Classifying and Explaining Life Conceptions: A Re-Analysis of the 14th Shell Youth Study 2002

This paper takes a second look at the data from the 2002 Shell Youth Survey, with the goal of building structure types that describe the life values held by young people in Germany and that line up clearly with socio-demographic background. The question driving the work is whether the value orientations of German youth can be grouped in a way that both makes statistical sense and matches how the original Shell researchers read their own material.

The analysis starts with 24 survey items. Factor analysis is used first to combine these items into workable scales, that is, sum indices. After the scales are built, the items are sorted into categories drawn from both sociological reasoning and statistical results, so the grouping does not rest on numbers alone.

To find the structure types themselves, we rely on configural frequency analysis (CFA) together with multiple correspondence analysis. These methods let the types be examined both as figures in a table and as points in a graph, and the picture that emerges corresponds to the interpretations offered by the authors of the original Shell study. Four categories of life values come out of this step.

Once the four types are in hand, the second part of the work asks how far they can be accounted for by background characteristics: sex, age, class position, and whether a respondent lives in the eastern or western states. The short answer is that these characteristics explain the types only in part. The links exist, but they are far from complete, and much of the variation in value orientation sits outside the standard demographic markers.

On the methodological side, we show that three tools work well together rather than competing: classification trees in the form of CHAID, prediction CFA built on Fuchs-Kenett residues, and the Kimball-Castellan test. Each one catches something the others miss, and reading them side by side gives a fuller account of where the structure types come from.

A final practical point concerns software. Every stage of the work, from preparing the raw data through to the statistical procedures described above, was carried out in SPSS using standard syntax, without recourse to specialized packages.