Handling Missing Data

Institution: see Organisers & Acknowledgements

Program of study: International Research Workshop

Lecturer: Martin Spieß (University of Hamburg & SOEP at the DIW Berlin)

Date: 07.10.2009, 09:00 a.m. – 17:30 p.m.

Room: n.s.

Max. number of participants: 30

Semester periods per week: n.s.

Credit Points: 3 CP for participating in the whole IRWS 2009

Language of instruction: English


Most surveys are affected by missing data. Depending on the mechanism that led to the observed pattern of missing and observed data, inferences based on the observed part of the data set using standard analysis tools may be severely biased. Thus, in the first part of the lecture, the missing mechanisms leading to data that are missing completely at random (MCAR), missing at random (MAR) or not missing at random (NMAR) will be discussed. In the second part, we will look at different methods to compensate for missing units and missing items. The emphasis in this part is on weighting to compenaste for unit nonresponse and multiple imputation to compensate for item nonresponse.

You have to register for the 3rd International Research Workshop to participate in this course.