Effort-reward imbalance (ERI) is an established conceptualisation of work stress. Although a validated effort-reward questionnaire is available for public use, many epidemiological studies adopt shortened scales and proxy measures. To examine the agreement between different abbreviated measures and the original instrument, we compared different versions of the effort-reward scales available in 15 European cohort studies participating in the IPD-Work (Individual-participant-data meta-analysis in working populations) consortium. Five of the 15 studies provide information on the original ('complete') scales measuring 'effort' and 'reward', whereas the 10 remaining studies used 'partial' scales. To compare different versions of the ERI scales, we analyse individual-level data from 31,790 participants from the five studies with complete scales. Pearson's correlation between partial and complete scales was very high in case of 'effort' (where 2 out of 3 items were used) and very high or high in case of 'reward', if at least 4 items (out of 7) were included. Reward scales composed of 3 items revealed good to satisfactory agreement, and in one case, a reward scale consisting of 2 items only demonstrated a modest, but still acceptable degree of agreement. Sensitivity and specificity of a composite measure, the ratio of effort and reward, comparing partial versus complete scales ranged between 59-93 and 85-99 %, respectively. Complete and partial scales were strongly associated with poor self-rated health. Our results support the notion that short proxy measures or partial versions of the original scales can be used to assess effort-reward imbalance.