Search
Program Calendar
Browse By Day
Browse By Time
Browse By Person
Browse By Room
Browse By Committee or SIG
Browse By Session Type
Browse By Keywords
Browse By Geographic Descriptor
Partner Organizations
Search Tips
Personal Schedule
Sign In
Group Submission Type: Refereed Round-Table Session
Rarely are samples perfectly representative of the population they are meant to represent. Rarer still are students sampled using simple random sample (SRS). However, many researchers in our professional community are still analyzing multistage sampled data as though the students were sampled using simple random sample. If researchers do not calculate and applying the sampling weights, the estimates will be bias they will make incorrect inferences about the population. If researcher do not account for the sample methodology, they will be over confident with their incorrect estimates and make erroneous conclusions.
This panel aims to explain why weights are useful and why the sample methodology needs to be accounted for in any inferential analysis. The first panel presentation offers guidance on the basics of weighting, including step-by-step guidance on how to compute weights for a two-stage stratified cluster sample (one of the most common sampling designs in education sector). The second panel presentation uses illustrative examples with real data to highlight the dangers of presenting estimates that are not properly weighted and when the sampling methodology is not taken into account. The third panel presentation explores the pros, cons, and implications of analyzing assessment data using weights derived from student attendance data rather than student enrollment data. Together, these three presentations will provide attendees with practical information that can be immediately applied to their sampling and analysis work.
Fundamentals of survey weight computation - Michel Rousseau, Universite du Quebec a Trois-Rivieres
The dangers of analyzing evaluation data without accounting for the sampling methodology - Chris Cummiskey, RTI International
Exploring how estimates differ when using student attendance versus student enrollment as the basis for your sampling weights - Peter Cooper, School-to-School International