Introduction to Survey Weighting in Market Research

Ensure your survey insights are valid and represent your researched population

This course is now delivered online as a series of live sessions which fully meet the learning objectives of face-to-face training.

Course overview/description

Good survey design will ensure that the results of your research and insights arising from it are valid and bias-free, and that the people being surveyed represents the population.  That said, it is almost impossible to ensure that samples mirror population in terms of composition.  This means that when coming to analyse the data, some respondents would need to be given more weight than others.  Straightforward in principle, weighting is an area where the more one knows, the more questions are asked and the more challenging it gets.  We explain terms such as “Rim Weighting”, “Design Effect”.

This practical course, with worked examples provides a gentle introduction as to how the right weighting schemes may be achieved, which methods and principles to apply, what the common pitfalls are and how they may be avoided and what the knock-on effects are. Indeed, it even discusses whether weighting is appropriate in the first place.  Survey weighting is critical to get right and if not applied correctly, it can do work against the efforts carried out by researchers in ensuring results are accurate and insights are valid. 

Learning outcomes/learning objectives

• Evaluate whether weighting is appropriate a specific survey and determine an optimal scheme;
• Know that weighting has an effect on survey precision (ie the “Design Effect”);
• Appreciate the difference between “Rim” and “Cell” weighting;
• Understand how to build a pragmatic weighting scheme;
• Be aware of the pitfalls and risks of weighting;

Who will benefit?

Quantitative researchers at all grades / levels of experience who carry out large scale complex surveys where a high degree of measurement accuracy is needed.  This will be of particular interest to public-sector researchers, but would also be relevant to those carrying out general purpose MR work in FMCG, B2B and media / audience measurement.  Although limited in terms of mathematical content, a certain amount of comfort with basic calculations would be needed as a pre-requisite.  

Learning method

A certain amount of theory, but mostly worked examples / exercises, case studies and discussions.

Course delivery mode

Online