Why group jobs

In a psychosocial risk assessment, a unit of analysis makes it possible to examine shared conditions rather than relying on a company-wide average. It can combine jobs with similar tasks, responsibilities and organisation, provided the grouping makes preventive sense.

OSALAN recommends considering the nature of work, discretion and responsibility for supervising people. Combining jobs with very different conditions can produce an apparently moderate result that describes neither group adequately and obscures where intervention is required.

Define units before examining results

Base the design on actual work: sites, shifts, processes, reporting relationships and service or production arrangements. An organisation chart is a starting point, not an automatic answer. Identically named jobs may involve different conditions at different sites.

Agree and document units before analysing responses. Changing groups afterwards to improve a score or conceal a problem undermines the assessment. If a technical reason requires a change to the design, explain and record that reason clearly.

A decision example

A customer service operation includes on-site and remote workers. Even where tasks are similar, available support, interruptions and schedule control may differ. Combining them could hide a specific exposure. Conversely, reporting every individual job separately may prevent confidential feedback.

Choose comparisons that can inform concrete action. A sufficiently protected quantitative grouping may be combined with confidential qualitative methods. Do not publish breakdowns that make it possible to attribute answers to an identifiable person or reconstruct their responses indirectly.

Check coverage and limitations

Examine participation within each unit, not only across the whole organisation. A high overall response can coexist with one team almost absent from the data. State that gap before extrapolating conclusions to everyone in scope.

The report should describe grouping criteria, covered jobs and limitations. A suitable unit helps locate changeable conditions and check improvement while maintaining agreed safeguards. No single group size automatically guarantees anonymity, representativeness or technical quality; these require judgement about the actual data and workplace context.