Privacy-friendly analytics help us see which jobs people explore. They stay off unless you agree.Analytics off unless you agree. Details

Read the record
Back to the river
Skip to sources
Time Machine

Survey researchers

Also called Survey research analyst, Survey methodologist, Polling researcher, Survey research manager, Public opinion researcher

In brief

In May 2025, BLS counted 8,290 US survey researchers, with median pay of $69,460 a year. Of the BLS counts since 2003, the highest was 24,140, in May 2006. The earliest figure here is an estimate of 3,000 from 1910, not measured the same way as today's count. The last full technology era for survey researchers: Big data and social media analytics, 2015 to 2022.

What the sources say

Official outlook

BLS projects employment to decline 4.8% from 2025 to 2035.

BLS Employment Projections 2025-35

AI exposure studies

Data as of BLS OEWS May 2025; BLS Employment Projections 2025-35. BLS publishes these counts as yearly survey snapshots and cautions against reading them as a precise trend. In May 2021 BLS changed how it estimates these counts (a model-based method using three years of survey data) and moved fully to the 2018 occupation codes, so part of a change across 2021 can come from the method rather than the work.

Last updated .

Cite this page
APA

Future History. (2026, October 11). Survey researchers: jobs and pay since 1910 [Data set; data as of BLS OEWS May 2025; BLS Employment Projections 2025-35]. https://futurehistory.earth/timeline/19-3022

Chicago

Future History. "Survey researchers: jobs and pay since 1910." Data as of BLS OEWS May 2025; BLS Employment Projections 2025-35. Last modified October 11, 2026. https://futurehistory.earth/timeline/19-3022.

BibTeX

@misc{futurehistory_19_3022, author = {{Future History}}, title = {Survey researchers: jobs and pay since 1910}, year = {2026}, month = oct, howpublished = {\url{https://futurehistory.earth/timeline/19-3022}}, note = {Last updated 2026-10-11. Data as of BLS OEWS May 2025; BLS Employment Projections 2025-35}, }

The access date is today's date on your device.

drag to travel through time
19001925195019752000now
2026
Known then as Social investigator / charity visitor (reform-era social survey worker)
Known then as Social survey researcher / field investigator (social survey movement era)
Known then as Public opinion researcher / survey analyst (wartime and postwar institutionalization era)
Known then as Survey researcher / CATI researcher (telephone and computer-assisted era)
Known then as Survey Researchers (BLS SOC 19-3022)
Known today as Survey Researchers (BLS SOC 19-3022)
Projected to still be called Survey Researchers (BLS SOC 19-3022)
Data from the May 2025 survey
US Employment
8K
Est.

Stated in source: BLS OEWS

Official statistic
Report an error
Median Annual Wage
$69,460
Est.

≈ $67,679 in 2024 prices (derived by us, CPI-adjusted)

Official statistic
Report an error
Beat · 1907

The Pittsburgh Survey, funded by the Russell Sage Foundation in its founding year, deployed roughly 70 investigators across Pittsburgh for two years to document wages, hours, accidents, and housing conditions in the steel industry. It was the first systematic effort of its kind in an American city and became the template for the "social survey movement" that inspired more than 2,500 similar surveys nationally over the next two decades. Crystal Eastman wrote the labor and workers' compensation analysis; Lewis Hine documented conditions photographically. The findings helped end the twelve-hour day for steelworkers.

Beat · 1936

George Gallup's successful prediction of Franklin Roosevelt's re-election in 1936 directly refuted the Literary Digest's straw poll of 2.4 million respondents that had predicted Alf Landon's victory. Gallup used a probability sample of roughly 1,500 respondents; the Digest used a self-selected convenience sample drawn from automobile registration and telephone directories, which over-represented affluent Republicans. The episode became a founding case study in survey methodology pedagogy and established that sample quality matters more than sample size.

Beat · 1941

The National Opinion Research Center was founded at the University of Denver by Harry Field, with support from department-store heir Marshall Field III. Field had worked for the Gallup Organization and wanted to apply survey methods in the public interest rather than commercial contexts. NORC is today the oldest and largest university-based survey research organization in the US. After Field's death in a 1946 plane crash, NORC moved to the University of Chicago under Clyde Hart in 1947, the same year AAPOR was founded.

Beat · 1947

The American Association for Public Opinion Research (AAPOR) was founded on September 4, 1947, by a group that included George Gallup, Archibald Crossley, and Elmo Roper. The same meeting founded the World Association for Public Opinion Research (WAPOR). AAPOR became the home of Public Opinion Quarterly (which began in 1937 and became AAPOR's official journal in 1948) and established the field's first code of professional ethics. Its founding marks the moment survey research became recognized as a distinct profession with its own standards and scholarly literature.

Beat · 1971

Chilton Research Services developed the first CATI (Computer-Assisted Telephone Interviewing) system in 1971, integrating computers into centralized telephone calling operations. The technology roughly doubled interviewer productivity from four to eight completed interviews per hour. The US Census Bureau conducted its first major CATI test in 1982. By the mid-1980s, CATI systems supporting complex branching questionnaires were standard at commercial research firms and government statistical agencies, enabling the large-scale telephone survey industry that employed tens of thousands of interviewers through the 1990s.

Beat · 1999

SurveyMonkey launched in 1999, founded by Ryan Finley, with the explicit goal of making survey creation so simple and cheap that anyone could do it without a trained researcher. Greenfield Online, founded in 1994, had already pioneered the online access panel model, enabling self-administered quota samples without telephone interviewers. By the early 2000s, the economics of routine commercial survey research had fundamentally shifted: what previously required a professional researcher, a call center, and several weeks could now be fielded by a business user in days at a fraction of the cost. Online panels accounted for 29% of global quantitative research spending by 2011.

Beat · 2024

The AAPOR 2025 annual conference took the theme "Reshaping Democracy's Oracle: AI and the Future of Public Opinion Research," reflecting the profession's collective reckoning with large language models. Academic research published through 2024-25 (Versta Research, arxiv 2512.17455) documents systematic bias in AI-generated synthetic respondents: repetitive sequencing, uniform phrasing, and superficial personalization that undermine construct validity and external generalizability. The profession faces a paradox: AI tools reduce the cost of routine survey work, compressing employment, while simultaneously creating demand for methodologists who can specify when AI-assisted methods are appropriate and when they introduce unacceptable bias.

How to read this number

OEWS is a point-in-time survey snapshot, not a continuous time series; BLS advises against using it for year-over-year trend comparison.

Each dot is a cited figure over time; the dotted line only links them (values between aren't measured). Hollow dots are estimates.

Official statistic: Copied by a script from a statistics agency table or a published study's table, not retyped by hand. BLS figures and the forecast and exposure tables are also re-checked against that table on every data run. Researched context: Researched from the cited source rather than copied from it by a script, so it is not machine-checked yet. Spot a mistake? Each figure has a “Report an error” link.

Cited forecasts · present → 2035

Will AI replace survey researchers? What the research says

Each figure is stated by the named source; we publish no forecasts of our own. Where sources disagree, we show the disagreement rather than an average.

Each figure is stated by the named source; we publish no forecasts of our own. Where sources disagree, the spread is the honest signal.

What published forecasts say
Forecast change in the number of people doing this work, as published by each named source. Bars share one scale on every role, centred on zero and reaching the end at ±30%. Grey marks a small change, 3% or less either way.
BLS Employment Projections 2025-35
→ 2035
-4.8%
BLS National Employment Matrix, 2025-35 cycle, line for SOC 19-3022 (Survey researchers): 9.8 thousand jobs in 2025, projected 9.3 thousand in 2035, a change of -0.5 thousand (-4.8%). Matrix employment counts all jobs in the occupation, including self-employed workers (9.6% of this occupation in 2025). The average for all occupations in this cycle is +3.5%.
Official statisticReport an error
How exposed is this work to AI (research measures)
Scores as published by each study. Each row says what it measures. These measure what AI could technically do, not forecasts of jobs lost.
Eloundou et al. (2023), "GPTs are GPTs"
2023 study · Share of tasks exposed to LLMs (β, human-rated)
84.4%
of tasks
What this measures

Published occupation-level exposure from the paper's occ_level.csv for O*NET-SOC 19-3022.00 Survey Researchers. Shown: human-rated β = 0.844, the share of the occupation's tasks an LLM could do in at least half the time at equal quality, counting tasks that need LLM-powered software at half weight (β = E1 + 0.5 × E2). Other published measures for this occupation: human-rated α (LLM alone) 0.75 and γ (all LLM-assisted) 0.938; GPT-4-rated β 0.625. This measures how much of the work an LLM could touch, not how many jobs disappear.

Official statisticReport an error
Today, in this role

What's changing day to day for survey researchers

Three parts of your work where AI is already doing real lifting, and what stays yours.

AI is sitting alongside you hereDirect AI-assisted open-end response coding, defining the initial code frame and sentiment dimensions before fielding, running verbatim data through Caplena's LLM-based topic assignment and reviewing AI-generated theme clusters and confidence scores for accuracy and conceptual fit, identifying response categories the AI failed to classify correctly, and synthesizing the final coded results into a clean theme hierarchy for the analysis report.

Direct AI-assisted open-end response coding, defining the initial code frame and sentiment dimensions before fielding, running verbatim data through Caplena's LLM-based topic assignment and reviewing AI-generated theme clusters and confidence scores for accuracy and conceptual fit, identifying response categories the AI failed to classify correctly, and synthesizing the final coded results into a clean theme hierarchy for the analysis report.[6],[5]

Tools picking this up
Where your edge is

Caplena and equivalent tools (Yabble Count, SurveyMonkey Thematic Analysis) automate most open-end coding volume at human-level accuracy. The rest (responses that fall in unexpected categories, culturally specific references the AI misclassifies, and emergent themes that weren't in the original code frame) are where your analytical judgment matters. Build skills in grounded theory and thematic analysis: the frameworks that tell you whether an AI-surfaced cluster is a real insight or a statistical artifact of phrasing similarity.

AI is sitting alongside you hereDirect and quality-control survey data collection operations, managing interviewer briefings and training for complex instrument administration (government household surveys, sensitive-topic studies requiring trained interviewers), monitoring data collection quality using AI-powered response-quality flags (SurveyMonkey Response Quality, Qualtrics fraud detection) that identify straight-liners, speeders, and AI-generated bot responses, and updating field protocols in response to nonresponse pattern monitoring.

Direct and quality-control survey data collection operations, managing interviewer briefings and training for complex instrument administration (government household surveys, sensitive-topic studies requiring trained interviewers), monitoring data collection quality using AI-powered response-quality flags (SurveyMonkey Response Quality, Qualtrics fraud detection) that identify straight-liners, speeders, and AI-generated bot responses, and updating field protocols in response to nonresponse pattern monitoring.[2],[1]

Where your edge is

The interviewer-management layer of this role is contracting as AI-assisted self-administered surveys replace telephone and in-person fieldwork. Redirect your energy from managing field operations to managing data quality: AI flags suspicious responses, but human judgment determines when a pattern of flagged responses indicates a systemic sampling problem (a compromised panel source) vs. random noise. Frame yourself as a data-quality expert, not a field manager.

AI is sitting alongside you hereAnalyze survey data and produce research reports using AI-assisted analysis tools, directing SurveyMonkey's Analyze with AI or Qualtrics Insights Explorer to generate segmented cross-tabulations and pattern summaries, reviewing AI-generated findings for methodological soundness (sample size adequacy, significance thresholds, confounding), applying statistical weighting using R's survey or svyweight packages for probability samples, and writing the final narrative report that connects statistical findings to the policy or business question the study was commissioned to answer.

Analyze survey data and produce research reports using AI-assisted analysis tools, directing SurveyMonkey's Analyze with AI or Qualtrics Insights Explorer to generate segmented cross-tabulations and pattern summaries, reviewing AI-generated findings for methodological soundness (sample size adequacy, significance thresholds, confounding), applying statistical weighting using R's survey or svyweight packages for probability samples, and writing the final narrative report that connects statistical findings to the policy or business question the study was commissioned to answer.[5],[4],[1]

Where your edge is

AI analysis tools answer "what does the data show" but not "what does it mean for the decision." Every survey dataset has methodological limits (coverage gaps, nonresponse bias, construct validity questions) that a good researcher names explicitly in the report, and that an automated summary will paper over. Build the skill of methodological caveating: the ability to say "this finding is directionally reliable but not sufficient for a regulatory decision" is what separates a research scientist from a dashboard.

Where survey researchers go next

Natural next steps for someone with your foundation: not exits, evolutions.

What each move takes

Survey Researchers who move into commercial market research settings (marketing analytics teams, insights platforms, research agencies) find that most of their core skills (questionnaire design, sampling, statistical analysis, report writing) transfer directly. The gap is commercial framing: instead of designing for methodological defensibility in an academic or government context, you design for business decision speed and stakeholder narrative. The BLS projects stronger employment demand in the commercial research sector than in the academic/government survey context where the decline is concentrated. This is the highest-payoff pivot for survey researchers in declining government or nonprofit roles.

What you'd add
  • · Commercial research platform fluency: Qualtrics, SurveyMonkey, Quantilope for rapid quantitative methods
  • · Business framing: translating research findings into strategic recommendations, not methodological reports
  • · AI tool adoption: Yabble for open-end coding, Outset.ai for AI-moderated qual, Cint for commercial panel management
  • · Stakeholder presentation: communicating research confidence and limitations to non-methodologist audiences
  • · Market research pricing and project scoping for commercial engagements
What it takesMost of your skills carry over
Tools of the era

How has the work of survey researchers changed since 1880?

The tools that defined the work, one line per era. Select an era to see how it reshaped the work.

  • Paper forms and in-person interview (social survey movement era)

    The foundational technology of the profession was a paper form and a trained investigator willing to knock on doors. The Pittsburgh Survey of 1907 exemplified the model: investigators used structured schedules (early questionnaires) to record wages, hours, housing conditions, and nationality in a consistent format that could be tabulated by hand. Coding and tabulation were physical operations performed by teams of clerks working with mechanical counters and ledger sheets. Data quality depended entirely on interviewer skill, and there was no technology between the question asked and the answer recorded.

    Effect on the work

    The paper-and-canvass model set a ceiling on the scale of any single study. The Pittsburgh Survey employed many investigators and took two years to produce usable findings. This labor intensity was the primary reason most social surveys remained small, local, and project-funded rather than becoming routine organizational practices.

  • Probability sampling and telephone polling (Gallup era)

    George Gallup founded the American Institute of Public Opinion in Princeton, New Jersey, in 1935 and in 1936 successfully predicted Franklin Roosevelt's re-election using random-probability sampling methods, directly disproving the Literary Digest's straw-poll prediction based on a huge but biased sample. Gallup's innovation was methodological, not mechanical: he demonstrated that a carefully drawn sample of more than a thousand respondents could represent the entire US electorate more accurately than a self-selected sample of millions. The telephone became the medium of choice by the 1940s, enabling interviewers to reach a representative sample without traveling to respondents' homes. NORC was founded in Denver in 1941 by Harry Field specifically to institutionalize this model for public-interest research, moving to the University of Chicago after Field's death in 1947.

    Effect on the work

    The telephone-interview model dramatically reduced the cost per completed interview compared to in-person canvassing and enabled national-scale surveys to be run from centralized offices. It also created the template for a bifurcated survey workforce: professional researchers who designed and analyzed studies, and a larger pool of telephone interviewers who administered them.

  • CATI (Computer-Assisted Telephone Interviewing, from Chilton Research Services 1971)

    Computer-Assisted Telephone Interviewing originated in 1971 when Chilton Research Services integrated computers into centralized telephone calling operations, roughly doubling interviewer productivity from four to eight completed interviews per hour. The US Census Bureau conducted its first major CATI test in 1982. By the late 1980s, CATI systems had spread across commercial market research firms, government statistical agencies, and academic survey centers, supporting complex branching questionnaires with hundreds of items on networks of workstations. CATI removed the paper form from the interview: the interviewer read questions from a screen and typed responses directly, reducing data entry errors and enabling real-time quota monitoring. The technology fueled a boom in telephone survey research through the 1980s and early 1990s and supported the growth of large call-center-based research organizations.

    Effect on the work

    CATI enabled a substantial expansion of the commercial survey research industry. But it also concentrated the interview workforce in large call centers that were optimized for high throughput, laying the organizational groundwork for the coming transition to internet-mediated self-administered surveys.

  • Online self-administered surveys (SurveyMonkey 1999, Greenfield Online panels, Zoomerang)

    SurveyMonkey launched in 1999, founded by Ryan Finley, with an explicit goal of making survey creation so cheap and easy that it no longer required a professional researcher. Greenfield Online, founded in 1994, built the first scalable commercial online access panel, enabling researchers to field nationally representative samples without telephone interviewers. By 2006, online panels accounted for a sizeable share of global quantitative research spending. Response rates in telephone surveys, which had already been declining, collapsed further in the mid-2000s as caller-ID adoption and cell-phone-only households eroded the RDD sampling frame. Pew Research Center, which had built its reputation on telephone polling, began transitioning its national surveys toward online methods by the mid-2010s. The shift fundamentally disrupted the economics of professional survey research: studies that had required a team of researchers, a call center, and weeks of fieldwork could now be fielded by a single analyst through a self-service platform in days.

    Effect on the work

    The internet-panel transition eliminated the call-center model of professional survey fielding and with it a significant share of the mid-tier survey research workforce. BLS employment data for 19-3022 shows a decline from the early-2000s baseline that tracks this structural shift.

  • Big data and social media analytics (passive data era)

    The arrival of large-scale social media data and behavioral tracking from digital platforms began challenging the primacy of primary survey research. Platforms like Twitter/X, Facebook, and Google made it possible to analyze billions of opinions, behaviors, and preferences without fielding a single survey. Research firms like Ipsos launched Synthesio (monitoring 800 million sources), and the consulting industry began offering social listening reports as substitutes for, or supplements to, traditional survey programs. The BLS explicitly cites this development as a driver of the projected employment decline for Survey Researchers 2024-2034: "the ongoing adoption of data mining and collecting information from social media sites are expected to lessen the need for traditional survey methods." The professional value proposition of survey researchers became increasingly concentrated in the tasks passive data cannot address: causal inference, probability sampling, IRB-compliant research design, and politically and legally sensitive topics where self-report from a structured instrument remains non-negotiable.

    Effect on the work

    The passive-data era did not eliminate survey research but created substitutes for the most routine commercial research applications, accelerating the occupational decline that internet panels had begun. Researchers who could work across both active and passive data sources became more valuable; those whose expertise was confined to traditional fieldwork faced the sharpest displacement pressure.

  • AI-assisted survey design and synthetic respondents (Qualtrics AI, Caplena, Fairgen)

    Large language models accelerated automation within the core of the survey researcher's workflow. Qualtrics Edge Audiences (2023) offers a fine-tuned LLM for synthetic pre-testing that achieves 12x greater accuracy than general GPT/Gemini on attitudinal scales. SurveyMonkey Genius can draft a complete survey instrument from a plain-text brief in seconds. Caplena automates open-end response coding at human-level accuracy, which previously required a team of coders spending days on a single dataset. Fairgen offers synthetic respondent augmentation with statistical guarantees for hard-to-reach populations. The AAPOR 2025 annual conference was themed around AI as a paradigm shift in public opinion research. The emerging division of labor concentrates human expertise on the methodological decisions AI cannot make: specifying what population a study must represent, designing for IRB compliance, validating when AI-generated survey responses introduce systematic bias, and adjudicating what evidence standard is required for a given policy decision.

    Effect on the work

    Academic research (Versta Research 2024) documents systematic quality problems in AI-generated survey responses: repetitive sequencing, uniform phrasing, and superficial personalization. This finding paradoxically reinforces the value of trained survey methodologists who can specify when synthetic data is methodologically legitimate and when real probability samples remain non-negotiable. The profession is contracting in volume while the remaining core becomes more technically demanding.

Part of Science · see all 27 roles →
Different role?

See the same long-arc view for your own profession.

Browse the directory by industry, or search by title or SOC code.

Browse all roles

Survey Researchers employment and pay by year

On record since1880
Latest US employment8,290 (May 2025)Source: BLS OEWS May 2025Official statisticReport an error
Latest US median pay$69,460 (May 2025)Source: BLS OEWS May 2025Official statisticReport an error
Outlook-4.8% by 2035Source: BLS Employment Projections 2025-35Official statisticReport an error
View all 27 years of cited US data, each linked to its source
YearUS employmentAnnual pay (median unless marked)SourceReport
19103,000n/aEst., Wikipedia Researched contextReport
19478,000n/aEst., AAPOR Researched contextReport
1965n/a$7,200Est., University of Missouri Libraries Researched contextReport
198018,000n/aEst., Wikipedia Researched contextReport
200316,850$24,600BLS OEWS May 2003 Official statisticReport
200419,480$26,490BLS OEWS May 2004 Official statisticReport
200521,650$31,140BLS OEWS May 2005 Official statisticReport
200624,140$33,360BLS OEWS May 2006 Official statisticReport
200722,140$36,820BLS OEWS May 2007 Official statisticReport
200821,100$36,220BLS OEWS May 2008 Official statisticReport
200920,300$35,380BLS OEWS May 2009 Official statisticReport
201017,850$36,050BLS OEWS May 2010 Official statisticReport
201117,060$40,150BLS OEWS May 2011 Official statisticReport
201217,370$45,050BLS OEWS May 2012 Official statisticReport
201317,370$47,720BLS OEWS May 2013 Official statisticReport
201415,410$49,760BLS OEWS May 2014 Official statisticReport
201513,650$53,920BLS OEWS May 2015 Official statisticReport
201611,930$54,470BLS OEWS May 2016 Official statisticReport
201711,270$54,270BLS OEWS May 2017 Official statisticReport
201811,690$57,700BLS OEWS May 2018 Official statisticReport
20199,930$59,170BLS OEWS May 2019 Official statisticReport
202010,350$59,870BLS OEWS May 2020 Official statisticReport
20218,850$59,740BLS OEWS May 2021 Official statisticReport
20227,880$60,410BLS OEWS May 2022 Official statisticReport
20238,190$60,960BLS OEWS May 2023 Official statisticReport
20247,720$63,380BLS OEWS May 2024 Official statisticReport
20258,290$69,460BLS OEWS May 2025 Official statisticReport
Embed this timeline on your site

Free for any site. Paste this where the timeline should appear; it stays interactive, every datapoint stays cited, and it sets no cookies on your page. How embedding works

<iframe src="https://futurehistory.earth/embed/19-3022"
  width="100%" height="430" style="border:0"
  title="Survey Researchers, a Future History timeline"
  loading="lazy"></iframe>
<p style="font-size:13px;margin:4px 0 0"><a href="https://futurehistory.earth/timeline/19-3022">Survey Researchers: jobs and pay over time</a>, from Future History</p>

See all roles in Science