The NOM-035 questionnaire, in full
If you run a plant in Mexico, this is the instrument an STPS inspector expects to see applied. Both official forms — 46 items and 72 items — are published here with a spreadsheet download, and no email address is asked for. This page explains which one applies to you and how the file is judged; the instrument itself lives on the Spanish page, in Spanish, and the next section says why that is deliberate.
Read every item on the Spanish version of this page.
The questionnaire is in Spanish,
and that is not a gap
This is the question US HR teams ask first, and answering it wrong is expensive, so it goes above everything else on this page.
Why it is not translated here
NOM-035-STPS-2018 published its instrument with specific wording, in Spanish, in Mexico's federal gazette. That wording is the stimulus the worker responded to, and the response is the evidence.
Translate the items and you have administered a different instrument. Nothing stops you, and an inspector is entitled to ask why the applied questionnaire does not match the published one. The honest answer — "our head office reads English" — is not a compliance argument.
So the split that actually works is: administer in Spanish, report in English. The workforce answers the instrument as published; the corporate office reads the results in its own language.
What this means in practice
Bloomder ships both guides as Spanish templates, item for item, from the same source that generated the Spanish page — so the two can never drift apart.
The evidence report is issued in English or Spanish from the same administration: English for the US corporate office, Spanish for the comisión mixta that has to read it on site.
The instrument's language and the report's language are two separate decisions. Only the first one is fixed by law.
Which of the three guides
applies to your site
By headcount of the work centre, not of the company — a distinction that decides which form a multi-site manufacturer applies where. And one common trap: Guide I is not chosen by size, because it does not measure the same thing as the other two.
| Guide | What it identifies | Who applies it | Items |
|---|---|---|---|
| Guide I | Severe traumatic events experienced by the worker | Every work centre with 16 or more workers | Short screen |
| Guide II | Psychosocial risk factors | Work centres with up to 50 workers | 46 |
| Guide III | Psychosocial risk factors and organisational environment | Work centres with more than 50 workers | 72 |
Under 16 workers
No psychosocial risk questionnaire is required. Other obligations of the standard still are — publishing the psychosocial risk policy, handling reported cases, keeping the corresponding records — so "under 16" does not mean NOM-035 does not apply.
Several sites
The threshold counts per work centre. A 300-person company across four 75-person plants applies Guide III at each. The same company with one site of 240 and three of 20 applies Guide III at the large one and Guide II at the small ones.
The items do not all point
the same way
This is where a spreadsheet built in-house usually goes wrong, and it goes wrong quietly — the numbers still add up, they just describe the wrong risk.
0 to 4 points
Every answer scores on the five-point frequency scale the standard prescribes.
Some are reversed
Positively worded items are reverse-scored. Adding them the same way inflates or deflates the whole result.
Summed by domain
There is no single score. There are sums by domain and by category, each with its own threshold.
Compared to the tables
The standard's own tables set five levels, from null to very high risk. No tool replaces them.
The level obliges
At high and very high, the standard requires documented action. The level is an instruction, not a statistic.
What this page is not
It publishes the instrument and explains how it is applied. That is different from telling you that you comply:
- This is not legal advice. We cannot tell you whether your file satisfies your inspector. Ask a Mexican labour attorney or a compliance adviser.
- Scoring and interpretation are done against the tables of the standard, not against this page and not against any software.
- The employer remains the responsible party. Reproducing the questionnaire transfers no obligation.
The questionnaire is one piece
of the file, not the file
The common finding in an inspection is not a missing questionnaire. It is a questionnaire that was applied, with nothing around it.
What the file holds
- The questionnaires as applied, and to whom.
- Results by domain and by category.
- A documented action plan wherever risk came out high or very high.
- Evidence that results were communicated back to the workforce.
How long to keep it
- At least 12 months at work centres with up to 50 workers.
- At least 5 years above 50.
The clock starts at administration, and the standard expects the instrument to be reapplied on the cycle your headcount sets. A three-year-old file with no reapplication is an expired one.
When paper is enough,
and when it stops being
We just gave you the instrument for free, so the bias in this comparison is obvious. It is worth stating anyway, because the point where paper stops working is operational and recognises itself.
Paper is fine
One site, under 50 people, the 46-item Guide II, somebody with time to key in and score. Sealed boxes, no sheet carrying a name, the archive locked.
That is legal, it is sufficient, and buying software for it would be spending for the sake of it. Download the spreadsheet above and run it.
Paper stops being fine
At 72 items per person, several sites or several shifts. Keying in 200 questionnaires by hand is not hard, it is slow, costly and error-prone — and the reverse-scored items are exactly where the errors land.
Anonymity also gets hard to demonstrate. That nobody looked at the sheets is not evidence. That no record links an answer to a person is.
What Bloomder does with this
- Both guides ship as templates, with these same items.
- Distribution on the cycle your headcount requires, with a timestamped record of every send.
- Answers stored with no identity link, and nothing shown below five respondents.
- An exportable evidence report, in Spanish for the site and English for the corporate office.
What it does not do
- Score against the tables for you. That analysis belongs to the employer.
- Replace the attorney or the action plan.
- Guarantee the outcome of an inspection — the authority decides that.
Running it on 200 people takes weeks on paper
In Bloomder it takes an afternoon: both guides are loaded, sending and reminders are automatic, and anonymity is enforced in the data layer. Free up to 25 employees, no credit card.
Need the file audit-ready in 30 days? See the 035 Shield Program.