Pieces of the TSE-Puzzle: Measuring a theoretical notion in interrelation with data-proceedings.
Sprache des Vortragstitels:
VI European Congress of Methodology, Utrecht
Sprache des Tagungstitel:
The major sources of nonsampling-errors (cf. Biemer & Lyberg 2003: 37pp.) within the survey process can be divided into three areas: deficient concepts of measurement (specification- and measurement-errors), shortcomings within data-collection (frame- & nonresponse-error) and inaccurate data-handling (processing-error). This paper explores the interrelation between aspects of the first and last area using the example of student-participation. In this case the researcher?s challenge starts with the definition of the term ?student-participation? since a wide range of participation-forms exist. The multidimensional concept for quantifying participation based on 27 operationalized dimensions, as developed by Wetzelhütter, Paseka & Bacher (2013), rudimentarily reflects the corresponding difficulty. Subsequently, the respondents are facing (more or less) difficulties in answering (more or less) accurately formulated items which lead to (more or less) measurement-errors. E.g. a simple response-model already depicts four steps within the answering-process. The ?final? task consists of preparing the data collected in an appropriate way. Otherwise incorrect results are reported e.g. in case of open questions when coders work or instructions are deficient. However, what links these aspects together? One factor is ?satisficing? which is connected to motivational aspects. In order to test the consequences of motivation on i) response (measurement-error) and ii) coding operation (processing-error), different motivating strategies (incentives/appeals) were included in the survey. The survey population (n=1363 students) was randomly divided into (differently motivated) groups whereupon five students performed (differently motivated) the coding-task. Consequently, the impact of those experimental arrangements will be presented separately and combined, which should permit conclusions of accumulative effects.