New HESA research asks ‘should we weight’?
Following an assessment by HESA researchers working in consultation with the Office for National Statistics (ONS), the decision was made not to apply weighting to data and statistics from the 2017/18 Graduate Outcomes survey.
Weighting is used in the publication of survey results where a sample may not be representative. At the end of year one of the survey (2017/18 graduates), HESA assessed the final sample to examine whether weighting could reduce the consequences of non-response. High-level conclusions were published in the Graduate Outcomes methodology statement published on 30 March.
Researchers tested various weighting methodologies but found little difference between the weighted and unweighted estimates for the proportion of graduates in employment and/or study. Exceptions to this were found in instances where sample sizes were small and estimates were subject to greater variability. For the most part, the application of weighting led to a marginal rise in standard errors and a slight fall in the precision of the estimates.
HESA researchers have now published a detailed technical account of how they reached the decision not to apply weighting. The paper describes the research methodology and illustrates the results that were found from the analysis. The paper is mainly aimed at academics, statisticians, and other interested parties wishing to understand the weighting research and its conclusions.
Comments or questions on this study are welcome via [email protected].
For more information about statistical outputs from the survey see HE Graduate Outcomes Data