Description
Making background and substantive measures comparable across countries or waves by designing a common input, mapping local items to agreed target variables, or recoding existing data.
Harmonisation is the deliberate work of making measures comparable when they are collected in different countries, languages, systems or waves. Input harmonisation fixes a common master questionnaire, developed sequentially, in parallel or simultaneously, and translates it. Output harmonisation lets each context ask locally appropriate items and maps them to an agreed target variable, such as ISCED for education or ISCO for occupation; this is ex-ante when the target and mapping tables are fixed before collection, ex-post when existing datasets are recoded afterwards. Rigour comes from documenting every deviation and mapping rule, and from assessing quality: is the concept relevant in each context, is the documentation complete, does the coding scheme fit, can each harmonised value be traced to its source, and are harmonised measures reliable. Collapsing to common categories always loses detail, and ex-post work is limited by what was originally asked.
Attributes
- Discipline
- Survey & Psychometrics
- Loss type
- Siloed
- Stage
- Design, Post-launch
- Data
- Analytical / expert
- Effort
- Weeks
Steps
- Decide between input harmonisation (one master questionnaire) and output harmonisation (local items mapped to target variables).
- For ex-ante output harmonisation, define the target classification and require a mapping table from each country or product version before fieldwork.
- For ex-post harmonisation, document the source variables and write explicit recoding rules for existing data.
- Record every deviation from the master or target, with its rationale.
- Assess quality: concept relevance per context, completeness of documentation, fit of the coding scheme, traceability from local to harmonised values, and reliability of the harmonised measures.
- Publish the harmonised variables with their mapping documentation.
Sources
- The SAGE Handbook of Survey Methodology (Wolf et al., 2016)
Also known as
- Input harmonisation
- Output harmonisation
- Ex-ante and ex-post harmonisation
- Target variable harmonisation (ISCED, ISCO)
Categories
- Analytics & Behaviour
- Survey & Psychometrics