on a normal approximation.

This is a poor approximation for proportion estimates close to 0 or 1. However, it is a useful communication tool in many settings and is reasonable in places where the proportion of interest is close to 50%. We report this normal approximation for our surveys assuming a proportion estimate of 50%.

** The effective sample size adjusts for the weighting applied to respondents, and is calculated using Kish’s approximation (2).

Who took the survey?

As noted above, 617 likely voters participated.

Gender breakdown

  • 46% of respondents identified as male
  • 51% of respondents identified as female
  • 3% of respondents identified as non-binary/other

Age breakdown

  • 21% of respondents are between the ages of 18-34
  • 29% of respondents are between the ages of 35-49
  • 27% of respondents are between the ages of 50-64
  • 23% of respondents are 65 or older

Race/ethnicity breakdown

  • 82% of respondents identified as White / Caucasian
  • 7% of respondents identified as Asian or Pacific Islander
  • 5% of respondents identified as Hispanic or Latino/a
  • 3% of respondents identified as Black or African American
  • 1% of respondents identified as American Indian or Alaska Native
  • 2% of respondents identified as Other


“We’re delighted to once again be able to share credible, independent data about the 2021 Seattle races with the public,” said Northwest Progressive Institute founder and executive director Andrew Villeneuve in a statement released on October 19th, 2021.

“NPI is fiercely committed to sound public opinion research practices. We know that the answers you get depend on the questions you ask. That’s why we work hard to ask neutral questions of a representative sample. Each of the results we’ve released contains the text of the questions that we asked, like always. Anyone can read the questions and see that they are neutrally worded, with nothing put in to influence or bias respondents. NPI is not aligned with any of the candidates running for office in Seattle, and we aren’t involved in any independent expenditures or other electioneering projects. That frees us to focus on providing credible, independent research and analysis to anyone interested in understanding what’s happening in this election.”

“We are very grateful to have been able to work with Change Research on two consecutive polls of the Seattle electorate. Change Research shares our commitment to sound research practices that yield quality data, and they worked hard to ensure that our poll’s samples would be as representative as possible. The results of the August 2021 Top Two election demonstrated that Change’s modeling for our last project was sound. Our poll was able to anticipate most of the dynamics of that election despite the fact that it was conducted before voting began, at a time when huge numbers of voters reported being undecided. Seven of the eight finalists chosen by voters either placed first or second in our Top Two survey, and our poll provided the first indication that City Attorney Pete Holmes was in danger of being ousted in the Top Two election, which eventually did happen.”

“We’re now in the final phase of this election cycle. Unlike in the Top Two, there’s other polling available that we can compare our data to, which is a welcome development. All of the candidates leading in our poll also led in the Crosscut/Elway poll last month and in the Strategies 360/KOMO poll (which did not ask about Council races, only Mayor and City Attorney), which suggests that we were once again able to capture an authoritative snapshot of the dynamics in this election.”

“There’s a lot still in flux right now given that this is an election in progress, so it’s difficult to draw many conclusions about what’s happening in Seattle politics at this juncture. We’re going to know a lot more next month. However, one thing that did stand out to our team when we were assessing the results was that there seems to be a real openness in electing an ideologically diverse set of candidates for Seattle’s citywide positions among a critical mass of voters.”

“It often doesn’t take that many people to swing a result one way or another: small groups of voters can wield a lot of power. And there may well be enough voters interested in ideological diversity this year that we could see some truly interesting results. Our team thinks that the odds of either of the two major ‘slates’ getting elected in their entirety is pretty slim. We would not be surprised, for instance, to see both Bruce Harrell and Nikkita Oliver win next month, and for them to win with the support of some of the same voters.”