Staff Wellbeing Survey Template: Beyond the Generic Employee Satisfaction Form

Why Most Survey Templates Fail the WHS Test

The average off-the-shelf employee survey asks whether people feel valued, whether they understand the company's direction, and whether they would recommend the organisation as a place to work. These are Net Promoter Score cousins. They measure sentiment. They do not measure hazard.

Under the model Work Health and Safety Act 2011, a Person Conducting a Business or Undertaking (PCBU) has a primary duty of care to eliminate or minimise risks to health — including psychological health — so far as is reasonably practicable. The harmonised WHS Regulations now specifically list psychosocial hazards: job demands, low job control, poor support, lack of role clarity, poor organisational justice, remote or isolated work, and traumatic events, among others.

If your survey instrument cannot produce a score against those specific hazard categories, it cannot satisfy the risk identification step. You may be running surveys every year and still have no defensible evidence you have discharged your duty.

Safe Work Australia's 2023 data showed that 45% of workers' compensation claims with a mental health primary diagnosis had at least one identifiable psychosocial hazard that could have been detected through a properly designed survey. The cost per claim averaged $24,300 in lost-time wages alone, before legal and investigation costs.

The 14 Hazard Domains Your Template Must Cover

ISO 45003:2021 — the international standard for psychological health and safety in the workplace — identifies 14 work-related hazard domains. Any wellbeing survey template worth running should include at least two validated items per domain.

Two items per domain gives you 28 core questions. Add 4–6 demographic slices (work unit, tenure, shift pattern, employment type) and you have a survey that takes around 8 minutes to complete — below the threshold where completion rates drop significantly.

The Regulatory Framework Driving Survey Design in 2026

Three layers of Australian law now shape what a wellbeing survey must achieve.

First, the model Work Health and Safety Act 2011 (adopted as WHS legislation in all mainland states and territories except Victoria, which maintains its own aligned OHS Act 2004) requires PCBUs to identify hazards, assess risks, and implement controls. The consultation duty under s.47 means workers must be involved in the hazard-identification process — a survey is a direct mechanism for satisfying this.

Second, the harmonised WHS Regulations now include a dedicated psychosocial hazards chapter. Regulators in Queensland, South Australia, and the ACT have already issued improvement notices to organisations whose hazard identification relied solely on incident reports and return-to-work data, with no worker-reported hazard pathway. Survey data fills this gap.

Third, the Fair Work Act 2009 (Cth) creates indirect survey obligations through the general protections provisions. An employee who raises wellbeing concerns via a survey and is subsequently disadvantaged has a potential adverse action claim. This means your survey design must include: genuine anonymity (or clearly disclosed identifiable data collection), a documented response process, and a feedback loop to participants. Surveys that disappear into head office with no visible action are a liability, not a defence.

ISO 45003:2021, while not legislation, is increasingly cited by regulators as the reference standard for demonstrating good-faith compliance. Aligning your survey instrument to its hazard taxonomy is the single most effective way to demonstrate due diligence.

Designing for Action, Not Reporting

The failure mode of most wellbeing surveys is not bad questions — it is no action. The survey runs, a PowerPoint appears at the next all-hands, and nothing changes. Eighteen months later the next survey shows slightly worse scores, and no one is surprised.

Build your survey template around the control cycle, not the reporting cycle.

1. 01

   Pre-survey hazard register review

   Identify which of the 14 domains have already surfaced through incidents, complaints, or return-to-work data. These become priority analysis domains. You are not starting from zero.

2. 02

   Stratified sampling design

   Ensure the demographic cuts you care about — overnight shifts, remote workers, frontline versus office — have sufficient sample sizes (minimum 30 responses per subgroup) to produce meaningful results at that level.

3. 03

   Pre-commit to control thresholds

   Before the survey launches, document what scores will trigger what actions. A domain score below 55/100 triggers a manager-level workshop. Below 40/100 triggers an immediate risk assessment. This removes the politicisation of results.

4. 04

   Closed-loop reporting within 30 days

   Participants should see aggregated results and an action summary within 30 days of survey close. Anything longer and trust in the process — and your next response rate — drops sharply.

5. 05

   90-day control verification

   A short 5-question pulse on the lowest-scoring domains 90 days after the annual survey gives you evidence that controls have had measurable impact. This is what a regulator wants to see.

Response Scales, Anonymity, and the Questions You Cannot Ask

A 5-point Likert scale (Strongly Disagree to Strongly Agree) is the standard for psychosocial hazard surveys and allows you to benchmark against published norms. Some platforms use a 0–10 NPS-style scale for selected items — this is fine for pulse checks but produces different psychometric properties, so do not mix them in the same instrument without normalisation.

Anonymity design matters more than most organisations realise. Under the Privacy Act 1988 (Cth) and the Australian Privacy Principles, if your survey collects information that could be used to identify an individual — including combinations of demographic fields in small teams — you are handling personal information. A team of four people where one is the only woman over 50 on a night shift is not anonymous regardless of what the consent form says.

Questions you genuinely cannot ask in a standard workplace survey — because they require clinical training to interpret safely and may trigger mandatory reporting obligations — include direct questions about self-harm ideation, substance dependence, or diagnosed mental health conditions. Validated screening tools like the K10 or PHQ-9 should only be deployed with a clear clinical pathway for high-distress responders.

Instead, measure the work conditions that drive these outcomes. "I have enough time to recover between demanding periods at work" is a psychosocial hazard item. It tells you what you can control, and it does not expose your organisation to the downstream liability of holding sensitive health data without a clinical response capability.

Benchmarking: What Good Looks Like in Australian Workplaces

Without benchmarks, a score of 62/100 on job demands tells you nothing about whether that is a problem or an achievement. OccuSpan's population dataset — drawn from over 340,000 Australian worker survey responses across mining, construction, healthcare, logistics, and professional services — provides sector-specific percentile rankings for each of the 14 ISO 45003 domains.

Across all sectors, the three domains with the lowest average scores in the 2025 data cycle were:

• →Organisational justice — average 51/100 (lowest in healthcare and public administration)
• →Organisational change management — average 49/100 (lowest in mining and resources during transition periods)
• →Recognition and reward — average 53/100 (lowest in logistics and warehousing)

Knowing where your organisation sits relative to sector peers matters because a regulator is unlikely to accept "our scores are typical for our industry" as a defence if those typical scores fall below 50/100. What it does give you is a prioritisation tool — if your justice score is 48/100 against a sector average of 51/100, that is a targeted problem requiring targeted controls, not a culture-transformation programme.

Frequently Asked Questions