Swiss Summer School 2003

Eugene Horber
Making Data Work: Conceive, structure and prepare data for analysis

Eugene Horber

Eugène Horber is Professor of Applied Computer Science in the Section des Sciences Sociales, University of Geneva. He teaches social science methodology, applied computer science, and statistics. His research interests and publications are in the area of statistical methodology, aggregate data analysis, and applied computer science (didactical software, hypertext). He is the author of a software package for exploratory data analysis.

Workshop contents and objectives

The best analytical techniques will yield only non-sense, if they are applied to badly conceived and collected data. Data is not given, but produced. Preparing data for analysis is often much less trivial than it seems and requires not only - as too many seem to believe - technical skills, but above all a conceptual view and awareness of all aspects of a research process. Preparing data for analysis needs to be planned and requires many decisions that impact heavily on the analysis. Validating data and assessing data quality requires non-trivial skills... and a lot of experience.

The aim of this workshop is to give participants knowledge and skills that are required to prepare data for analysis. While many technical and practical aspects will be covered, conceptual and planning aspects will have a central place, as it is essential to understand that data collection, data management and data preparation are the key to good analysis and are an essential part of the research process, that cannot be reduced to its mere technical side that can be left to a "specialist".

The workshop will be centered around "case studies" that show the wide variety of problems, ranging from simple surveys to complex repeated surveys, hierarchical data sets (e.g. family studies where data is collected on children, parents and families etc) and or other nested structures. Although data entry problems will be treated, the main focus will be on how to conceive and prepare a dataset for a particular analysis. Most examples will be drawn from already existing data sets, namely

Besides solving practical problems, the workshop will also address conceptual problems linked to using survey collections (e.g. study the evolution of Swiss electoral behaviour using the Vox survey), combining data from different sources (e.g. combining survey data with aggregate data).

The choice of the "case studies" will however be adapted to the participant's needs and backgrounds, as they will provide the link between the technical aspects and the conceptual, research oriented ones. There will be room to work on problems brought in by participants.

Participants will acquire skills with SPSS in the following areas:

Although the workshop will be based on SPSS (or more precisely its data management and preparation facilities), the skills and knowledge acquired will be useful for all data preparation and management tasks, no matter what software you will be using.


Familiarity with SPSS or any other major statistical software package (basic analysis and transformation skills).