Conferences

Measurement error in survey research 2008

Miniconference on Measurement error in survey research (Berlin, 23–24 June 2008)

Different proposals have been made for the estimation of measurement errors in survey research. The most popular design is the split ballot design (see for example Schuman and Presser1981). Many big studies using this design have been done by Harris Interactive nowadays (Thomas et al.). Frank Andrews (1984) and Saris and Gallhofer (2007) have suggested to use MTMM experiments. This approach has been used for many different substantive variables (Scherpenzeel and Saris (1996). Harry Ganzenboom (2007) applied the same approach on objective variables like occupation. Alwin (2007) has suggested using quasi simplex models. Van der Veld (2006) has compared the advantages and disadvantages of the different approaches. Besides that one can ask the question whether one can ignore the categorical character of most of the data or that one should take this characteristic into account in the analysis. This point has been introduced by Oberski et al. (2007). In this miniconference we would like to discuss these issues which are essential for comparative research.

The following papers have been presented:

Willem Saris:

Different designs for estimation of measurement error in survey research

Harry Ganzeboom:

Multitrait-Multimethod models for occupational status

Randall Thomas:

Life satisfaction across countries: does the measure matter ?

William van der Veld:

Quality of survey items in a cross- national context: An illustration with social trust and political trust

Eldad Davidov:

A cross-country and longitudinal test of invariance : The human values questions in the third round of the European Social Survey

Daniel Oberski:

Discretization, unequal intervals, nonlinearity and non-normality in MTMM models: a latent class approach

Frauke Kreuter:

Good Item or Bad – Can Latent Class Analysis Tell? Examining the Effectiveness of a Latent Class Analysis Approach to Item Evaluation

Willem Saris:

Evaluation of the different designs for estimation measurement error in survey research