Training and Events
Statistical Analysis with Missing Data
|Department of Methodology, London School of Economics|
Professor Jouni Kuha
View in Google Maps (WC2A 2HD)
Department of Methodology, London School of Economics. 02071075212. firstname.lastname@example.org.
Very many statistical analyses suffer from the problem of missing data, where not all of the data that we wanted to collect was actually observed. This raises the methodological question of how to analyse such data, to avoid the biases and minimise the loss of information which arises from the missing data.
In this workshop we consider this question. We discuss different types of missing data and assumptions about them, and review the possible ways of handling them in the analyses. The main focus of the workshop is on the specific method of multiple imputation of missing data in individual observations of variables (item nonresponse). The basic theory, assumptions and methods of multiple imputation are described in the lecture, and in the computer class we then practice carrying out multiple imputation in statistical software.
Intermediate (some prior knowledge)
Website and registration
Quantitative Data Handling and Data Analysis, Descriptive Statistics
Related publications and presentations