Workplace climate questionnaire: 12 questions to open a dialogue.

Before sharing the questionnaire.

Set a collection period, a feedback date and an owner for the follow-up. Only launch a survey when you can explain how the responses will be used.

1. Choose the scope: the whole organization or a work group large enough for protective reporting. Use the same reference period for everyone: the past four weeks.
2. Review the questions with people from the relevant roles to check understanding. Adapt the wording before the first wave and record later changes.
3. Explain that responses are optional, the collective purpose, who can access data and how findings will be shared. Only promise anonymity if the arrangements genuinely support it.
4. Offer comparable access for everyone in scope, announce the closing date and plan a collective reminder. Do not try to identify non-respondents from the results.

The complete questionnaire.

Instruction to display: “Thinking about the past four weeks, indicate your agreement with each statement. You may choose ‘Not applicable / Prefer not to answer’ or leave an answer blank. Do not include names, medical information or identifying details in the comment.”

Use the 1–5 scale for all statements. Exclude NR and blank responses from calculations; neither counts as zero or as a neutral response.

Read results without inventing an overall score.

For each question, show the valid response count, the distribution of responses from 1 to 5 and the share answering 4 or 5. Keep NR and missing response counts separate. Discuss each dimension with teams, without an overall average or individual ranking.

Favorable share = (number of answers 4 + number of answers 5) / number of valid answers 1 to 5 × 100. If no valid answers exist, display ‘not calculable’. A change in percentage points is the difference between two percentages, using a comparable question and scope.

Calculate participation separately: respondents submitting at least one closed question / people invited × 100. Define deduplication before collection without adding unnecessary individual identifiers. High participation alone does not prove that responses represent every situation.

Illustrative interpretation example.

Fictional data, unrelated to any Brainmood client. For Q03, 50 people were invited and 40 responded to the questionnaire. Of these, 36 answered Q03 on the 1–5 scale, 3 selected NR and 1 left the question blank.

Turn the finding into a discussion.

In the example, participation is 40 / 50 = 80%; the favorable share for Q03 is (12 + 6) / 36 = 50%. This describes the respondents, not the cause of a difficulty or anyone’s psychological state.

Useful feedback might say: ‘Half of valid responses indicate a workload that fits the time available. Which situations make work manageable, and which need a decision on priorities?’ Discuss work situations, choose one or two actions and announce a follow-up date.

Do not directly compare teams with different roles, periods or collection methods. For repeated measurements, record changes in headcount, context and participation before interpreting trends.

Prepare feedback that protects people.

Removing names does not guarantee anonymity: a small group, combined categories or a comment may allow someone to be identified. Assess aggregation in your context; no single minimum group size is a guarantee.

Before collecting data, define recipients, access rights, retention periods, comment handling and rules for combining or withholding results. Avoid individual variables and details that do not help the decision. Have your organization’s data protection lead review the arrangements.

Share findings about work and the decisions taken. Do not publish individual rankings, identifying quotations or medical details. These templates support collective organizational steering, not individual psychological assessment.

For background on removing identifiers and anonymization: CNIL guidance on data anonymization (in French).