17 March 2017
14:30
Louvain-la-Neuve
ISBA C115 (Seminar Room Bernouilli)
Applied statistics workshops
Practical concerns in Survey research: Weighting
Abstract:
In many practical settings, taking a (simple random) sample and then applying a statistical model on that data is often not sufficient. For instance, because of non-response some groups can be overrepresented and some other groups can be underrepresented. Another example is when a sample design has been used that requires special treatment of the data . A commonly applied correction technique is applying weighting adjustments. While this subject is well covered in statistical theory, in practice weighting is not always easy and the consequences of the different approaches are not always well understood by practitioners. In this talk I will describe how weighting is done in a market research context. I will give an overview of the different approaches that are used (‘normal’ weighting, RIM weighting, linear weighting) followed by a few remarks on software. I will discuss weighting efficiency and cover some of the pitfalls and challenges market researchers face.
In many practical settings, taking a (simple random) sample and then applying a statistical model on that data is often not sufficient. For instance, because of non-response some groups can be overrepresented and some other groups can be underrepresented. Another example is when a sample design has been used that requires special treatment of the data . A commonly applied correction technique is applying weighting adjustments. While this subject is well covered in statistical theory, in practice weighting is not always easy and the consequences of the different approaches are not always well understood by practitioners. In this talk I will describe how weighting is done in a market research context. I will give an overview of the different approaches that are used (‘normal’ weighting, RIM weighting, linear weighting) followed by a few remarks on software. I will discuss weighting efficiency and cover some of the pitfalls and challenges market researchers face.
Categories Events: