Population Mean Study - Statistics

Part I:

The population mean study I’m choosing is a happiness index that rates citizens of Denmark the happiest. In the published articles it was stated that 65% of all Danes rated themselves very happy (or very satisfied). (Comparative graph at the end of the document.) I found a methodology page that explains the questions used and the process used to conduct the study. In general randomly selected citizens of surveyed countries where surveyed, face-to-face, by trained interviewers in a series of closed question interviews. The number of persons surveyed in each country varies based on the number of persons in the country. Luxemburg and Denmark have fewer persons interviewed than Germany.

This study is done by a firm that has been measuring public opinion in the EU since 1973. The questionnaire and question design seem to be appropriate with only closed questions. There is no mention of how the list of persons to be interviewed is obtained and it is seems likely that the initial list per county is generated in an inconsistent manner in each country. For example: Is the German list based on driver’s licenses and the Danish list based on nation health insurance register? Obviously a standard method across countries could eliminate this nonsampling error. Since I don’t suspect nonresponse is higher on a per country basis in the EU that is not much of a concern.

The non-sampling error that is most interesting to me is one of language. Since the interviews take place in the native language of the citizen members the issue of translation can be a factor in the results. I suppose this is exactly the same challenge as trying to ask balanced questions. In this case though languages don’t always directly translate - in the case of Danish there are 2 words that roughly translate to happiness. One is very happy and one is more like content or satisfied, the interview uses the one closer to content. I suspect the results be the same with a different...