Loading page content
Preparing content…
SOC 19-3022
Survey researchers design and conduct surveys, clarify the research question, the sample and the questionnaire, monitor collection quality, and then turn the answers into checkable tables, analysis and method notes. Interviewer execution, questionnaire design and statistical interpretation work together but cannot be mixed into a single responsibility. Use the fictional records first to test whether you are willing to handle missing data, check the denominator and constrain the conclusion.
Model impact index for three equally weighted representative core tasks; not a measured share of working time or a job-loss probability. floor((50+25+75)/3+0.5)=50; fermatmind-ai-impact-v1; 2026-10-18.
See whether you are willing to turn a research question into a testable questionnaire and sample design, trace missing data and the denominator, and state clearly what the material supports and what it does not. Communication involves clients, interviewers and analytical colleagues; patience in revision and evidence discipline are more worth experiencing than simply liking to ask questions.First choose the task object: survey design emphasises measurement and sampling, collection management emphasises interviewers and completion status, and data analysis emphasises coding, statistics and interpretation. This occupation may cover several stages, and conducting interviews is not the same as independently owning the research design.
This is a career exploration screening, not hiring, a statistics competence certification or a personality diagnosis. It also needs to be combined with training background and real post requirements.
The signal that this is worth trying first is a willingness to turn 'what do I want to find out' into a clear research question and to check whether the wording, the sample source and the answers really support the conclusion, rather than chasing attractive charts.More specific signals are: being able to state the target population and the question first, then check whether the items make respondents understand the same thing; checking completion status, coding rules and the analytical denominator, and going back to the material when conflicts appear instead of hiding them. On receiving feedback, being able to say which rule was changed and why, and leaving unconfirmed samples and explanations as unknown. Organising the method record on time and being honest about your own statistical experience are also important signs of reliable cooperation.
Survey researchers plan, develop and conduct surveys, analyse and interpret the material, clarify objectives, test questionnaires, monitor collection and produce result summaries and method records. The duties may cover design, collection management and analysis; interview execution and population inference must be distinguished by the actual division of work and the evidence.
Survey design and survey implementation are two connected work lines: the first determines measurement and sample, the second puts the plan into collection, quality checks and analysis. What follows is an editorial process based on the tasks of this occupation, not a uniform national operating procedure.
Career direction comparison
First distinguish the tasks inside this occupation from independent adjacent occupations. Design concerns what to ask, whom to ask and how to infer; collection management concerns obtaining material through the procedure and its quality. Clinical coordination, management consulting, market research and financial quantitative analysis each have their own objects and deliverables; sharing tables or interviews does not mean they can be swapped directly.
| Career direction | Core work and outputs | Key difference | Which choice fits better |
|---|---|---|---|
| This occupation: survey design and methods | Clarify the research question, measurement and sample design, test the questionnaire and record the sampling and weighting decisions. | The focus is whether the material can answer the question, which requires stating the target population, sampling and measurement limitations; it is not just writing the items smoothly. | Willing to trace the problem definition, test comprehension differences and accept several rounds of revision. |
AI's impact on the occupation
Text, coding and analysis assistance can be explored, but questionnaire validity, sample source, missing-data handling and the scope of conclusions still need human checking. Cross-occupation exposure is not a substitution rate for this occupation, and this page has no representative adoption share or reviewable FermatMind task score. Test the specific task first, then decide how to use it. AI Impact Score 50% is an editorial estimate. Model impact index for three equally weighted representative core tasks; not a measured share of working time or a job-loss probability. floor((50+25+75)/3+0.5)=50; fermatmind-ai-impact-v1; 2026-10-18.The three tasks, ratings and sources are listed in the FermatMind method source record on this page.
FermatMind AI Impact Score estimation method v1
FermatMind / 费马测试 · 2026-10-18
Mainland China salary reference
China salary basis
Classification of public materials
Insufficient evidence: a nationally comparable wage distribution for survey researchers in China. Two historical institutional announcements were checked, one CGSS internship with an undisclosed allowance amount and one adjacent research assistant post disclosing a basic salary; they are not a national sample and cannot be compared directly with the U.S. statistics for this occupation.
Basic salary 8,000 yuan / monthThis is the basic salary of the adjacent Ocean University research assistant post, with performance calculated separately, after-tax pay not given, and the recruitment closed.
U.S. salary reference
The U.S. annual wage median for this occupation in 2025 is 69,460 dollars, and the O*NET wage page notes the BLS 2025 source; the same page gives an hourly median of 33.40 dollars. The table below uses the directly published annual values and does not work backwards from rounded hourly rates, and it is not a take-home amount or an income promise.
| Wage position | Annual wage reference | Personal income boundary | How to understand this position |
|---|---|---|---|
| 10th percentile | 39,260 dollars / year | The source publishes the annual value directly; it is not a personal take-home amount | The 2025 10th percentile annual wage is 39,260 dollars; this describes a lower position in the wage distribution of employees in this occupation, not a newcomer level, a statutory minimum or a hiring guarantee. |
| National annual wage median | 69,460 dollars / year |
Career fit guide
Fit requires looking at interest in the research question, rules and detail, statistical understanding, accepting revision and communication arrangements together. Personality and ability tests can only provide discussion clues; the clues below are turned into observable tasks, with no type or score threshold.After the experience, see whether you are willing to trace inconsistencies, record revisions and admit unknowns. Extraversion, personality type and a single test score cannot replace questionnaire, data and method training.
Worth exploring further： Observe whether you are willing to check questionnaire and data rules repeatedly and can explain the basis for each revision; liking to ask questions alone cannot decide suitability.
First check the real conditions： Research interest cannot replace a real living budget. Judge sustainability from the real contract, the pay composition and the work arrangements, and do not promise based on unconfirmed performance pay or a future project income.
First do a small trial experience： Use the six fictional records in F1 to judge the eight claims independently, then revise separately for the F2 A and B feedback. Observe whether you are willing to recheck the denominator, admit the missing data and constrain the conclusion; pen and paper are enough.
Work pressure, risks and occupational boundaries
The main things to check are participant and material protection, measurement and sampling bias, denominator and coding errors and over-interpretation. Fieldwork and employment arrangements are checked against the post, and method changes and publication authority are confirmed by the project division of work; a results table does not prove the survey is reliable.
The following is compiled from material on this occupation, the two historical recruitments and AAPOR professional ethics; the operational suggestions are editorial applications, not an accident rate or a personal score. AAPOR is not a Chinese law or a competence certification body, and the U.S. occupational description cannot replace specific Chinese rules.
Typical scenarios：Planning to publish raw rows or upload them to tools after obtaining answers, without stating the purpose and the permitted scope separately.
Main posts：Research and collection staff handling questionnaires, interview records and identifiable material
Entry, certificates and career development
First make the question, the coding, the denominator and the conclusion explainable, then progressively learn questionnaire, sampling and statistical methods and revise with real feedback. When entering a specific internship or post, check the duties, conditions and data permission; one small exercise cannot replace formal training and whole-project experience.
Career outlook and related occupation transitions
The precise U.S. 2025—2035 projection for this occupation falls 4.8%, with average annual openings of about 600; a decline and replacement opportunities can coexist, and this does not mean annual net growth or guarantee personal hiring.Moving to clinical coordination, consulting, marketing, financial quantitative analysis, statistics, social science assistance, sociology or data science all requires verifying the target output, the professional training and the post authority again; using data does not mean they can be swapped directly.
United States · Occupational statistics for survey researchers, with the SOC corresponding to the O*NET code
−4.8%
Overall employment change
BLS notes that data mining, social media data and big data improve market research efficiency and may reduce the need for traditional telephone and face-to-face surveys; the background explanation cannot on its own be converted into a generative AI substitution figure or an equivalent change in China.
Design and test questionnaires, coordinate and monitor collection, analyse material, and produce chart summaries and method records. The specific division depends on the project, and interview execution is not the same as independent research design.
This occupation centres on the survey question, sample, questionnaire and interpretation; market research faces business questions, statisticians focus on methods, and clinical coordination handles clinical procedures. Shared tools do not make the responsibilities directly interchangeable.
Choose a real task from this page for a small practice project or a conversation with someone in the field. Use the planning guides to set a date, an output and feedback questions before deciding whether to continue.
Judge research interest and working conditions separately. Questionnaire testing, data cleaning and method records often need repeated revision; travel, interviewer coordination and project timing should be checked against the specific post, without assuming that everyone does long-term fieldwork.Signals needing caution include: equating more answers collected with representativeness, treating a significant result as the only value, filling missing answers in at will as the expected answer, or wanting only to produce charts without explaining the process. Also assess public communication, repeated quality checks and fieldwork arrangements separately; liking social contact cannot replace questionnaire and statistics training, and disliking a lot of social contact cannot automatically rule out the analysis direction.
For the safe experience, use the six fictional book-exchange records in the entry module F1, judge the eight claims independently, then revise the original draft in two places each for the F2 A and B feedback and fill in F3. Pen and paper are enough; do not contact real residents, collect private information or fabricate interviewees.After finishing, record which denominator judgment went wrong, what specific revisions the two sets of feedback produced, and whether you are willing to check one more set of material. A single fictional exercise only supports the next exploration step and cannot impersonate a real survey or a professional qualification.
An executable questionnaire, sample design and method notes
Checkable data, summary charts and an analytical report
The U.S. occupational identity on this page is O*NET-SOC 19-3022.00 Survey Researchers: planning, developing and conducting surveys and analysing and interpreting the material. It may include social science researchers focused on questionnaire design or survey team supervision; market research, statisticians and interviewer execution each have their own boundaries. The design side decides what must be measured, whom to ask and how to ask; interviewers contact, ask and record according to the confirmed procedure, and collection management handles coordination and quality tracking. Survey researchers may supervise these stages, and the specific division depends on the project; the table below compares responsibilities and cannot be used to define a uniform job grade.
| Dimension | Main directions | Comparison directions |
|---|---|---|
| Core purpose | Form a measurement and sample design that can answer the research question, clarifying the survey subjects, the items and the interpretation boundary. | Obtain answers under the confirmed procedure and record completion status accurately, without changing what respondents express. |
| Main object and basis | The research objective, the target population, the sample frame, existing research and questionnaire pretest results. | The confirmed questionnaire, contact procedure, collection arrangements, interview records and participant notes. |
| Core process | Background research, designing and testing the wording, selecting the sample and mode, then reviewing the measurement and analysis plan. | Receive training, contact and ask according to the procedure, record answers and refusals, and report anomalies and progress. |
| Main outputs | Questionnaire versions, sampling and collection plans, coding rules and analysis method notes. | Raw answers, complete or partially complete status, anomaly records and collection progress material. |
| Responsible subject | The research and method leads confirm the design and interpretation, and survey researchers propose, test and record decisions within the actual division of work. | Interviewers are responsible for accurate execution and recording; questions beyond the procedure go to the collection supervisor or research lead for confirmation. |
| How support works | Clarify the needs with the client and verify with the collection team whether the items, the procedure and the resources are feasible. | Report comprehension difficulties and execution anomalies through the procedure, and without confirmation do not change items or choose answers for respondents. |
| Delivery boundaries | A sound design does not guarantee the absence of non-response or measurement error, and the analysis must still check the data and the limitations. | Completing one interview does not prove that the sample represents the population and does not grant authority to publish research conclusions independently. |
Understand client needs and existing research, and determine the questions to answer, the data sources and the survey scope.
Suggest and test wording, options and procedures, and record revisions after finding comprehension or measurement problems.
Plan and coordinate single or multiple surveys, organise training and procedure communication, and review the collection work.
Check sample completion and response calculations, evaluate interview material and follow up anomalies.
Process the data according to the objective and data rules, produce tables, charts and fact summaries, and explain limitations.
Record the questionnaire, collection, sampling and weighting decisions, keep researching survey methods, and do not replace method notes with charts.
Every judgment must be traceable to the research objective, the sample and the records. What follows is an editorially designed work self-check; you cannot jump from one rate, model result or chart straight to a conclusion about the whole population.
Questionnaire judgments must return to the research question: are the wording, the options, the routing and the time frame consistent, and does an item lead the respondent or mix several questions? The editorial exercise checks the task text and the coding first and does not treat your own preference as the respondent's actual answer.
Material inputs include the research objective, the target population, the available sample frame, the questionnaire, the collection method and existing data. First distinguish the population you want to describe from the population you can actually reach; convenience-collected answers cannot automatically represent the population.
When monitoring collection, check the sample completion status and the response calculation, distinguishing complete, partially complete, refused and unknown eligibility. A rate must have a clear numerator, denominator and rule, and a high completion rate alone cannot declare that the material has no bias.
| This occupation: collection management and data quality | Coordinate collection, review interview work, check sample status, coding and response calculations, then produce a reviewable summary. | The focus is plan execution and material consistency; interviewers completing interviews cannot directly stand in for population interpretation. | Willing to handle repeated checks and anomaly communication and to record rules and versions clearly. |
|---|
| Clinical research coordinators | Coordinate the procedures, records and compliance tasks of clinical research; general survey experience cannot replace clinical project training and authorization. | It involves clinical research procedures and compliance records, and ordinary questionnaire project experience does not automatically grant clinical task authority. | Interested in research execution and participant management, and willing to check the target institution's clinical training requirements. |
|---|
| Management consulting analysts | Study organisational problems and processes and propose efficiency and management improvements; research communication shares a foundation, but the delivery object and responsibility differ. | The output is advice on organisational processes and management, and one survey report cannot prove all consulting ability. | More focused on organisational problems and implementing improvements, and willing to build process diagnosis and advisory communication. |
|---|
| Market research analysts specialising in marketing | Study markets, consumer demand and competition to support marketing decisions; a questionnaire is only one possible information source. | Facing consumer demand, competition and marketing decisions, where a survey is only one method; market and business interpretation ability must be built. | Wanting to apply research results to business problems, and willing to learn the market context and decision-making use. |
|---|
| Financial analysts | O*NET 13-2099.01 refers to financial quantitative analysis tasks focused on investment, valuation and risk models; the local name 'financial analyst' should not be expanded to all finance posts. | The O*NET object linked here is financial quantitative analysis, delivering investment, valuation or risk models; it is not the same as direct promotion from a survey occupation. | Willing to build finance, mathematics and modelling and to verify the transition with the target outputs. |
|---|
Use the six fictional records first to experience denominators, missing data and conclusion boundaries, then judge whether you prefer design, checking or interpretation. When moving to an adjacent job, prove the new ability with the target tasks' outputs; the original method experience can serve as a base but does not automatically meet the professional knowledge, project authority or employer conditions.
After choosing the task, judge whether AI assists the questionnaire draft, the coding or the analysis, and check the input permission, the errors and the human responsibility separately; a general trend cannot replace analysis of the specific post.
Editorial ratings use the existing occupational responsibilities and AI task analysis on this page. They are not adoption rates, measured effects, job-disappearance or personal-unemployment probabilities, or official scores from Stanford or another external institution. No measured time weights or future robot deployment are assumed.
At the technical ability level, you can try using text tools to organise question lists, propose alternative wording or write coding drafts. These are editorial suggested uses and do not prove that the questionnaire measures better. You must return to the research objective, the original records and human testing; generating reasonable text is not the same as a representative sample or correct analysis.
At the actual adoption level, this page obtained no separate representative AI usage rate or net time-saving study for survey researchers. Professional discussion, cross-occupation exposure indicators and one institution's tool trial cannot replace an adoption share. The task, the input, the human check and the deployment context should be stated separately, and technical feasibility cannot be used to infer widespread use.
At the employment outcome level, the precise U.S. projection for this occupation falls 4.8% over 2025—2035, with average annual openings of about 600. BLS notes that methods such as data mining and big data may reduce the need for traditional surveys; the projection includes many factors and cannot isolate job losses caused independently by generative AI. Personal opportunity and the Chinese market need separate evidence.
| Work direction | Task | Current change | Human control points |
|---|---|---|---|
| Questionnaire preparation | Organise the research question and candidate items | Editorially suggested use | Check the item meaning, options and pretesting; a draft does not prove the measurement is valid. |
| Coding open-ended answers | Propose a classification and coding draft | Must be validated by a person | Check classification consistency against the actual records, keeping the misclassifications and revisions. |
| Data cleaning | Propose checking scripts or an anomaly list | Editorially suggested use | Check the rules, the missing data and the original records; real answers cannot be changed automatically. |
| Statistical analysis | Draft the calculation steps and the interpretation | The method must be reviewed | Check the denominator, sample, weighting and uncertainty; it is not enough that the code runs. |
| Report writing | Organise chart captions and the method summary | Editorially suggested use | Check every sentence against the data and the methods, and do not generate sources or conclusions that do not exist. |
| Collection management | Organise the interview status and the follow-up items | Responsibility is not transferred | Record complete, partial and refused according to the confirmed procedure, and refer changes to the responsible person. |
| Occupational judgment | Compare tasks, authority and post conditions | The net impact is not quantified | Distinguish exposure, adoption and the employment projection, and make no promises about personal opportunity. |
Below, the tasks of this occupation, the employment projection, cross-occupation research and generic risk methods are listed separately. Check the research object, the year and the scope first; global exposure, occupational decline and tool capability cannot be converted into one another. All operational suggestions for this occupation need real data and human review, and matters without figures remain unknown.
Questionnaire drafts, open-ended coding and simulated answers are different tasks. The first two must be verified for their effect on measurement and classification; simulated material must be kept separate from people's actual answers. The editorial exercise uses only the fictional records in the task text and cannot add generated answers into the completion count.
What AI mainly changesGenerate candidate wording, which still needs the objective and testing.
Still the responsibility of peopleCheck the item meaning, the options and the sample design, and record the pretest revisions.
What AI mainly changesPropose code or classifications, which may introduce misclassification and wrong interpretation.
Still the responsibility of peopleCheck the original records, the rules, the denominator and reproducibility, and state the limitations.
What AI mainly changesCan organise status and text but does not replace participant and project responsibility.
Still the responsibility of peopleRecord people's answers accurately and state the data source and research scope.
First determine the research question and the available materials, check the output, record the human changes, and then decide whether it can be delivered. Real data, procedure changes and public reports must be confirmed by the responsible person with authority on the project; the control process below is an editorial application, not a new statutory licence.
First state which questionnaire, coding or report task is being assisted.
Check the input source, permission and identifying information, and use only the approved scope.
Real rule or purpose changes go to the research and data lead for review.
Check the records, formulas and methods, and do not treat fluent wording as evidence.
Record the actual tools, changes and verification; if AI was not used, write truthfully that it was not.
Generated content mixed into real completion counts distorts the data source and the conclusion.
Control method:Mark simulated and actual material separately; this page adds no fictional respondents.
The input may exceed the participant notes, institutional authority or the tool's permitted scope.
Control method:Check the purpose and permission and control identifying information; a service of unknown status is not treated as a private space.
A plausible-looking calculation may mix missing, partially complete and valid answers.
Control method:A person rechecks the rules and the original rows, keeping the calculations before and after the change.
Generated citations or population conclusions may have no material support.
Control method:Open the original source and mark facts, inferences and unknowns separately.
Cross-occupation indicators and total projections cannot determine an individual's hiring or loss.
Control method:Compare by task and evidence scope, and keep the gap in adoption and net impact for this occupation.
Complete the fictional record exercise, the two revision versions and F3 first, showing your real judgment process; record tool use truthfully and do not impersonate actual survey experience.
Clarify the assisted task and the material permission, spot-check the coding, formulas and interpretation, and record the model's use and the human revisions; method decisions are still confirmed by the project responsibility chain.
Check the variable rules, the missing data and the denominator first, and be able to explain each processing step; without confirmation do not change questionnaires or invent real answers.
Arrange task-level validation, material permission and review roles, and explain the methods and limitations in external reports; a full-time decline projection is not treated as a project substitution basis.
It can propose a draft, but validity must return to the research objective, the item meaning, the options and pretesting. O*NET lists suggesting and testing wording as a task of this occupation, and generated text cannot be treated as a test result. The specific way it assists is an editorial suggestion, with no verified improvement percentage.
Source:O*NET survey researchers 19-3022.00Original material ↗
No. The ILO cross-occupation task exposure indicator describes potential technological impact, not actual adoption or a probability of unemployment for this occupation. This page does not enter the global share as a score for this occupation and does not infer from exposure that a particular person will lose their job.
Source:ILO generative AI and work: task exposure researchOriginal material ↗
No. The precise occupational projection includes many factors and cannot separate out the generative AI effect on its own; the average annual 600 openings come from replacement and other demand and cannot be treated as net annual growth or a substitution figure.
Source:BLS EP precise occupational projection tableOriginal material ↗
The O*NET tasks for this occupation include questionnaire testing, collection monitoring, material analysis and method records, but they do not provide improvement percentages after AI use. Being able to draft questions still requires checking the actual measurement, sample and interpretation.
Source:O*NET survey researchers 19-3022.00Original material ↗
BLS describes the background of methods such as data mining and the change in traditional collection needs; it does not prove that all surveys will be eliminated. Compare the specific research question, whether the material can cover the target population and the method limitations, and do not mix big data and generative AI as the same kind of evidence.
Source:BLS OOH survey researchersOriginal material ↗
This exercise does not allow generated answers to be treated as actual completion records. Methodologically, the data source, coding and simulation purpose must be stated; this limitation is an editorial exercise rule, not a specific law that NIST issued for this occupation. Keep the missing data and the human review records.
Source:NIST AI 600-1 generative AI risk profileOriginal material ↗
Check the processing purpose, the material permission, the identifying information and the tool's storage purpose first, then arrange output checking and confirmation by the responsible person. NIST is a general risk reference; the specific law, participant agreement and institutional rules still have to be checked separately. Tool access does not prove the right to upload.
Source:NIST AI 600-1 generative AI risk profileOriginal material ↗
Next, test the tool against the specific task: does it help find omissions, does it introduce errors, can the human review be traced? Keep the record of the model's actual use and non-use, then observe the working conditions, and do not treat tool output as a transfer of research responsibility.
The Renmin University CGSS internship announcement gives no allowance amount, and the formal labour dispatch post salary at Ocean University cannot be applied to it or inferred from the number of working days.
Status of the historical cases
| Project and scope | Quotes and disclosure status | Applicable scope and gaps |
|---|---|---|
| Historical Ocean University research assistant recruitment | Labour dispatch, master's degree or above; basic salary 8,000 yuan per month, with performance calculated separately. | The master's degree or above requirement belongs to that round of Ocean University research assistant recruitment; the dispatch unit and the using unit are checked separately and it does not represent a national threshold. |
| Renmin University CGSS project assistant internship | The allowance is basic pay plus performance, with the specific amount not given; the published attendance period has ended. | The Renmin University announcement is an internship whose attendance period has ended; the allowance composition is stated but the amount is unknown, and it is not used as a current post count or a national wage. |
| Case comparability | Two institutions, different tasks and employment forms; this is not an education-pay band. | The two announcements do not constitute a third institutional sample. There is no national wage table for the same duties and the same population, so no uniform bands by education can be set. |
| Precise national occupational wages in China | Insufficient evidence: no comparable national distribution was obtained. | Insufficient evidence: a national precise occupational wage distribution. One historical announcement from each of two institutions can only describe specific posts and does not compute recruitment heat, a national median or an income trend. |
The Renmin University CGSS 2026-05-27 internship announcement lists tasks such as questionnaire standardisation, system testing, training and liaison, with an allowance consisting of basic pay plus performance but no amount given. The published attendance period runs from 2026-06-01 to 09-30 with at least three working days per week; that period has ended, so it cannot be used to confirm that it is still recruiting.
When comparing posts, check the contracting entity, the employment form, the basic salary or internship allowance, performance pay, social insurance and housing fund and the work requirements separately; similar job titles do not guarantee the same duties or statistical object.
Basic salary, separately calculated performance pay and personal take-home income are different quantities. The Ocean University announcement gives no after-tax value or uniform performance amount, so 8,000 yuan per month cannot be treated directly as take-home income or extrapolated to national master's-level pay.
The employment form, workplace and time should be checked item by item. The Renmin University internship and the Ocean University labour dispatch arrangements differ, and one public amount cannot be used to infer a continuing contract, promotion or the pay of every project.
This page has no material supporting a national pay reduction share for Chinese survey researchers caused by AI. A tool assisting one step does not mean the whole post's pay changes by the same proportion; check the specific task, the division of staff, the quality requirements and the pay agreement first. U.S. projections and cross-occupation exposure cannot be converted into a Chinese personal income change.
| The source publishes the annual value directly; it is not a personal take-home amount |
| About half of the statistical employees are below and half above the median; this cannot be used to infer starting pay, after-tax pay or each person's actual full-year income. |
| 90th percentile | 130,860 dollars / year | The source publishes the annual value directly; it is not a personal take-home amount | The 2025 90th percentile annual wage is 130,860 dollars; this is a percentile position, not a guaranteed rate for experienced people, and it does not guarantee the total income of any contract or self-employed project. |
|---|
Pay · 10th percentile：39,260 dollars / year；The 10th percentile annual wage for this occupation, not a novice grade.
Pay · 25th percentile：51,270 dollars / year；The BLS 2025 annual values published directly on the O*NET wage page, not education or a job grade.
Pay · Median：69,460 dollars / year；The annual wage median for this occupation, not take-home pay or starting pay.
Pay · 75th percentile：99,760 dollars / year；The BLS 2025 annual values published directly on the O*NET wage page, not education or a job grade.
Pay · 90th percentile：130,860 dollars / year；The 90th percentile annual wage for this occupation, not a guarantee for experienced people.
This page uses the annual values published directly on the wage page. The hourly value can serve as a separate reference, but it cannot be used to assume 2,080 paid hours per person per year; your own income still depends on the contract, hours, taxes and fees and employment status.
OEWS wages exclude the self-employed, and regular wages differ from personal take-home income; the employment projection covers a different population from the wage statistics, and the two cannot be mixed to estimate total occupational income.
A percentile describes the wage distribution of employees in this occupation, not years of service, education or job grade. Institutions, industries, regions and individual contracts differ, and the 90th percentile cannot be labelled directly as a guaranteed rate for experienced people; jobseekers still need to check the real tasks and conditions.
Compare the contracting entity, the duties, the annual or hourly basis, the actual hours and the taxes and fees. Published annual wage statistics, post quotes and personal receipts are different quantities; income choices must first keep the occupation and the population scope consistent.
| Industry | Annual wage median / statistical status | How to understand it |
|---|---|---|
| Research and development services | 98,470 dollars / year | The annual wage median for survey researchers in research and development services is 98,470 dollars, and the statistical object is employees in this occupation within the industry, not all research and development posts or an individual quote. |
| Educational services (public and private) | 67,000 dollars / year | The annual wage median for this occupation in educational services is 67,000 dollars and cannot be treated as a teacher's wage; only comparisons within the same occupation across industries are meaningful. |
| Other professional, scientific and technical services | 65,880 dollars / year | The annual wage median for this occupation in other professional, scientific and technical services is 65,880 dollars, not a Chinese equivalent post or a current wage at any employer. |
| Self-employed survey researchers (coverage note) | Statistics exclude the self-employed | OEWS wages exclude the self-employed. This row explains the employment status and the coverage gap and is not an industry wage ranking; the employee median must not be substituted for self-employment business income. |
First confirm whether the target post is an employee, self-employed or another contractual form, then check the pay, hours and continuity arrangements. BLS describes most people in this occupation as full-time on regular hours, and this does not support writing everyone as short projects or long gaps.
Survey design, collection coordination, analysis and project affairs may carry different duties. Confirm the post object and the pay basis before comparing conditions; the same job title or the fact that both tabulate data does not automatically mean the same statistical scope.
Industry median differences cannot decide the choice on their own; location, travel, working hours and the target tasks also matter. Published U.S. industry data do not represent a quote in any Chinese city or a contract an individual can obtain.
The 2025 employment size, which does not cover exactly the same population as the wage statistics.
The total change over the projection period, not an individual's probability of getting a job.
Mainly replacement demand from people changing careers or leaving the labour market, not annual net growth.
Can the IC interest code show that I am suited to survey research?
Investigative clues can indicate whether you are willing to question the problem and the method, and conventional clues can open a discussion about whether you like coding, organising and checking. The O*NET IC describes occupational activities, not two letters an individual must match; testing the interest with the actual exercise on this page is more concrete.
Do not use it like this:Liking research may still coexist with disliking repeated checking or fieldwork coordination, and these should be observed separately. Do not let an interest code replace real ability, training and living-condition judgments.
Explore my career interestsWhich work habits are worth observing together with the test results?
The results can be turned into questions about preparation, checking and collaboration: do you organise material by the rules, do you keep checking when results are unfavourable, can you make a specific revision on receiving method feedback. This page has no evidence for setting any dimension high or low as a pass line for survey research, and behaviour records are more checkable than labels.
Do not use it like this:The Big Five results are only behaviour discussion clues. Observe rule organising, feedback revision and honest recording, and do not turn a high or low dimension into a hiring threshold or a statistics competence certification.
Explore my Big Five traitsCan introversion or extraversion decide suitability for survey research?
No. Independent analysis, interview communication and team coordination are different tasks; social preference cannot directly decide questionnaire and statistics ability. You can discuss separately how you prepare, take feedback and restore energy, then confirm with specific tasks, and do not rank posts by type.
Do not use it like this:Do not explain every project or application outcome with four letters, and do not rule out a research direction you have not actually tried.
Explore my personality preferencesHow do you observe your own motivation to do survey research?
You can distinguish interest in real problems and methods, expectations of influencing conclusions and the need for recognition. Several motivations coexist; the key is that when results are unfavourable you still protect the authenticity of the material and the method record, and do not change answers or hide limitations for an attractive conclusion.
Do not use it like this:Do not judge research talent from a motivation label or infer other people's psychology. Record observable tasks and feedback first, then discuss possible motivations.
Explore my motivation cluesHow are numerical, reading and reasoning abilities verified?
This page first checks the six records, the completion status, the valid answers and the denominator, and explains how the calculation produces the conclusion. A generic test is not a sampling or statistics competence certification; errors should be located to the rule, the comprehension or the calculation, and then targeted training is arranged, with no fictional hiring score set.
Do not use it like this:One error cannot summarise overall ability. Recheck the rule and the material comprehension, then retry; mastering this exercise is not the same as mastering all survey methods.
Explore my cognitive performanceHow do you listen to answers and maintain the research boundary at the same time?
You can observe whether you listen patiently to the explanation, restate the feedback accurately and ask answerable questions about unclear requirements. Real interviews must respect participants' expression and refusal under the procedure; this page uses fictional records only, does not test the ability to persuade real people and does not require exposing private experience.
Do not use it like this:The research exercise needs no real respondents or personal materials. Wanting to finish a piece of work does not mean you may collect, publish or upload other people's information; the fictional material in the task text is enough.
Explore my communication cluesBetter match:Willing to clarify the objective, test the wording and record the basis for revision.
Note:Good written expression is not valid measurement, and sampling and statistics training still have to be built.
Better match:Willing to check status, coding and anomalies and to communicate with interviewers and analysts.
Note:Being able to interview is not the same as being able to interpret the population independently, which needs verification against the real division of work.
Better match:Interested in research execution, procedure records and participant management.
Note:Clinical project training, compliance procedures and authorization are checked separately, and ordinary survey experience does not apply directly.
View related occupations:Clinical research coordinatorsBetter match:More willing to study organisational problems and propose executable process suggestions.
Note:Organisational diagnosis and improvement communication must be built and verified with consulting outputs.
View related occupations:Management consulting analystsBetter match:Willing to apply research to demand, competition and marketing decisions.
Note:The market context and business interpretation must be built, and not every questionnaire is the same as market research.
View related occupations:Market research analysts specialising in marketingBetter match:Willing to learn financial mathematics, risk and valuation models.
Note:This belongs to different objects and outputs and requires independently building finance and modelling ability.
View related occupations:Financial analystsThe six fictional records on this page first require distinguishing completion status from valid answers to specific items. After remembering the numbers you still have to explain why the denominator is six, four or three; when an error occurs, locate the record and the rule, and do not treat one miscalculation as a conclusion about overall ability.
On receiving feedback, restate the problem first, then revise the original draft separately per the A and B requirements. A tests the denominator and the missing data, B tests the interpretation scope and the simulated material; each version needs at least two changes with the basis stated, and you cannot just make the sentences smoother.
Facing project or application feedback, separate what you can revise in your preparation from the screening conditions you do not yet meet. Without feedback, keep the reason unknown and do not invent it; choose one next step that can give specific method feedback, then assess the time and investment.
Reading interest is only a starting point. First experience item-meaning checking, status coding, denominator checking, the two revisions and the method record, and note which stages you are willing to repeat and which keep draining you. Appreciating a report cannot directly prove that you like collection and quality management; one exercise cannot promise career success either.
First distinguish independent analysis, interviewing and team coordination. Quiet people still have to state problems and take feedback; sociable people equally have to check rules and material. Turn the preference into specific behaviour, then judge by the target post's proportions and arrangements, and do not draw a hiring conclusion straight from extraversion.
No. Running the software still requires a clear question, sample, variables, missing-data handling and inference conditions. Use the exercise on this page to explain each denominator, then progressively train questionnaire and sampling skills; the work should show methods and review, not only attractive charts. Real post requirements are still checked against the announcement.
List the locations, times and travel range clearly, then look for matching tasks. BLS describes office analysis and possible travel side by side, and it cannot be assumed that everyone does continuous fieldwork. If information about the target post is insufficient, check the specific collection and analysis arrangements; living constraints are not personality flaws.
Yes. The six records and eight claims in F1, the two feedback sets in F2 and the record in F3 can all be completed on paper without real participants. Make clear that the task text is original fiction, which judgments and revisions are yours, and whether AI was actually used, stating truthfully if it was not; this cannot impersonate a formal survey or an employer comment.
Treat the result as a clue to be verified, and return to the denominator check, the feedback revisions and the real conditions. Turn one description into a behaviour question and use task records to support or refute it; keep the uncertainty, and you do not need to test repeatedly until you get a desired label, nor ignore real interests or constraints.
Organise the work you have completed, the rules you are still unsure about and your willingness to keep training, and choose only one next step that can give method feedback, such as questionnaire measurement or sampling basics. One small exercise only supports continued exploration; also record the time, cost and help needed, and do not promise professional qualification or employment from it.
Possible consequences：This may expose personal material or exceed the research agreement; publishing results is not the same as publishing raw identifying data.
Reducing risk：Check the data source, the processing purpose, the identifying information and the institutional authority. Participating in a survey or being able to read a file does not automatically mean it may be published, uploaded to external tools or used for another purpose; confirm under the project rules and applicable law.
Basis：AAPOR professional ethics and disclosure standards
Typical scenarios：Changing the location, contact method or adding unconfirmed tasks at short notice, so the person executing is unclear about the procedure and the support.
Main posts：Staff undertaking fieldwork contact, training or collection coordination
Possible consequences：This affects execution, participant communication or one's own safety judgment, and confirmation must not be skipped to save time.
Reducing risk：Interview execution follows the confirmed procedure and the institutional safety arrangements. For unclear sites, contacts or extra tasks, explain promptly and refer them to the responsible person for checking; this page does not organise real interviews or provide safety certification.
Basis：BLS OOH survey researchers
Typical scenarios：Treating an adjacent post's basic salary or U.S. statistics as a guaranteed take-home amount for all survey researchers.
Main posts：People comparing internships, dispatch, employee or self-employed opportunities
Possible consequences：Expectations then conflict with the contract, performance pay and actual arrangements, affecting living decisions.
Reducing risk：Distinguish wage statistics, historical post quotes and personal take-home pay. Check the contracting entity, the pay composition and the continuity arrangements, and do not promise income from an industry median or an adjacent research assistant amount.
Basis：Ocean University Global New Quality Productive Forces Research Centre research assistant recruitment · Renmin University CGSS 2026 project assistant internship announcement
Typical scenarios：Treating a typical master's degree, an internship certificate or a professional association membership as a uniform global research licence.
Main posts：People choosing courses, internships or specific survey projects
Possible consequences：Investing in irrelevant training, or missing the real employer, project and regional requirements.
Reducing risk：Distinguish typical education, training, employer conditions and statutory licensing. The BLS description of a U.S. occupation requiring no licence or certification cannot be extended to mean that all survey activity in China needs no formalities.
Basis：BLS OOH survey researchers · AAPOR professional ethics and disclosure standards
Typical scenarios：Uploading identifying material to write code or a report, changing the purpose without confirmation or accepting tool retention.
Main posts：People handling unpublished questionnaires, data and analysis results
Possible consequences：The material may leak, and the model may also misclassify, invent citations or over-interpret.
Reducing risk：Real data and unpublished questionnaires are used according to law, institutional authority and the project agreement; a tool may make errors or retain inputs, so identifying material is not uploaded without confirmation.
Basis：NIST AI 600-1 generative AI risk profile · AAPOR professional ethics and disclosure standards
Typical scenarios：To prove the expected result, recording missing as no, ignoring the convenience-sample limitation or inventing answers.
Main posts：Design, quality management, analysis and external reporting staff
Possible consequences：The method and the conclusions are distorted, affecting how clients and the public understand.
Reducing risk：Recheck the items, the sample and the original records; missing does not mean no, simulation is not passed off as actual, and the report states the limitations. When the procedure or the basis needs to change, refer it to the research lead for confirmation.
Basis：AAPOR Standard Definitions · O*NET survey researchers 19-3022.00
| Entry direction | What you mainly do at first | Employer requirements and basis for observation | How to prove it with no experience | Difference from other directions |
|---|---|---|---|---|
| Questionnaire and method practice | Clarify the question, the options and the sample scope, and organise the testing and revisions. | O*NET lists tasks such as testing wording; the exercise below is an editorial design. | Show the rules, judgments and revisions for the six fictional records, without falsely claiming a real survey. | An ability entry point, not a recruitment grade or a hiring promise. |
| Data and quality assistant | Organise the coding, status and calculations, and record anomalies. | Post requirements follow the announcement; the method and the denominator should be traceable. | Present the explanation and review of 4/6, 2/3 and 2/4. | Being able to tabulate is not the same as being able to interpret the population independently. |
| Historical Renmin University CGSS internship case | Questionnaire standardisation, system testing, training and liaison affairs. | 2026-05-27 announcement; attendance from 06-01 to 09-30 with at least three working days per week. | Explain your experience and time according to the actual application requirements; the small exercise is only your own practice. | The internship allowance amount is not given and the attendance period has ended, so it is not confirmed to be still recruiting. |
| Historical Ocean University adjacent research assistant case | Research data analysis and visualisation, reports and project affairs. | 2025-06-06 announcement, master's degree or above, dispatch, deadline 06-20. | Real research and analysis work cannot replace that round's education and other conditions. | An adjacent post is not an exact mapping of this occupation; basic salary and performance pay are kept apart. |
| Method and ethics learning entry point | Learn questionnaire, sampling, missing-data, disclosure and material protection. | O*NET, BLS and AAPOR are reference materials, not a third recruitment. | Organise the method notes and the review process, and mark the editorial exercise clearly. | Not a third institutional sample, a qualification certificate or a uniform course. |
The Renmin University announcement involves questionnaire standardisation, system testing, training materials and liaison; the Ocean University announcement involves data analysis, reports and project affairs. Each supports only its own announcement and does not constitute national recruitment statistics.
Make an A and a B version from the same F2 original, each with at least two changes, stating the supporting record and rule.
F3 makes clear what is original fiction on this page, your own judgments, the given feedback and the actual AI use, without inventing real participants or employer evaluations.
State the time, place, skills and methods needing help separately; a historical announcement does not represent current requirements.
Use F1 and write supported or unsupported for each item with the rule stated; list complete 4/eligible 6, valid willing 2/valid 3, and willing 2/complete 4.
Per F2, A and B each revise two points from the same original, keeping the before and after and the reasons; the two versions may focus on different errors.
Write the fictional source in the task text, what you completed, the given feedback, the original and the two versions, and the actual AI use, with no real-person material.
Check the completion of the four work products, one error, one interest and the next method course question; do not declare professional qualification.
| Situation | What to check | How to observe | What cannot be inferred from it |
|---|---|---|---|
| Before applying | Check the announcement duties, education, location, employment form, current status and materials. | Experience is genuine and matches the requirements one by one, with unknowns left pending. | A historical announcement and a complete preparation do not guarantee hiring. |
| Receiving quality feedback | Restate the item number, the record and the calculation rule, then revise. | The before and after are comparable, and the explanation shows how the change affects the conclusion. | The editorial exercise is not a uniform examination for all employers. |
| Confirming cooperation | Check the contracting entity, the task, the pay, the attendance and the material permission. | Distinguish internship, dispatch and other employment forms and keep the agreement. | Applying, interviewing and the final contract are not the same commitment. |
| Work review | Explain the source, sample, coding, denominator, inference and limitations. | The first version, the A and B versions and F3 are traceable. | The fictional exercise does not prove real survey experience or a licence. |
| Qualification / preparation item | Who it mainly suits | Practical value in job hunting | Whether it is a universal entry threshold | Application / processing conditions | Investment considerations | Important limitations |
|---|---|---|---|---|---|---|
| U.S. typical education | People exploring this occupation in the United States | BLS: usually at least a master's degree, with some entry posts possibly accepting a bachelor's. | Typical occupational background, not a legal requirement for all posts. | Check the conditions against the real school and post. | First check the measurement, statistics, computing and social science foundations and the practice feedback. | The U.S. scope cannot be applied automatically to China; a degree does not guarantee hiring. |
| U.S. licensing and certification | People checking the occupational preparation for this occupation | BLS states that this occupation requires no licence or certification. | A U.S. occupational description; it does not imply that projects worldwide need no formalities. | The specific institutional and regional requirements are still checked separately. | Do not buy unknown certificates because of the occupational name. | It is not a Chinese legal conclusion and does not replace data permission or project review. |
| Historical internship in China | People considering tasks similar to the Renmin University CGSS work | Learn about questionnaire testing, coordination, training and internship preparation. | The conditions of that historical announcement only. | The original announcement required at least three working days per week and similar; the attendance period has ended. | Assess the time, place and allowance again against a new announcement. | An internship certificate is not a statutory survey licence or a national qualification. |
| Historical Ocean University research assistant | People meeting the adjacent post conditions | State the master's degree, dispatch and specific research task requirements. | That round's conditions, not a national threshold for survey researchers. | Read the original announcement item by item; 2025-06-20 has passed. | Build research and analysis outputs for the current target task. | An adjacent occupational sample cannot establish an education or pay grade for this occupation. |
| AAPOR professional ethics and methods | People learning survey standards | Provides references for participant protection, disclosure and status definitions. | A professional standard, not a uniform Chinese licence. | Membership or a compliance statement is not a competence certification. | Read the applicable clauses and methods first, then check the project arrangements. | The real law, institutional requirements and professional standards are understood separately. |
| Your own method work | Explorers lacking real experience | Show the question, the calculation, the revisions and the limitation notes. | Editorial preparation suggestions, not a uniform recruitment certificate. | Use only the fictional records in the task text and write your own contribution truthfully. | Pen and paper are enough; get specific method feedback first. | The work does not guarantee hiring, a degree or real project authority. |
All editorially fictional, not a real survey. The objective is only to check the rules; all six records are known to be eligible and come from convenience practice material, with no probability sampling, weighting or population representativeness. Question: if a book exchange event were held next month, would you be willing to take part? Options: willing, not willing, not sure yet, prefer not to answer.
Records: 01 complete / willing; 02 complete / not willing; 03 complete / this item missing; 04 complete / willing; 05 partially complete / this item unanswered; 06 refused / this item unanswered. Rule: completeness follows the status in the task text, and a missing item does not change the status. The valid answers to this item are only willing or not willing in the complete records; missing, partial and refused are not recorded as no.
First judge the eight claims item by item: (a) eligible records 6; (b) complete 4; (c) valid answers to this item among the complete ones 3; (d) willing 2; (e) the willing proportion of valid answers 2/3=66.7%; (f) missing may be recorded as not willing; (g) the six records prove that most of the community or the country is willing; (h) two AI-generated records may be recorded as actual completions. Write supported or unsupported with the basis for each item.
Draft to be revised: 'the willing proportion of valid answers is 50%; missing is treated as not willing; this proves that most community residents are willing; AI is used to fill in two completed questionnaires.' A feedback: check the valid denominator and the missing-data handling, one change each. B feedback: independently check the interpretation scope and the simulated material from the same draft, one change each. At least two changes per version, keeping the before and after and the basis; do not use version A as the draft for version B. F3 records the source, your own outputs, the given feedback, the original, the A and B changes and reasons, and the actual AI use (write none if unused), with no real-person information collected.
(a) to (e) supported; (f) to (h) not supported. The completeness proportion is defined as 4/6=66.7%, which is only the completion proportion customised for this exercise, not a general official AAPOR response rate. Among the complete records, valid answers are 3 (01, 02, 04) and willing is 2 (01, 04), so the valid willing proportion is 2/3=66.7%; 2/4=50% is another quantity using all complete records as the denominator and cannot be labelled the valid answer proportion. Missing for 03 is not the same as not willing, and the partial 05 and refused 06 cannot be added to the valid denominator for this item.
Version A, two points: 50% changed to 66.7%, noting the 01, 02, 04 denominator; missing changed to keeping the missing data rather than coding it as no. Version B, two points: community majority changed to describing only the three valid fictional records without extrapolating to the population; AI filling in completions changed to not treating generated material as actual participants. The two versions focus on different problems, the unrevised items still need integrating, and the final draft must correct all four points together. F3 reference: original fiction on this page, your own paper judgments and the two revisions, feedback from the task text, no real people or third-party material, and write none if AI was not actually used.
Pen and paper are enough; do not contact real people, collect private material or add simulated participants. Four work products: (a) the eight first judgments and three denominator calculations; (b) the A and B versions with two changes each and the final integrated draft; (c) the F3 source and actual tool record; (d) the seven-day review and the next method learning question. The reference answer is for checking; answer independently first, and do not impersonate a formal survey, an employer evaluation or a professional qualification.
Read the six records in F1 and list the complete, partial, refused and item-validity rules; produce the input and coding cards.
Without reading the answer first, judge the eight claims item by item and list the numerator, the denominator and the reason; produce the first version.
Compare with the five supported and three unsupported and check the meaning of 4/6, 2/3 and 2/4; produce the error location without changing the status in the task text.
Revise the valid proportion and the missing-data handling from the original, with the before and after and the basis for two points; produce version A.
Still from the same original, revise the population inference and the AI-filled records, with the before and after and the basis for two points; produce version B and the final draft with all four corrections.
Fill in the original fictional source, your own outputs, the given feedback, the before and after versions and the actual tool use; write none if unused.
Check the four work products and record one rule error, one interest and the next questionnaire or sampling question; do not judge talent or promise hiring.
Typical responsibilities：Clarify the research objective, item meaning, options and time frame, and do a small paper check first.
Promotion evidence：Can point out an item-meaning or denominator problem and explain the basis for the revision.
Qualification boundary：Check the statistics, computing and social science background against the course and the post; do not assume uniform tuition.
Next step：Learn questionnaire pretesting and measurement basics and seek specific feedback.
Typical responsibilities：Support questionnaire standardisation, system testing and documentation in a clearly supervised project.
Promotion evidence：Show the version and test-issue records and distinguish your own contribution.
Qualification boundary：Internship conditions follow that announcement; the Renmin University case is only a historical task.
Next step：Build sample frame and sampling basics and clarify how testing affects the variables.
Typical responsibilities：Study the sample, weighting and inference conditions within the actual division of work and record the decisions.
Promotion evidence：Explain the method, limitations and sensitivity, not just show the output.
Qualification boundary：Training and confirmation by the project lead are required; software ability is not a research licence.
Next step：Gradually take part in complex design and obtain method review feedback.
Typical responsibilities：Propose complete method suggestions combining the objective, resources and material quality.
Promotion evidence：Has method records and specific feedback from several real projects.
Qualification boundary：The post and publication authority are determined by the institution and are not obtained automatically by years of service.
Next step：Deepen the professional direction, or build the abilities an independent adjacent occupation requires.
Typical responsibilities：Check complete, partial, refused and variable rules from fictional material.
Promotion evidence：Can explain each status and denominator and keep the missing data.
Qualification boundary：A paper exercise is not a real interviewing qualification or experience.
Next step：Learn collection procedures and basic data quality.
Typical responsibilities：Support training, contact arrangements and interview status organising under guidance.
Promotion evidence：Record anomalies, procedure versions and confirmation by the responsible person.
Qualification boundary：Real sites, material permission and safety procedures are checked separately.
Next step：Build knowledge of non-response, interview quality and collection modes.
Typical responsibilities：Plan or coordinate surveys, review collection and track response and anomalies.
Promotion evidence：Show problem location, the basis for handling and the effect on the version.
Typical responsibilities：Complete the eight judgments and three denominators for F1 and constrain the population interpretation.
Promotion evidence：The four original fictional work products are traceable and do not impersonate a real survey.
Qualification boundary：Check the statistical foundation first; a generic test cannot replace training.
Next step：Learn missing-data handling, coding and descriptive statistics.
Typical responsibilities：Organise data, visualisations and research material according to a clear method.
Promotion evidence：Present the process, the rules and your own revisions, and show the material within its permission.
Qualification boundary：The Ocean University adjacent research assistant case has specific education and dispatch conditions.
Next step：Build questionnaire, sampling and analysis methods and check the target post.
Typical responsibilities：Analyse the material with the research objective, produce summaries and explain the method limitations.
Define the objective, sample and variables, and be able to explain the input and the coding.
Check status, denominator and errors under the rules, keeping the original records.
The calculation is reviewable and explains the limited sample and the applicable conditions.
Learn measurement, sampling and statistics, combining real feedback and authority to expand.
Cannot show：The precise projection does not mean that every region or sub-task is the same, and it cannot be split into an independent generative AI loss or a personal probability.
Professional standard reference · AAPOR survey professional standards
Sources, methods and generated content must be distinguished
Research ethics and disclosure
The AAPOR 2026 professional ethics update requires disclosing the material and methods and distinguishing people's answers from AI-generated content; this is a professional standard response and does not prove industry-wide adoption rates or income growth.
Cannot show：Professional standards and material disclosure requirements are not the same as a statutory competence certification and do not prove an industry adoption rate or a personal income change.
Historical announcements from two Chinese institutions · A specific project or employer, not the whole occupation
Sample of internship and adjacent assistant tasks
Entry routes and evidence gaps
The historical announcements from the two institutions show internship and adjacent research assistant tasks, with the status explained. A small number of announcements cannot constitute the current post count, a change in demand or a national wage sample.
Cannot show：The two historical institutional announcements have no representative population and cannot compute current recruitment heat or a national demand trend.
These are ability adjacency relationships and do not mean automatic promotion or that no further learning is needed.
Transferable：Understand the research procedure, the material records and participant communication.
Needs to be filled in：Build clinical protocol, compliance and target institution training and authorization, and verify clinical project outputs.
View occupation detailsTransferable：The research question, material interpretation and client communication.
Needs to be filled in：Build organisational process diagnosis, efficiency analysis and implementation of improvement suggestions.
View occupation detailsTransferable：Questionnaires, material analysis and report interpretation.
Needs to be filled in：Build consumer, competition, pricing and marketing uses and verify with market problems.
View occupation detailsTransferable：Numerical checking, interpretation and method records.
Needs to be filled in：Build financial mathematics, investment valuation and risk modelling; the denominator exercise here is not quantitative ability.
View occupation detailsTransferable：The foundations of data quality, sampling and statistical problems.
Needs to be filled in：Deepen probability, statistical methods and inference and prove it with real method outputs.
View occupation detailsTransferable：Questionnaires, material organising and research collaboration.
Needs to be filled in：Build literature, experiment or database management for the assistant duties and clarify the support scope.
View occupation detailsTransferable：Social problems, survey material and interpretation.
Needs to be filled in：Build sociological theory, research design and argument about social mechanisms.
View occupation detailsTransferable：Data cleaning, analysis and visualisation.
Needs to be filled in：Build programming, machine learning and large-scale structured or unstructured data methods.
View occupation detailsUse the six fictional records in F1 to judge the eight claims independently, revise two points each for the two F2 feedback sets and fill in F3; the four work products can be done on paper, without collecting real-person information or impersonating project experience.
BLS: usually at least a master's degree, with some U.S. entry posts accepting a bachelor's, and it describes no licence or certification requirement; this is the U.S. scope for this occupation. In China, check the target post and project requirements, and the two historical announcements are not a uniform national threshold.
Insufficient evidence: this page has no representative adoption rate or generative AI independent substitution rate for this occupation. The 4.8% U.S. decline is a multi-factor total projection; drafts, coding and analysis must be verified by people, and actual answers cannot be invented. BLS2025 through 2035 United States SOC19-3022 overall projection for survey researchers in this occupation: 9,800 people in the 2025 base period, 9,300 at the end of 2035, an official net decrease of 500, a change of −4.8%; average annual openings of 600 include transfers, exits and replacements and are not net additions or current hiring. The headcounts are converted from thousands and keep the original table's precision, and the growth rate cannot be recomputed from the two rounded endpoints; the projection covers the self-employed and others and so differs from the OEWS wage-earner population, and it cannot be extended to a generative AI independent substitution rate, Chinese projections or an individual's probability of unemployment.
Insufficient evidence: no nationally comparable wage distribution for survey researchers in China was obtained this round. No. The 2025 wage statistics for U.S. employees in this occupation and the Chinese internship and adjacent research assistant announcements cover different populations and pay bases; the evidence for national Chinese wages for this occupation is insufficient, and no currency conversion is used to fill the gap. 2025 pre-tax annual average wage for the broad professional and technical personnel post in China's above-scale enterprises: 155,491 yuan per year. The official definition covers personnel engaged specifically in scientific research and professional technical work; the author references it as a condition only when the person actually carries research professional duties such as survey design, measurement or research analysis and the unit is included in the statistics, and interview execution, liaison or administrative affairs are not automatically included because of a similar name. The coverage is enterprise legal persons in 16 industry categories included by size, qualification or quota conditions, not all universities, research institutes or all urban units; it includes individual income tax, social insurance and housing fund deductions, and it is not the median, starting pay or a personal range for survey researchers and does not replace the Ocean University dispatch basic salary or the Renmin University internship allowance.
Check the questionnaire item meaning, the sample source, the missing data and the denominator first, then confirm the material use and publication authority. A high completion rate, attractive charts and fluent AI text alone cannot prove the research is reliable.
Observe whether you are willing to trace inconsistencies, record the rules, accept feedback and admit unknowns. IC is an occupational interest clue, and personality or cognitive scores are not a hiring line.
Keep the first version of the eight judgments and calculations, the A and B versions and the final draft, F3 and the seven-day review. Mark what is original fiction, your own contribution and the given feedback, and write none if AI was not used.
Confirmed by the research, collection, data or publication leads in the project's actual division of work. Readers of the exercise only make judgments about the task text and do not change the real questionnaire, collect real-person material or publish research conclusions externally.
FermatMind career content editing
2026-10-12 author check; the statistical period, the historical announcements and the professional standard versions are kept separately. The complete AAPOR method handbook and the two 2025 supplementary reports were not fully read and no full formulas are listed; this round includes no independent review, HTML or publication.
2026-10-12 author check of the nine modules: O*NET tasks for this occupation and the eight adjacent occupations, the IC and Job Zone5, the BLS wages by industry and the 2025–2035 projection for this occupation, the historical Renmin University CGSS and Ocean University announcements, the AAPOR2026 ethics web page and the Standard Definitions page and the 2024 generative AI discussion, and the AI600-1 generic risk clauses applicable to NIST. The NBS broad professional and technical post and the ILO summary reuse the applicable original text from this batch; the editorial exercise is not an official course, a national sample or a legal conclusion.
Usage limits：Not the same as the duties of this occupation, an automatic transfer or a Chinese qualification.
Usage limits：Not the same as the duties of this occupation, an automatic transfer or a Chinese qualification; the finance link is limited to quantitative analysis.
Usage limits：It does not represent a Chinese or personal probability and cannot be split into an independent AI effect.
Usage limits：A professional standard cannot replace the specific jurisdiction, institutional authority or hiring qualification, and does not prove an adoption share.
Usage limits：Not the same as a Chinese qualification, a personal quote or a generative AI independent loss rate.
Usage limits：The internship attendance period has ended and it is not confirmed to still be recruiting; it is not a national wage or a uniform threshold.
Usage limits：Not the same as the duties of this occupation, an automatic transfer or a Chinese qualification.
Usage limits：It is not a representative adoption or time-saving experiment, and the AI evidence table on this page does not compute a share from that paper.
Usage limits：It is not a law, certification, adoption rate or accident rate for this occupation.
Usage limits：It provides no Chinese wages, a person's actual paid hours or take-home amount; an annual wage percentile is not treated as a job grade.
Usage limits：Occupational interest is not a personal threshold and does not establish Chinese admission or an AI effect.
Usage limits：Not actual adoption, net loss or a personal probability of unemployment for this occupation.
Usage limits：An adjacent research assistant post, not an exact mapping of this occupation; it has closed and does not represent the national picture or a current position.
Usage limits：The internship attendance period has ended and it is not confirmed to still be recruiting; it is not a national wage or a uniform threshold.
Usage limits：An adjacent research assistant post, not an exact mapping of this occupation; it has closed and does not represent the national picture or a current position.
Usage limits：A professional standard cannot replace the specific jurisdiction, institutional authority or hiring qualification, and does not prove an adoption share.
Usage limits：A professional standard cannot replace the specific jurisdiction, institutional authority or hiring qualification, and does not prove an adoption share.
Usage limits：Not all supplementary PDFs were read and no full formulas are listed from them; this is not an occupational licence.
Usage limits：Not the same as the duties of this occupation, an automatic transfer or a Chinese qualification.
Usage limits：Occupational interest is not a personal threshold and does not establish Chinese admission or an AI effect.
Usage limits：Not the same as the duties of this occupation, an automatic transfer or a Chinese qualification.
Usage limits：Not the same as the duties of this occupation, an automatic transfer or a Chinese qualification.
Usage limits：Not the same as the duties of this occupation, an automatic transfer or a Chinese qualification.
Usage limits：Not the same as the duties of this occupation, an automatic transfer or a Chinese qualification.
Usage limits：Not the same as a Chinese qualification, a personal quote or a generative AI independent loss rate.
Usage limits：Employee wages exclude the self-employed and are not take-home pay, starting pay or a job grade; the annual figure is not worked back from a rounded hourly rate.
Usage limits：It is not a law, certification, adoption rate or accident rate for this occupation.
Usage limits：A professional standard cannot replace the specific jurisdiction, institutional authority or hiring qualification, and does not prove an adoption share.
Usage limits：Not a real survey, a representative sample, a qualification certification, an AI score or an employment guarantee.
Usage limits：A broad multi-occupation post, not the precise median, starting pay or range for survey researchers, and not used to fill in nearby wages, internship allowances or the national occupational distribution; it is not an official Chinese-U.S. occupational mapping.
For career exploration; the specific post, contract, material permission and applicable law are checked separately. The fictional exercise does not constitute real research or a qualification certification, and the historical samples do not represent the whole industry. The Chinese and U.S. occupations, statistics, employment and qualification scopes are checked separately. The directly published annual wage is not take-home pay; a percentile is not a job grade and excludes the self-employed. A historical announcement is not a current post; a fictional exercise is not real research; AI exposure is not a substitution rate.
MBTI (English) · Big Five (English) · MBTI (Chinese) · Big Five (Chinese)
Locate the item number, record or calculation rule first, then explain how the revision affects the conclusion. Keep the original and the reasons, and do not merely polish the wording; changing the real survey procedure requires confirmation by the project lead.
Check the data source, identifying information, processing purpose and tool permission separately; a new purpose cannot be assumed to be allowed merely because the original file is accessible. AAPOR provides a professional ethics reference, and the actual legal and institutional requirements must be checked per project.
At delivery, record the questionnaire development, collection, sampling and statistical weighting decisions, and explain the limitations and unanswered questions. A public report does not mean that raw identifying material may be published; data use must be checked against the law, institutional arrangements and the research agreement.
Qualification boundary：Specific procedure changes are confirmed through the responsibility chain, and items cannot be changed at will.
Next step：Connect collection quality with the later analysis and method notes.
Typical responsibilities：Compare the execution conditions and quality limits of different projects and form improvement suggestions.
Promotion evidence：Explain the contribution with checkable monitoring records and real feedback.
Qualification boundary：Management authority and responsibility depend on the post, and years of service do not guarantee promotion.
Next step：Deepen the implementation method, or move with verification to adjacent project management.
Promotion evidence：The analysis is reproducible and separates facts, inferences and unknowns.
Qualification boundary：The method and external publication are confirmed by the project's actual division of work.
Next step：Learn weighting, uncertainty and more complex research designs.
Typical responsibilities：Propose evidence-constrained interpretations and improvement directions for complex material.
Promotion evidence：Method notes and real feedback from several projects, rather than the number of charts.
Qualification boundary：Complex tasks require training and verified responsibility and are not granted automatically by AI output.
Next step：Continue method development in this occupation, or build statistics, sociology and other target occupational abilities.
Usage limits：Editorial ratings use the existing occupational responsibilities and AI task analysis on this page. They are not adoption rates, measured effects, job-disappearance or personal-unemployment probabilities, or official scores from Stanford or another external institution. No measured time weights or future robot deployment are assumed.
