| Variable | Type | Definition | Used in multivariate models* |
| Loge(Y) | Dependent variable | Loge transformation of individual measurements of inhalable dust (mg/m3) or CSF (μg/m3) | X |
| Country | Categorical | Country where survey was performed (Germany, Poland, the United Kingdom, Sweden or the Netherlands | X |
| Survey purpose | Purpose of sampling survey | – | |
| Sample type | Personal, stationary or source-oriented measurement | X | |
| Sample strategy | Representative or worst-case sampling | – | |
| Department | Department (BRMA coding) where measurement was done | – | |
| Groups | Grouped departments | X | |
| High-exposure | Dummy | A priori defined high exposure processes in a department | X |
| Industrial sector | Production of tyres or general rubber goods | ||
| BRMA/HSE | Measurement collected by BRMA or HSE (UK only) | X† | |
| 4–8 h TWA | 8 h time weighted average measurement | – | |
| <4 h TWA | 4 h time weighted average measurement | – | |
| Dust fraction | Inhalable or respirable dust measurement | X | |
| General trend | Continuous | One trend in time for occupational exposure in all departments and country | – |
| Nested trend | Distinct time trends for group of departments per country | X |
*All variables included in the multivariate models (marked “X”) as interaction terms with aggregated department groups (see online supplemental material for details of codes).
†Added as a main effect in the “inhalable dust” model.
CSF, cyclohexane soluble fraction; TWA, time weighted average.