This is one of the most overlooked areas in cross cultural analysis. Some of the questions to ask are: How was the questionnaire used in each country developed? Does the translation attain equivalence of meaning? The issue of cultural equivalence also applies to the adaptation of questions within a language. For example, the adaptation of the questionnaire for use in France, Switzerland and Canada will mean that the questions are different. Translation and adaptation (or localisation) of questions can introduce important differences in meaning as seen in the IALS.

In developing a questionnaire to be used across languages or cultures it is important to try and establish linguistic equivalence. This means that the concept being measured is the same in both target languages. Being aware of the difficulties of achieving linguistic equivalence is an important first step. It is also important to recognise that questions in the source language(s) are often improved by putting them through the process of translation. In developing a questionnaire it is better to work in multiple languages from the start, in an iterative process, with all versions of the questionnaire being improved at the same time. Many questionnaires for large studies are now developed simultaneously in more than one language with several translators. In the past questionnaires were often developed in a single language, usually English and then translated or adapted. Back translation was usually the recommended method for checking the accuracy of the translation. This is no longer recommended.

If you are looking at variables which are outside the range of standard classificatory variables or are new questions in an ongoing survey you should ideally look at the questionnaire in each country that you want to include in your analysis in the original language if you have sufficient awareness of the language. This is to ensure that the questions are asking about the same concepts and about the same phenomenon. It is very easy for subtle changes to be made in translation that can change the meaning or intent of the question. In many surveys it is difficult to get access to the various language versions of the questionnaire.

More attention is now paid to trying to achieve cultural equivalence of questionnaires and this is a quickly developing area with new guidance being developed about translation processes. The European Social Survey translation guidelines are a good example of how practice is changing in relation to survey instruments as are the PISA guidelines. The questions raised by countries show some good examples of the practical problems encountered in translating survey questionnaires. For more on survey translation see Harkness, Pennell and Schoua-Glusberg (2004).