About this Resource

An important question is whether the samples in each country are sufficiently similar that you are confident in making comparisons. Has the survey used probability sampling at all stages - that is that each element of the population has a known probability of selection that is not equal to zero.

In-field sampling procedures, that is where the final stage of sample selection is conducted in the field, are particularly important - pay attention to whether the final stage of sampling was carried out by the interviewer and the method used. Two main types of infield sampling may occur - where the address is being selected from within an area selected at random by the centre and where a household is being selected at a multi-household address or where an individual is being selected from within a household. Where possible it is better to separate the functions - in particular, where an address is being selected in the field this should be carried out by someone other than the interviewer to avoid difficult to interview addresses being substituted. It is more difficult to apply this to selection of households within addresses or to selection of individuals within households since contact with the households before the approach for an interview may jeopardise response rates. It is more usual therefore for these types of infield sampling to be conducted by the interviewer.

Timing and reference period - Were the surveys conducted at around the same time and if not does that matter - for some topics you might conclude that a difference in timing of 2 or three years might not be a problem since the topic of interest is unlikely to be subject to fluctuation. For other topics, such as social or political attitudes even small differences in timing might be problematic depending on the context. For example whether there were elections held or whether there were political events that might impact on the variables of interest. Similarly, are the reference periods in the questions the same - where they ask questions about expenditure for example do they ask about expenditure in the week ending last Sunday, in the past 7 days or in the past month or a longer period. This is particularly important when the question requires the respondent to recall events. The longer the time period the more difficult it is for the respondent to answer accurately.

The University of Manchester; Mimas; ESRC; RDI

Countries and Citizens: Unit 3 Making cross-national comparisons using micro data by Siobhan Carey, Department for International Development is licensed under a Creative Commons Attribution-Non-Commercial-Share Alike 2.0 UK: England & Wales Licence.