Conducting analysis that seeks to describe or explain differences between countries or groups of countries provides a more complex set of challenges that analysing data from a single survey source in a domestic context. The analyst needs to exercise additional caution to ensure that the data is suitable for the analyses intended. It is your responsibility as an analyst to ensure that the analyses you undertake are valid. Perfect data doesn't exist so you will need to apply your professional judgement as to what is "good enough" to use for a particular purpose and what is not. This will mean making an assessment of the data and the extent to which it is fit for the purpose you intend. For example, if the topic is one in which there may be an element of competition or national pride such as educational attainment, or healthy eating, then you will want the sample to be representative of the whole population and the measurement to be fairly robust. For topics where countries probably wouldn't mind being at the bottom or top of a league table you might be less concerned about coverage. Ask yourself the questions,"How would I defend my conclusions if someone challenged them?" "Am I confident in the results of this analysis?"

Unit 2 in this series has already dealt with issues around making international comparisons using aggregate data. This chapter will look at issues when using micro data or primary data sources to do cross-national analysis. Many of the issues are similar since they may be on the same primary data although remember that the aggregate data may have been transformed in some way.

Many of the examples are draw from the International Adult Literacy Survey . This survey aimed to measure the literacy skills of adults through a household survey using a direct assessment more commonly associated with tests carried out in schools. The survey results were controversial and as a result a methodological review was conducted to investigate whether the results were subject to some form of cultural bias. The findings from this review were important in developing an understanding of the issues in cross-national surveys and in improving the design and implementation of surveys such as the European Social Survey.