Opposite
Psychosocial
risk profile
Work design diagnostic · Wave 1
Opposite
April 2026
Issued 9 September 2026
Work design diagnostic · Wave 1
Two measures place each domain: intensity, the mean frequency across every hazard item in that domain, and the system of work rating, from Non-existent to Well-established. Status is set separately, by the impact and duration people reported.
The two domains with the highest intensity are the two with the weakest system of work behind them, both rated 2.0 out of 5. The organisation has controls for the hazards it has already named, and none for the two generating the most exposure.
Status combines how often a hazard occurs with how much it affects people and how long the effect lasts, the same likelihood-by-consequence logic used in a WHS risk assessment. Three domains cleared that threshold.
Volume, cognitive load, emotional labour and rosters. There is more work than time, and the roster does not leave room to recover from it.
Q01“You have too much work to do, or it is hard for you to complete your work within the time available.”
Procedural fairness, consultation and job security. Change lands before the conversation about it does.
Q07“Changes are introduced in a way that leaves you without enough consultation or practical support.”
Recognition, manager availability and peer support. Effort goes unnoticed more often than help goes missing.
Q08“Your efforts go unnoticed, or you do not receive the feedback you need to improve.”
The remaining three domains read Monitor or Stable and need no action this wave. Their exposure, intensity, respondent ratings and system scores are on the Detail page.
Four strain indicators and one engagement indicator, then the six protective factors that reduce the harm a hazard causes even when the hazard itself cannot be removed.
Mean of the four strain items, lower is better. Above the scale midpoint. Emotional exhaustion is the highest single item at 3.6, with 3 of 5 reporting it at 4 or above.
Enthusiasm and immersion at work, reverse-scored so higher is better. Holding up.
Strain elevated while engagement holds is the pattern of a committed group absorbing overload personally. This is the point at which redesign is cheapest.
Team backing and sense of meaning are the assets to build on. The weakest, having the right tools, technology and information at 2.8/5, is also the most directly fixable, and it feeds straight back into the demands domain.
How well the systems controlling each domain are working, and where each respondent places the organisation on the four-rung maturity ladder. Together these tell you whether a hazard is an isolated event or a control failure.
| System | Cohort mean | Mean | Range |
|---|---|---|---|
Systems for managing workload and demands |
Non-existentWell-established |
2.0 | 1–3 |
Systems for managing change, fairness and job security |
Non-existentWell-established |
2.0 | 1–3 |
Systems for leadership support and feedback |
Non-existentWell-established |
2.8 | 2–4 |
Systems for supporting role clarity and employee autonomy |
Non-existentWell-established |
3.0 | 2–4 |
Systems for addressing relationships and preventing harmful behaviour |
Non-existentWell-established |
3.4 | 3–4 |
Systems for managing physical safety, isolation and conditions |
Non-existentWell-established |
3.4 | 3–4 |
Each respondent picks the rung that best describes their experience. Bars show how the 5 responses fell.
Maturity clusters at Level 2, Aware: “we have policies and training, but nothing really changes in how we work.” The system ratings agree, and they locate the gap precisely. Relationships and environment score 3.4 and 3.4 out of 5, so those controls are landing. Workload and change score 2.0 and 2.0, and those are exactly the two domains carrying the exposure.
Read together, this is not a policy problem. The organisation has controls for the hazards it has already named, and none for the two that are actually generating harm.
Ordered by the domain ranking on page 01, and pitched at the level of the work rather than the level of the individual. Under the WHS Regulations the duty is to eliminate psychosocial risk so far as is reasonably practicable and to minimise it where elimination is not, which puts job and system redesign ahead of resilience training in every case.
Demands is the highest-intensity domain at 3.00/5 with 80% exposure. Both Priority ratings cite major impact lasting several days or recurring. Make current commitments visible in one place and sequence them explicitly, so saying yes to new work visibly requires deferring something else.
Change, fairness and security carries the second-highest intensity and its control system is rated 2.0/5. Consultation (Q07, mean 3.2) scores worse than transparency (Q06, mean 2.8), and these are hazard items, so higher is worse. A standing pre-announcement step addresses that mechanism directly.
The weakest protective factor is having the right tools, technology and information at 2.8/5. The cheapest item on the list, and it feeds the demands domain directly, since time lost to missing information reappears as workload.
Maturity sits at Level 2, Aware. Moving to Level 3, Supportive, means a standing mechanism that notices overload without someone having to raise it: a recurring load review, agreed trigger points, and a named response. Without it, action one decays back within a quarter.
At 5 responses the ranking is directional. Repeating in three to six months gives a trend on the domains you have acted on, and lifts the sample past the threshold where team-level cuts become releasable under the suppression rule.
Protect what is already working. Relationships and environment are the two Stable domains, and team backing is the strongest protective factor at 4.0/5. The common way to lose them is to answer workload by fragmenting the team through individual targets, reallocation, or removing shared time. Treat that as a control to preserve, not a cost to cut.
Every figure used in this report, with the full item set and the rules that produced each status. Nothing here changes the story on the earlier pages; it is the audit trail for anyone who wants to check a number.
| Domain | Exposure | Intensity | Respondent ratings | System of work | Status |
|---|---|---|---|---|---|
Demands & workload Volume, cognitive load, emotional labour, rosters |
80% 4 of 5 | 3.00 of 5 |
2× Priority1× Monitor2× Stable |
2.0 / 5 |
!!Priority |
Change, fairness & security Procedural fairness, consultation, job security |
80% 4 of 5 | 2.80 of 5 |
1× Priority2× Monitor2× Stable |
2.0 / 5 |
!Elevated |
Control & role clarity Autonomy over how and when work is done; role conflict |
60% 3 of 5 | 2.50 of 5 |
1× Monitor4× Stable |
3.0 / 5 |
~Monitor |
Support & leadership Recognition, manager availability, peer support |
60% 3 of 5 | 2.20 of 5 |
1× Priority1× Monitor3× Stable |
2.8 / 5 |
!Elevated |
Environment & conditions Occupational violence, trauma exposure, isolation, physical conditions |
80% 4 of 5 | 1.95 of 5 |
1× Monitor4× Stable |
3.4 / 5 |
~Monitor |
Relationships & behaviours Conflict, bullying, sexual harassment, gendered hostility |
20% 1 of 5 | 1.35 of 5 |
5× Stable |
3.4 / 5 |
✓Stable |
| Impact reported | Duration reported | Rating |
|---|---|---|
| Major or serious ongoing | Several days, or ongoing / recurring | !!Priority |
| Major or serious ongoing | Briefly, or part of the day | !Elevated |
| Minor or moderate | Any duration | ~Monitor |
| Agreement, or neither agree nor disagree | ✓Stable | |
A domain reads Priority when 40% or more of respondents rate it Priority or Elevated. Across the six domains, 4 respondent-level ratings reached Elevated or Priority; demands and change account for 3 of them. Colour never carries the meaning on its own; every status is labelled.
Section 2 of the instrument. One summary statement per domain, rated for agreement. Where a respondent disagrees, the form opens two follow-ups on how much the issue affects them and how long the effect lasts. Agreement closes the branch, so impact and duration exist only for the rows that signalled a problem.
“Reflecting on your overall demands and workload over the last month, my workload has been manageable.”
| Resp. | Agreement | Impact | Duration |
|---|---|---|---|
| R1 | Disagree | Major impact | For several days |
| R2 | Strongly disagree | Serious ongoing impact | Ongoing or recurring |
| R3 | Neither agree nor disagree | — | — |
| R4 | Disagree | Moderate impact | For most of the day |
| R5 | Agree | — | — |
“Reflecting on fairness and stability over the last month, I have felt that changes and decisions have been handled fairly.”
| Resp. | Agreement | Impact | Duration |
|---|---|---|---|
| R1 | Disagree | Moderate impact | For several days |
| R2 | Strongly disagree | Major impact | Ongoing or recurring |
| R3 | Disagree | Moderate impact | For part of the day |
| R4 | Neither agree nor disagree | — | — |
| R5 | Agree | — | — |
“Reflecting on the support you received over the last month, I have felt well-supported by my manager and colleagues.”
| Resp. | Agreement | Impact | Duration |
|---|---|---|---|
| R1 | Agree | — | — |
| R2 | Disagree | Major impact | For several days |
| R3 | Neither agree nor disagree | — | — |
| R4 | Disagree | Minor impact | Briefly |
| R5 | Agree | — | — |
“Reflecting on your sense of control and clarity over the last month, I have felt clear about my role and able to influence how I work.”
| Resp. | Agreement | Impact | Duration |
|---|---|---|---|
| R1 | Neither agree nor disagree | — | — |
| R2 | Disagree | Moderate impact | For part of the day |
| R3 | Agree | — | — |
| R4 | Agree | — | — |
| R5 | Strongly agree | — | — |
“Reflecting on your work environment and conditions over the last month, I have felt physically safe and well-connected in my work setting.”
| Resp. | Agreement | Impact | Duration |
|---|---|---|---|
| R1 | Agree | — | — |
| R2 | Neither agree nor disagree | — | — |
| R3 | Agree | — | — |
| R4 | Disagree | Minor impact | For part of the day |
| R5 | Agree | — | — |
“Reflecting on workplace relationships over the last month, I have felt safe and respected in my interactions at work.”
| Resp. | Agreement | Impact | Duration |
|---|---|---|---|
| R1 | Agree | — | — |
| R2 | Neither agree nor disagree | — | — |
| R3 | Agree | — | — |
| R4 | Agree | — | — |
| R5 | Strongly agree | — | — |
Section 1 of the instrument. Twenty items, each naming a specific psychosocial hazard, answered on a five-point frequency scale. Domains appear in order of intensity.
Respondents may name a hazard the instrument does not cover and rate its frequency. None of the 5 added one. This wave also ran the standard instrument only; custom items are set up per organisation and appear with the same distribution treatment as the hazard items.
Sound at this sample: which domains carry the most exposure and their order, whether a pattern is broad or concentrated, which control systems are rated weakest, and whether protective factors are intact. Not yet supported: quoting exposure as a percentage, any cut by team, site, role or shift, or claiming a trend.
Suppression rule. No cell with fewer than five responses is reported, and sensitive-domain items are only ever released at whole-of-organisation level. This protects respondents, and it is a precondition for getting honest answers to the relationships and environment questions in the first place.
Sydney Trains · cleaning and maintenance
Opposite was engaged under the Foundations in Health Program to assess psychosocial risk in the Sydney Trains cleaning and maintenance workforce, and to design the interventions that followed.
Claims analysis and risk-rating workshops set the baseline that shaped the survey and interview design. Five factors accounted for almost all recorded occurrences.
responses across Transport for NSW and Sydney Trains. 58 leaders, 91 employees. A 42-item instrument on a five-point scale, built from the psychosocial risk framework.
24 individual and situational items for everyone, 6 on engagement and psychological capital, and 12 organisational items answered only by leaders.
face to face, semi-structured. 14 cleaners, two cleaners in charge, three rail maintainers, two train technicians, and one each of electrician, engineer, shunter and delivery manager. Tenure from under a year to 24 years.
Opposite's psychosocial framework, built on the Safe Work Australia categories and mapped to Transport for NSW's psychological risk management procedure. Every survey item traces to one of these sub-factors, asked at up to three levels.
Only staff who identified as leaders completed the organisational items. Non-leaders would not usually have visibility of organisation-wide practice, so for them the same hazard is experienced at the situational level instead. This is also what makes the leader-versus-staff comparison on page 06 possible.
Four risk factors, each scored out of 5, alongside the share of people answering favourably. Support and relationships is the clear focus area, at an overall mean of 3.41 across the instrument.
Each spoke is a sub-factor and each ring is one point on the five-point scale, so a short spoke is a weak score. Hover or tap a segment to read it.
Support and relationships holds three of the four lowest spokes. Job design and organisational systems sit furthest out.
Each sub-factor was asked at the individual, situational and organisational level. Organisational items were answered only by the 58 leaders in the sample.
| Sub-factor | Individual | Situational | Organisational |
|---|---|---|---|
| Organisational change management | 3.28 | 3.64 | 2.74 |
| Organisational justice | 3.85 | 3.83 | 3.37 |
| Workplace relationships | 3.14 | 2.62 | 2.72 |
| Recognition and reward | 2.81 | 3.20 | 2.69 |
| Support | 3.11 | 3.47 | 3.24 |
| Job control | 3.17 | 3.88 | 3.43 |
| Role clarity | 3.95 | 3.62 | 3.13 |
| Type of work | 3.48 | 3.52 | 3.15 |
| Job demands | 3.09 | 3.05 | 2.63 |
| Environmental conditions | 3.68 | 3.16 | 3.74 |
| Remote work | 3.40 | 3.00 | 3.48 |
| Work-life balance | 3.23 | 3.17 | 3.48 |
The organisational column is the lowest of the three at 3.15, and it is the one only leaders answered. Recognition and reward, workplace relationships, job demands and change management all sit below 2.8 there, which is a group of leaders saying the organisation has not given them the systems to manage the things their teams are raising.
Five sub-factors showed a statistically significant gap, ranked by size of difference. Leaders consistently rated conditions and clarity more favourably than the people doing the work.
Across job design as a whole, employees rated favourably 67% of the time and leaders 83%. A gap this consistent means leaders are likely to underestimate the conditions their staff are working in, and staff are likely to read that as being unheard.
The 12 organisational items, answered by 58 leaders. Percentage rating each favourably.
Read alongside the ranked gaps above, this is the useful tension in the data: leaders rate their own working conditions more favourably than staff do, and rate the organisation's systems harder than staff rate their day to day.
Six items sat outside the risk framework, covering how engaged people were and the personal resources they had to draw on. Both held up better than the conditions around them.
Most people rated their engagement positively. Leaders found it easier to hold attention across a whole shift than staff did, which points at task variety rather than commitment.
Optimism, hope, self-efficacy and resilience. Moderately positive overall, and uneven between individuals.
Weakest was optimism at 3.18. Leaders reported significantly more self-efficacy than staff, so skill development and mentoring were the obvious lever.
The two engagement items scored 3.74 for holding attention across a shift and 3.66 for feeling absorbed in the work. Both showed a leader-staff difference on the first item only.
Each interview was themed and also scored, so the qualitative work maps back onto the survey. Scores ran higher than the survey throughout, which is what a non-anonymous conversation with a volunteer sample tends to produce.
| Area | Question asked | Mean |
|---|---|---|
Organisational systems | How is change managed at work? | 3.13 |
Environment | How would you describe your work conditions? | 3.71 |
Support & relationships | How well are you supported in your role? | 3.79 |
Negative events | How frequent are harmful events? | 3.96 |
Job demands | How would you describe your workload? | 4.08 |
Engagement | How engaged or enthusiastic are you about work? | 4.09 |
Wellbeing | Overall, how would you rate your wellbeing? | 4.13 |
Job design | How much control do you have in your job? | 4.46 |
Counts are the number of times a theme appeared across the 24 interviews.
Flexible rostering and grace periods to cancel shifts were valued. Against that, limited information about coming changes, no consultation, and leaders with little visibility of what staff actually do.
Peer and team support was consistently good. Supervisor and manager support was the single most raised improvement, including how complaints get handled.
Culture and atmosphere were positives. Safety concerns, break and quiet spaces, and uneven access to resources between sites were the issues.
Harmful events were rare and staff felt capable of defusing customer aggression. Targeted harassment and the reporting and resolution process were where the volume sat.
Workload was manageable when resourced. Fluctuation from weather and special events, rostering and staffing, and breaks were the recurring pressures.
The strongest area. Role clarity, autonomy and task rotation were all appreciated. Dependency on others and pay discrepancies between departments were the drag.
The diagnostic was phase one. Survey and interview data were analysed together to design interventions against specific sub-factors, rather than a general wellbeing program.
Both data sets pointed at the same three: workplace relationships, recognition and reward, and job demands. Interventions were scoped against those rather than against psychosocial risk in general.
Leaders rated their own conditions more favourably than staff and their organisation's systems harder. Both halves of that gap were treated as findings, since one drives underestimation and the other names the missing systems.
Engagement at 3.7 and psychological capital at 3.5 were the buffer holding the workforce steady. The sequencing put condition changes first so that buffer was not spent.