30221

SoSe 16: Advanced Data Analysis with SPSS

Dieter Ohr

Hinweise für Studierende

Participation in last semester’s course „Elementary Data Analysis“ or comparable knowledge is a prerequisite.

Kommentar

„Right answers and wrong answers are easy to obtain“ (Long 1997, editor’s introduction) when doing multivariate data analysis by utilizing one of the major statistical packages (e.g. SAS, SPSS, Stata). In order to find the right answers one has to choose – correctly – from a large number of options. How to do this with SPSS can be learned in this preparatory course. Beginning with elementary linear regression methods we will soon move on to more advanced methods of regression analysis which will be in the centre of this course. It will be shown, first, for the linear regression model how to specify an explanatory model (e.g. modeling interactions, non-linear relationships), how to make use of advanced regression diagnostics, which remedies can be applied when problems in the data and/or in the specified model are detected. Since in the social sciences true quantitative variables – required for the dependent variable in the linear regression model – are rather the exception, alternative approaches such as logit or probit regression models often have to be used. These models will be dealt with in the second part of the course. Whereas regression methods focus on the estimation of causal effects, methods of dimensional analysis (e.g. factor analysis, correspondence analysis) seek to identify structures in the data. These dimensional methods will be treated in the third part of the course. Schließen

Literaturhinweise

Agresti, Alan, 2007: An Introduction to Categorical Data Analysis. Hoboken, NJ: John Wiley & Sons.