In producing international datasets there are different approaches to harmonisation. In some surveys such as the ESS, the inputs are harmonised as much as possible - so the design in each country is as similar as possible and the questionnaire is standardised in advance. In other surveys, the international dataset set is harmonised at output level, so the specification is set as to the output variable and countries can arrive at that through their own national processes. In this instance the questions may be different by design.

For some international datasets it is useful to be aware that there are also sometimes corresponding national datasets available which may contain additional variables, slightly different versions of the international variables or indeed may contain additional sample. For example, many international data files use various international classifications such as the International Standard Classification of Occupations (ISCO) or the International Standard Classification of Education (ISCED). The national data files may contain the usual national classification for that country which may provide more depth. For example, an international survey may contain household income variables that are based on quintiles or deciles. The national survey micro data may contain the variable either as a continuous variable showing the actual amount of household income in the local currency, it may contain banded income in the local currency and it may contain the international percentiles variable.

Surveys may also have added questions of national interest which were not required for the international survey. Similarly, the population defined as in scope for the national sample may have been broader than that required. For example, an international survey on fertility may direct questions only at women aged 16-49 or in some countries only to women who are married or sexually active. The national samples may have used a more extended age range so additional sample may be available on a national version of the datafile. Similarly questions on drinking and smoking may be directed at different populations reflecting legal age for drinking, questions on employment may be asked of different ages reflecting the differences in the retirement age. The national data files may also contain more detail on income or education.