Research · Protocol JRI-NOSACQ-v7 · Participant information and consent notice v8, November 4, 2026 (Eastern time)
Safety climate by industry
Optional company code and participant reference
A management account can create one shared company code per quarter. No worker account is required. Enter or follow that shared code, check the company name, and separately agree to include your response in that company’s protected grouped climate report. You can remove the code and participate independently without losing access or benefits. Do not let anyone pressure you or watch your answers. The company name is supplied by its management account; shared codes do not verify your employment or prevent duplicate participation.
Only new responses under this notice contribute to the participant reference, whether independently submitted or separately company-linked. A company-linked response adds an opaque company survey reference to the private answer object. The website also stores seven person-level dimension means, questionnaire language, quarter, random response reference, optional company survey reference and expiry in a private database to prepare grouped comparisons. This is separate from ordinary company working records and the older optional company participation counter. No personal account identifier, name, email, demographic field, exact location or individual completion status is added to this derived database. Authorized research and hosting personnel may have technical access. Earlier answers are not backfilled, linked or added to this new comparison pool.
A company receives only a reviewed whole-company quarterly report: seven dimension means, usable response counts and descriptive differences against other newly consented participants in the same quarter and questionnaire language. English and Spanish remain separate. All industries are combined in this first comparison; industry-matched comparisons are not offered. Responses linked to that company are excluded from its reference group. Independently submitted responses may include people from the same company; unique workers and distinct workplaces cannot be verified. The participant reference is a self-selected sample, not a national baseline or proof of representative coverage, causation, compliance, statistical significance or shared workgroup climate.
No live count, raw answer, individual score or completion list is provided. A report requires quarter close plus seven days, at least 30 usable responses in every dimension in each displayed company and reference language group, and research-owner consent, quality, comparability and privacy review. Thirty is a project disclosure floor, not a power calculation, anonymity guarantee or NFA requirement. A whole language comparison is withheld if any dimension is too small; no team, state, citizenship or demographic filters are offered to employers. Rare groups, complementary totals, overlapping releases and information known by coworkers can still create identification risks. The research owner may withhold a numerically eligible report.
The shared code appears in the survey link when used and may appear in browser history and hosting logs. It is a shared group identifier, never a personal invitation tracker. Your employer can know that you were given a code; this does not tell them whether you submitted. Never disclose your private withdrawal receipt to your employer. Before release, withdrawal removes the retained response and derived scores from future preparation and invalidates a pending snapshot. After release, individual records can still be removed, but a frozen grouped report cannot be recalled one answer at a time. Individual derived records follow the response’s 730-day expiry; company report snapshots expire 730 days after quarter close. Provider logs/backups have separate provider-controlled retention. No automatic numeric release is authorized.
Your choice comes first
You may participate if you are 18 or older and currently work at a U.S. workplace. Participation is voluntary, free and separate from accounts, purchases, employer training and ordinary services. An invitation is not consent. You may decline, skip any questionnaire item or leave before submitting. There is no compensation and no promised personal benefit. The participation policy does not permit employers to make employment benefits or ordinary services conditional on taking part; the research team cannot guarantee how an employer will behave. Answer privately if you choose to participate.
Purpose and procedure
The purpose is to explore how participating workers and leaders experience safety and to identify questions for learning and improvement across industries. Think about one current workplace or work unit throughout. You will see the 50 items from your selected official version: NOSACQ-50 English 2019 or Spanish–Colombia 2020 in their original order, with four response options. Some items are phrased negatively. You can skip an item. If you wish to refrain from the entire questionnaire, leave without submitting. Completion time has not yet been pilot-tested. Please submit once for the same workplace per calendar quarter, even if you receive several invitations. Each quarterly collection closes at midnight Eastern when January 1, April 1, July 1 or October 1 begins. A new collection opens immediately; new participation requires a fresh form and consent. Closed-quarter responses are retained under their original notice and cannot be silently moved into the next quarter. Without identity or workplace linkage, we cannot reliably detect repeat participants; repeat participation is a reported limitation.
The 50-item instrument is preserved. Additional context questions are separate. This online collection omits the paper form’s optional birth year, gender and comment fields. The English 2019 and Spanish–Colombia 2020 questionnaires use NFA’s existing official item wording and four-point response labels. “Colombia” identifies the regional instrument edition, not a participant’s origin or citizenship. You choose the language before beginning; to change it, start a fresh form and consent. Do not submit again just to answer in another language. The research team’s Spanish instructions, added context and consent are AI-assisted translations and have not received independent bilingual or target-population comprehension review; they are not NFA-authored or independently validated. The site navigation and detailed official NAICS directory remain in English; broad sector labels and survey controls have Spanish explanations. Leer este aviso en español.
What is collected
The retained response contains the 50 selected answers or blanks; optional NAICS sector and detailed code with the 2022 edition; optional worker/managerial role, broad workplace-size and tenure bands, employment arrangement and shift pattern; optional self-reported U.S. citizenship category; selected questionnaire and consent language; consent and instrument versions; UTC submission date and Eastern-time collection quarter; a random response reference, withdrawal-secret hash, integrity digest and expiry. No names, email addresses, employer names, exact work locations, employee IDs, health details or open comments are requested in the response.
Independent responses have no account, company or invitation identifiers. If you separately choose the company-code pathway below, your response includes an opaque company survey reference. No personal account or invitation identifier is attached to answers. Invitations are delivered and administered separately. We do not provide employers with individual research answers or lists of who completed, skipped or declined the survey. Connected employees may separately choose the optional company counter described below. Because public participation does not include a verified workplace identifier, distinct-workplace counts and workplace clustering cannot be verified from these responses.
Privacy and foreseeable risks
Questions about leadership and work conditions may feel uncomfortable. Unusual combinations of industry and background categories may still make someone recognizable. Confidentiality cannot be absolute. This is not a guarantee of complete anonymity.
Answers are kept in private research objects, outside ordinary account and company working records. An index contains random object references and expiry dates for retention maintenance, without answers or membership identifiers. Todd Jerome Jenkins administers the research records; authorized hosting/service personnel may have technical access. Ordinary employer and learner accounts have no individual response-reading interface; company managers receive only reviewed grouped reports under the separate choice below.
The website runs on ChatGPT Sites and its Cloudflare hosting infrastructure. Technical systems may process network addresses, request times, security information and ordinary website logs. Survey pages and submission routes are excluded from the website’s ordinary analytics-event collection. A daily hash derived from the request network address is used for rate limits; the raw address is not copied into survey records, and these rate-limit records expire after 48 hours and are removed by maintenance. A hash and hosting logs can still create privacy risk. No guarantee about a particular storage country or absolute anonymity is made. Only the optional broad U.S. citizenship category described below is collected. Nationality, immigration or work-authorization status, birthplace and supporting documents are not requested.
Submission verification and accessibility
When enabled on the submission screen, Cloudflare Turnstile checks for automated traffic before a new response can be saved. It may show a checkbox. This is a security step, separate from the 50 questionnaire items. It does not verify your identity, eligibility, citizenship, truthfulness or unique participation.
Turnstile processes browser and network signals, including your IP address, user-agent, TLS fingerprint and the site origin. Cloudflare uses these signals to protect websites and improve bot detection under its own Turnstile Privacy Addendum. We do not send questionnaire answers, background choices, private receipts or account/company identifiers to the verification API. Verification tokens are used transiently and are not saved with research responses, company counters or ordinary analytics. Cloudflare controls its own signal and operational-log retention; the research dataset’s 730-day limit is not a promise about provider logs.
If verification expires or fails, retry it while keeping this page open. Answers, consent and any pending receipt remain in this browser tab; reloading or closing the page can lose unfinished answers. No name, identity document, phone number, purchase or extra account is required. If you cannot complete the check, use accessibility support; do not send answers, sensitive background details or your private receipt through a support form. Asking for help does not submit a response or bypass verification, and no alternative submission method is promised. Existing duplicate and shared-network rate controls remain. A passing check does not establish response quality.
Retention and withdrawal
Responses are excluded from new analysis after 730 days and removed by daily retention maintenance. Temporary service interruptions may delay physical deletion; expired records remain ineligible for analysis. Hosting backup and operational-log retention are controlled by the service providers and are not promised to end at the same instant.
Before sending, your browser creates a private receipt. Copy it when shown. The withdrawal page uses this receipt to remove the retained individual response. We cannot recover a lost receipt by name or email, because those identifiers are not attached to the response. Already published grouped findings cannot be recalled one response at a time. No survey drafts are saved in browser storage; losing or reloading an unfinished form may lose its contents.
Optional citizenship comparison — research purpose and risks
We ask whether people who currently hold U.S. citizenship and people who do not describe different safety experiences, and whether observed patterns suggest possible gaps in support worth investigating. The optional question is “Do you currently hold U.S. citizenship?” Choices are “Yes — including dual U.S. citizenship,” “No — I do not hold U.S. citizenship,” and “Prefer not to say / skip.” Skipping does not affect survey participation, account access, benefits or services. People working at U.S. workplaces can be eligible regardless of citizenship.
This is self-report for research, not an eligibility or employment check. Citizenship does not determine immigration status, work authorization, language, ethnicity or birthplace. We do not ask for those statuses or documents, and do not infer citizenship from accounts, names, email, language or industry. We do not distinguish naturalized, birthright or dual U.S. citizens. Earlier responses remain under their original notices; missing citizenship in an earlier version is “not asked,” not a citizenship group or a refusal.
Citizenship is sensitive. A confidentiality failure could create workplace, social or immigration-related concerns. Rare combinations of industry and background categories can increase identification risk. The researcher and authorized hosting personnel may have technical access to private records. Absolute anonymity is not guaranteed. The response can be withdrawn using the same private receipt under the existing retention rules.
Citizenship is never attached to a company counter, invitation, account profile or operational jSMS record. Employers receive no citizenship breakdown, individual status or answer, and no company-by-citizenship table. Only separately reviewed grouped research comparisons may be published. A comparison requires at least 30 usable responses per dimension in each citizenship group plus privacy/quality review, including small cells, complementary totals, rare combinations and repeated releases. Thirty is a project floor, not a power calculation, anonymity guarantee or NFA requirement. Choosing “Prefer not to say” is reported separately where safe; it is never recoded as noncitizen.
Comparisons are exploratory. Differences could reflect industry, job role, tenure, shift, recruitment, response patterns or measurement differences; they do not establish that citizenship caused an experience or prove a community is underserved. Earlier responses were English-only. New English/Spanish participation still misses other language communities and does not establish complete or equitable coverage. Language is never used to infer citizenship. Differences between questionnaire editions, recruitment and comprehension may affect comparisons; measurement equivalence in this U.S. sample has not been established. Language/version and collection-period comparability must be assessed before pooling or comparing results, with small language cells protected. The separate optional company report keeps questionnaire editions separate under the explicit company-code choice; no standalone demographic language breakdown is provided. No finding of “no difference” establishes equal access or adequate service. Group comparability and these coverage limits must be assessed before interpreting a difference. Follow-up with workers about specific access and support needs would be necessary before choosing interventions or claiming improvement.
Workplace state · optional
We ask which state or U.S. territory you mainly work in to understand participants’ geographic coverage. Choose the workplace you are describing, rather than your residence or employer’s headquarters. You can select Multiple states / no main state, or Prefer not to say / not sure / skip. State and territory names are shown in English. We do not ask for a street address, city, ZIP code or exact location.
The choice is kept only in your private research response and excluded from accounts, invitations, company counts and employer reports. We do not infer state from your network address or account. Earlier records remain unchanged: an absent field means not asked. Combinations of state, industry and other categories can increase identification risk. Any public state summary requires at least 30 usable responses per dimension plus privacy and quality review, including suppression or combination of small or rare groups and protection of complementary totals. This volunteer sample does not represent all workers in a state. Results are not published automatically.
Reporting and data use
We plan reviewed quarterly summaries, not automatic publication of raw data. NOSACQ-50 is scored as seven 1–4 dimensions with specified reverse coding and missing-answer rules. There is no individual score, compliance percentage, employer ranking or automatic judgment of a worker. Public volunteers are a self-selected sample. These responses do not establish shared workgroup climate, representative industry performance, causal effects, or that a safeguard functions.
For any published numeric industry or subgroup cell, the initial project minimum is 30 usable responses per dimension plus a disclosure review. Thirty is a reporting floor, not a statistical power calculation or a guarantee of privacy. Suppress or combine small/rare cells and complementary totals. Do not publish detailed crossed background tables by default. Verified distinct-workplace coverage is unavailable in this public collection; we will say so, will not claim that a five-workplace threshold has been met, and will not present confidence intervals that assume independently sampled workplaces. The first release may be a methods/participation update if safe, interpretable dimension estimates are unavailable.
Only grouped findings and methods are intended for public release. Individual response downloads will not be given to employers or sold. No transfer to NFA’s international database or unrelated future reuse is authorized by this notice; those would need separate disclosure and an appropriate consent basis. Results and practical follow-up decisions should be accessible to participating workers.
Optional company participation count — a separate choice
After a quarterly response under a notice that already disclosed the counter (v2, v3, v4, v5, v6, v7 or v8) is received, a signed-in employee with an active company connection can separately agree to be counted for one company in that quarter. Declining does not change the research response, ordinary access or paid benefits. This is not automatic enrollment and responses under notices that did not disclose this counter remain ineligible. A company total counts only connected people who separately opted in; it is not a complete count of all employees who may have answered publicly, a company response rate or a company safety-climate score.
The website checks the current company connection and private receipt. The separate counter stores the company connection, quarter, a keyed account hash to prevent a second count, a different secret-derived proof, count-consent version and coarse expiry. It stores no name, email, research response reference, item answers, exact response timestamp or background categories. The research answer object has no counter join key. Only the separate company-code choice described above adds a company survey reference, never the counter’s account/proof keys. Technical hosting/security logs may still create linkage or inference risk; authorized operators and service personnel may have technical access. Absolute anonymity is not guaranteed.
Employers receive no live count, individual status, identity list, completion timestamp, answer, demographic or language breakdown or citizenship information. A protected total is finalized only after the quarter closes and a seven-day buffer. Counts below 20, and counts leaving fewer than 5 currently eligible nonparticipants, are withheld. Withheld results do not disclose whether a group was small, near-total or unavailable. These project choices reduce disclosure risk; they are not NFA rules or a guarantee that an employer cannot infer participation using information volunteered elsewhere. No individual reminders are based on participation.
Before finalization, withdrawing with the same private survey receipt removes both the retained research response and its separate counter record. After finalization, individual records may still be removed but a finalized protected total cannot be individually recalled. Counter records, private hash keys and company totals expire 730 days after the collection quarter ends; this coarse expiry avoids copying an exact answer timestamp. Company counting is available only before that quarter closes. Optional company-count page. Notice: JRI-NOSACQ-company-count-2026-10-27.1.
Independence and commercial interests
Todd Jerome Jenkins also offers safety products, training and jSMS services. This financial interest is disclosed. Purchases and company relationships must not determine findings, participation or access to individual answers. This survey is separate from JRI Study 001, doctoral research, university resources and federal funding, as stated by the project owner. That statement is not an independently issued legal or regulatory determination.
Questions, concerns and changes
Contact the research team about methods, withdrawal difficulties, participant rights or privacy. Do not send your answers or sensitive employee information through the general inquiry form. The researcher handles these concerns; there is currently no independent research-ethics complaints office for this project. The study may be paused for privacy, scientific or ethical concerns. Material changes require a new notice and consent version before new collection; previous consent does not silently cover broader use.
Historical v1 participant notice. Earlier records remain governed by their original notice; the optional company counter is not added retrospectively.
Historical v2 notice — quarterly collections before the citizenship question. Its already-authorized optional company counter remains available under the original separate notice; no citizenship is added to earlier records.
Historical v3 notice — English collection with optional citizenship. No language is retrospectively added or inferred in earlier records; their existing instrument version remains unchanged.
Sources and consent
NOSACQ-50 was developed by the Nordic occupational-safety research network, with NFA hosting its official versions. Official English 2019 questionnaire · Official Spanish–Colombia 2020 questionnaire · NFA language directory · Permission for academic/commercial use · Kines et al. (2011), original instrument-development article. Use does not imply NFA endorsement. AI assistance supported implementation and drafting; no independent technical or ethics review is claimed.
The online consent requires an affirmative eligibility confirmation and your freely selected agreement to this information. Reading this page, creating an account, receiving an invitation or belonging to a company does not give consent.
Earlier bilingual v4 notices remain available: English v4 · Español v4. Earlier responses retain their original versions and are not retrospectively verified.
Earlier v5 notices remain available: English v5 · Español v5. Earlier responses retain their original notice.
A minimal withdrawal marker stores only the random response reference and an expiry for 730 days after withdrawal, to stop an old retry from restoring derived scores. It contains no answers, company or identity.