Is SOEP-IS a representative study / sample?

SOEP-IS is a representative study for the German population. The “normal” samples of SOEP-IS are random probability samples. While interviews are only conducted with people aged 18 and older, all residents of a surveyed household are included at gross sample level. So, the data also provides basic information about individuals that are below 18. These individuals are also incorporated into the weighting process of SOEP-IS.

Anyway, of course there is some kind of selection and self-selection. The survey is only conducted in German, which limits the representativeness of possible respondents who do not speak German. Some sub-samples also have a special design in specific survey years (e.g. web interviews only), which of course also influences the selection process and can bias the sample. Furthermore, participation in the study is completely voluntary. Not all people who are selected in the sampling process actually take part. To compensate for the non-responses that bias the sample, survey weights are generated and included in the data. SOEP-IS provides weights on individual- and household-level. These weights are needed to be able to draw conclusions for the total population.

In general, the same basic conditions regarding representativeness apply to SOEP-IS as to SOEP-Core. However, SOEP-Core has some special samples and different survey designs to reach certain groups of people (such as the “migration samples” or the “high-income sample”). These special samples do not exist in SOEP-IS, so the representativeness might be more limited for certain groups of the population than in comparison to SOEP-Core.