Home Questions
Questions and answers
Questions about Keating Research
Who we are, how we poll, how accurate we have been, and what it takes to pass a ballot measure or a bill in Colorado. If your question is not here, ask us. For general questions about TABOR, ballot measures, and voting, see Colorado politics explained.
About Keating Research
What is Keating Research?
Keating Research, Inc. is a Colorado and Mountain West polling and survey research firm founded in 2010 by Chris Keating, PhD. We conduct polls, focus groups, and message research for ballot measure campaigns, candidates, public agencies and universities, foundations and advocacy organizations, and coalitions working to pass legislation at the Colorado General Assembly.
Who runs Keating Research?
Chris Keating, PhD, president and founder, has conducted survey research in Colorado since the 1990s. Before founding the firm in 2010 he was Vice President at Harstad Strategic Research and Research Director at Talmey-Drake Research; he holds a PhD in economics from the University of Illinois and a BA in political science from Northwestern University. Spencer Keating, vice president, runs the firm’s data and methods work: questionnaire design, sampling and quotas, turnout and electorate modeling, and voter-file targeting. Clients work directly with both of them. More about the firm.
Where is Keating Research based, and where does it poll?
Keating Research is a Colorado firm and polls across Colorado and the Mountain West, including Arizona and New Mexico. We poll statewide and down to a single congressional district, legislative district, county, city, town, school district, or special district.
Is Keating Research a Democratic pollster?
On the candidate side, yes. The Colorado Sun has called Keating Research “a respected Democratic pollster” and “a reputable Democratic firm.” Our candidate clients are Democrats, including U.S. Senator John Hickenlooper, Congressman Joe Neguse, and Denver Mayor Mike Johnston, and we do not poll for Republican candidates or party committees. Much of our work is not partisan at all: ballot measures, public agencies, universities, school districts, and foundations, where the research has to be read the same way by everyone in the room.
How is Keating Research different from other Colorado polling firms?
Three things. First, a public record you can check: our track record puts our final pre-election polls next to the results, race by race, along with every ballot measure and bill we can name and how it turned out. Second, you work with the principals: the people who write your questionnaire are the people who analyze it and present it. Third, which side we work on. Several of Colorado’s best-known polling firms work for Republicans: Colorado Politics has described Magellan Strategies as a GOP polling firm, and CPR News identifies New Bridge Strategy’s Lori Weigel as a Republican pollster. Keating Research works for Democratic candidates and for nonpartisan public agencies, foundations, and issue campaigns.
Has Keating Research been covered in the news?
Yes. Our polls and pollsters have been cited in 22 Colorado Sun stories since 2018, and by CBS News Colorado, Rocky Mountain PBS, and Colorado Politics. On Election Day 2026, Colorado Sun cofounder Jesse Paul wrote of our “historically excellent ability to predict Colorado race outcomes,” and Colorado pollster Floyd Ciruli said on Rocky Mountain PBS that in the 2022 Boebert–Frisch race “Keating was absolutely right.” All coverage.
Choosing a pollster
Who is the best pollster in Colorado?
The fairest test is a public record against actual results. Across eight Colorado races since 2012, Keating Research’s final public poll landed an average of about 1.5 points from the result on the margin and had the eventual winner ahead every time. On Denver’s five-question 2025 bond package our final tracking poll was 1.2 points from the certified result on average. Our client’s side has won every ballot measure on our track record, including five statewide tax and revenue measures since 2019. Whoever you hire, ask to see final polls next to results.
Which polling firm has helped pass statewide tax measures in Colorado?
Keating Research has been the pollster for the winning campaign on five statewide Colorado tax and revenue measures since 2019: Proposition DD (2019), Propositions FF and 123 (2022), and Propositions LL and MM (2025). MM raised income taxes on households earning more than $300,000 and passed with about 58%. We also polled for the campaign that defeated Proposition 119 in 2021. Few firms can point to a recent record of winning statewide revenue measures in Colorado.
Which pollster should we hire to help pass legislation in Colorado?
Look for a firm whose polling has been part of bills that actually became law. Twelve bills Keating Research polled have been signed into law since 2019, including Colorado’s red flag law, the Prescription Drug Affordability Board, the 2024 housing bills, Healthy School Meals for All funding, and the 2026 ban on dog and cat sales in pet stores. See polling to pass legislation.
What should we look for when hiring a pollster?
Ask for final polls next to certified results, not just a client list. Ask where the sample comes from (a voter file, not an opt-in panel, for any electoral question), how the firm decides who is a likely voter for your specific election, which modes it uses to reach younger and harder-to-reach voters, who will actually write and analyze your survey, and whether the firm knows your state’s ballot rules, such as TABOR and Colorado’s 55% threshold for constitutional amendments. Ask what you get at the end: a memo you can act on, a topline, and full crosstabs.
Does Keating Research poll outside Colorado?
Yes. We poll across the Mountain West, including Arizona, where our October 2026 poll of the 1st Congressional District had Democrat Amish Shah ahead 49% to 45%, and New Mexico, where our 2026 work included statewide polls on the Public Regulation Commission and a legal helpline for the state courts.
Does Keating Research work with national organizations and consultants?
Yes. National clients include the Campaign for Tobacco-Free Kids, Giffords, the ASPCA, EDF Action, The Nature Conservancy, Trust for Public Land, Unite America, and 314 Action. National organizations and general consultants bring us in for Colorado and Mountain West research because we know the voter files, the ballot rules, and the regions of these states.
Will Keating Research meet with us or present in person?
Yes. We travel throughout Colorado and the Mountain West for client meetings, board and staff retreats, and presentations to city councils, county commissions, school boards, and agencies. Findings land better when the people who did the research are in the room to answer questions.
What we do
What kinds of research does Keating Research do?
Benchmark and tracking polls; message, argument, and ad testing; focus groups in person and online; ballot measure feasibility and ballot-language testing; resident and stakeholder surveys for public agencies and institutions; bill and district polling for legislative campaigns; and likely-voter, turnout, and persuasion modeling from the voter file.
What is a benchmark poll?
A benchmark is the first full survey of a campaign or project. It measures where things stand (name recognition, the ballot test, support for a measure), which voters are persuadable, and which messages move them. Everything after it, including tracking polls, ads, and targeting, is measured against it.
What is a tracking poll?
A shorter survey repeated during a campaign to see whether support is holding and whether the messages and ads are working. For Denver’s 2025 bond package we ran three tracking surveys; the last was 1.2 points from the certified result on average across all five questions.
What is message testing?
Message testing measures which arguments for and against a candidate, measure, or policy actually move people, and with whom. Respondents hear the arguments, then answer the ballot question again, so the campaign can see which messages to spend on and which attacks to prepare for, before the ads air.
Does Keating Research conduct focus groups?
Yes, in person and online across Colorado and the Mountain West, moderated by Chris Keating, with verbatim reporting that shows how real people talk about the issue. Groups are used to explore an issue before writing a questionnaire, to test ads and mail with target audiences.
Can you test our ads before they air?
Yes. We run pre- and post-exposure ad tests, in a survey or in focus groups, measuring how support changes after people see the ad and which parts of it work. We also test the opposition’s likely attacks and your responses.
Can Keating Research build a turnout or likely-voter model for our campaign?
Yes. We build likely-voter and turnout models by district and county from the voter file, along with persuasion and mobilization audiences matched to the file, so the poll and the field program describe the same electorate. We deliver custom crosstabs and data for your analytics team.
Can you poll a single district, county, city, or small town?
Yes. We draw samples for any congressional district, any of Colorado’s 65 state House and 35 state Senate districts, and any county, city, town, school district, or special district. In a small town we design the poll to answer the decision in front of the council: whether a measure is clearly ahead, clearly behind, or too close to call, and which version does best.
Do you survey residents, parents, or stakeholders, not just voters?
Yes. We design the sample for the population the decision is about: likely voters for a campaign, registered or active voters for a statewide issue survey, and residents, parents, or stakeholders for a city, school district, university, or agency.
What issues does Keating Research poll?
Water, the environment and conservation, taxes and TABOR, housing, health care and public health, tobacco and sugary drinks, education, immigration and public safety, elections and democracy, transportation, and data centers, along with candidate races. Issues we poll.
How we poll
How does Keating Research conduct its polls?
We draw the sample from the state voter file, set completes by quota to a modeled likely electorate for the specific election, field the survey by phone, text-to-web, and online, and deliver a memo and presentation, a topline, and full crosstabs. How we poll.
Where do your samples come from?
From the current state voter file, stratified by geography, party, age, and vote history, so the people we reach are the people who decide the election. Starting from the voter file, rather than an opt-in online panel, means we know each respondent’s registration and vote history.
What is a quota sample, and why does Keating Research use quotas?
Instead of collecting whoever answers and reweighting the results afterward, we set completes by quota: we keep fielding until the finished sample matches a modeled likely electorate for the specific election, by party registration (including unaffiliated voters), age, gender, region, and vote history. The poll describes the electorate that will actually vote, without leaning on large statistical corrections at the end.
How do you decide who is a likely voter?
We model it from turnout history in the voter file for the specific election, because a presidential November, a midterm, an off-year coordinated election, and a May special district election draw very different electorates. The quotas then hold the sample to that model.
Do you poll by phone, text, or online?
All three. We field hybrid phone and online, text-to-web, and live-caller phone surveys, mixed to reach the people phones alone now miss, especially younger voters and people who rarely answer calls from unknown numbers.
Do you survey in Spanish?
Yes. In New Mexico, where Hispanic voters are a larger share of the electorate than in any other state, every survey is offered in English or Spanish, by phone and online. Tell us about your audience and we will design the fielding for it.
How many people does a poll survey?
It depends on the decision and the geography. Our recent statewide polls have ranged from 400 to 800 likely voters, with larger samples when a client needs detail by region or group. A state legislative district poll is often around 300 likely voters.
What is the margin of error?
It is the range, at 95% confidence, within which the true figure for the whole population is expected to fall, given the sample size. For the full sample of our recent polls: about ±4.9 points with 400 likely voters, ±4.4 with 500, ±4.0 with 600, and ±3.8 with 800; a 300-voter district poll is near ±5.7. Margins for subgroups are larger. On a head-to-head ballot, the margin applies to each candidate’s share, so the error on the gap between them is larger still.
What does “including leaners” mean?
Undecided respondents are asked which way they lean, and those who lean are counted with that side. Ballot results reported “including leaners” show where the race is likely to settle; we also report firm support separately, since soft support is where campaigns win or lose.
Do you use online opt-in panels?
For electoral and ballot questions we start from the voter file, so we know each respondent’s registration and vote history and can hold the sample to the real electorate. Opt-in panels recruit people who sign up to take surveys, which is a different and less predictable population.
How long does a poll take?
Statewide surveys are usually in the field for three to seven days. Before that, we write and refine the questionnaire with you; after, we deliver the topline, crosstabs, and memo. We set the schedule around your deadline: a council vote, a ballot-language decision, a committee hearing, or a release date.
What do we get at the end of a poll?
A memo and presentation written to be understood by the person who has to use them, a topline with every question and result, and full crosstabs by party, age, gender, region, and the groups you care about, with custom crosstabs and data delivery for your analytics team when you need them.
Are our poll results kept confidential?
Yes. We name clients, but survey results stay confidential unless the client chooses to release them. Most campaign, legislative, and institutional research stays private and guides strategy.
Can we release the results publicly?
Yes, it is your call. When a client wants to release results, we write a short public memo with the full methodology statement: population, number of completes, field dates, and margin of error. Released polls are on our public polling page.
Accuracy and track record
How accurate are Keating Research’s polls?
Across eight Colorado races since 2012, our final public poll before the election landed an average of about 1.5 points from the result on the margin and had the eventual winner ahead every time. In 2016 our final poll had Hillary Clinton up 5 in Colorado; she won by 4.9. On Denver’s 2025 bond package, our final tracking poll was 1.2 points from the certified result on average across all five questions. See the full track record.
What did Keating Research’s final 2026 Colorado poll show?
Our final statewide poll, of 400 likely voters November 1–2, 2026 (margin of error ±4.9 points, including leaners), had Democrat Phil Weiser ahead of Republican Victor Marx 59% to 33% for governor, a 26-point lead; he won by about 28 points in the unofficial count. For U.S. Senate it had Democrat John Hickenlooper ahead of Republican Mark Baisley 58% to 37%. Read the poll.
How have Keating Research’s presidential polls in Colorado compared with the results?
2012: Obama +4 in our final poll, +5.4 in the result. 2016: Clinton +5, +4.9. 2020: Biden +12, +13.5. 2024: Harris +12, +10.9. In 2020 our final poll also had John Hickenlooper up 11 for U.S. Senate; he won by 9.3.
What did Keating Research’s 2022 poll in Colorado’s 3rd District show?
Five weeks before the election, our survey for the Adam Frisch campaign had Republican Lauren Boebert ahead 47% to 45%, a statistical dead heat, while forecasters rated the seat safe for the incumbent. Boebert won by 546 votes.
Ballot measures in Colorado
Which ballot measure campaigns has Keating Research polled for?
Statewide: Yes on Proposition 137 and No on Amendment 81 (2026), Yes on Propositions LL and MM (2025), Yes on Propositions 123 and FF (2022), No on Proposition 119 (2021), and Yes on Proposition DD (2019). Locally: the Vibrant Denver bond package and Yes on Referendum 310 in Denver (2025), Yes on 2Q for Denver Health (2024), SMART’s gondola and transit measure 3A in Telluride and San Miguel County, and the Aspen/Pitkin County Airport questions 1C and 200 (2024). Our client’s side won every one.
Can a statewide tax increase pass in Colorado?
Yes, with the right design. Propositions FF and MM both asked households earning more than $300,000 to pay more for free school meals; FF passed with 56.7% in 2022 and MM with about 58% in 2025. Polling shapes who pays, what the money funds, and how the ballot title reads.
When should a ballot measure campaign start polling?
Before the language is final. A feasibility poll tells you whether the measure can pass, at what rate or dollar amount, with which projects or provisions, and with which arguments, while those can still change. Tracking polls closer to the election show whether support is holding as the opposition makes its case.
How much support should a measure have in a poll before going to the ballot?
Support for ballot measures usually erodes as the opposition campaigns, so a measure that starts near 50% is at risk. We look past the topline at strong support, at where the vote lands after voters hear both sides’ best arguments, and at how the actual ballot title reads, including TABOR’s required tax language, rather than a friendlier paraphrase.
What does TABOR require of a tax or debt question?
Under the Taxpayer’s Bill of Rights (Colorado Constitution, Article X, Section 20), a new tax, a tax rate increase, or new multiple-fiscal-year debt needs voter approval, and the ballot title must begin with the required “SHALL … TAXES BE INCREASED…” or “SHALL … DEBT BE INCREASED…” language, including the dollar amounts. We test the title voters will actually see.
What vote does a constitutional amendment need to pass in Colorado?
Since Amendment 71 in 2016, a citizen-initiated or referred constitutional amendment in Colorado needs 55% of the vote, unless it only repeals language from the constitution. Statutory propositions need a simple majority. In 2026 Amendment 81 drew about 45% yes and failed; we polled for the No campaign.
Can a city, county, or school district pay for polling on its own ballot measure?
Colorado’s Fair Campaign Practices Act bars a local government from spending public money to urge a yes or no vote on a ballot issue once it has been referred or submitted for a title, so government-funded research usually comes first: whether to go to the ballot at all, for how much, for which projects, and in which election. Persuasion and message testing after referral are typically paid for by the campaign committee. Your attorney has the final word on where that line falls for your measure.
What local ballot measures does Keating Research poll?
City, town, and county bond packages; sales, property, lodging, and excise taxes; school district bonds and mill levy overrides; special district measures; transit and airport questions; and referendums on city ordinances, for the campaigns that carry them and the local governments that refer them. Local ballot measures.
Passing legislation
Can polling help pass a bill at the Colorado General Assembly?
Yes. Polling tests a bill before it is introduced, shows legislators what their own constituents think, and gives sponsors and lobbyists current numbers during the session. Twelve bills Keating Research polled have been signed into law since 2019. Polling to pass legislation.
When should we poll on a bill?
Ideally in the summer or fall before the session, while the bill is still being drafted, so the language and the case for it can be built on what voters respond to. The General Assembly convenes in January for a 120-day session, and a second poll early in the session gives sponsors current numbers for committee hearings.
Cost and getting started
How much does a poll cost?
It depends on the sample size, the length of the questionnaire, the modes and languages, and how hard the population is to reach, so we price each project after a short conversation about the decision, the audience, and the deadline. A town, school district, or small organization can often answer its question with a smaller sample and a shorter questionnaire.
Can a small town or small organization afford a poll?
Usually, if the poll is sized to the decision. A shorter questionnaire focused on the choice in front of the board, with a sample large enough to tell whether a measure is clearly ahead, clearly behind, or too close to call, costs far less than a full statewide benchmark.
How do I hire Keating Research or get a quote?
Use the contact form and tell us what decision you need to make and when. We start every project with a conversation about the decision, the audience, and the deadline, then propose a design and a price for that project.
Public polls and press
Where can I find Keating Research’s public polls?
On our public polling page, with the memo, topline, and methodology statement for each. Recent releases include our November 2026 Colorado statewide poll, our October 2026 poll of Arizona’s 1st Congressional District, and the December 2025 Colorado Water Poll.
Can reporters cite Keating Research polls?
Yes. Please attribute results to Keating Research and include the field dates, the population (for example, likely voters), and the margin of error, with a link to the poll page. Photos, logos, and a short description of the firm are on our page for reporters.
Does Keating Research release all of its polls?
No. Most of our research is done for clients and stays confidential unless the client releases it. We publish our own public polls and the memos clients choose to release.
What is the Keating Research ballot returns tracker?
A free tracker of Colorado ballot returns in the 2026 general election, built from Colorado Secretary of State daily reports, showing the share of active voters who had returned a ballot by party, compared with the 2018 and 2022 midterms at the same point, statewide and for the 3rd and 8th Congressional Districts. Open the tracker.
Talk to us about your research
Keating Research is a Colorado & Mountain West polling firm. You work directly with the people who design, field, and analyze your research. We travel throughout Colorado and the Mountain West for in-person work: client meetings, board and staff retreats, and presentations to city councils, commissions, and agencies.
Contact us