Recommendations for reporting regressionbased norms and the development of free-access tools to implement them in practice

Abstract

Norm-referenced tests compare individuals to a reference or source population. Norms usually depend on individual characteristics (norm-predictors) like age, gender, etc. Regression-based norming, a type of continuous norming, allows for exact evaluation of the test-taker’s score for any combination of the norm-predictors. Regression-based norms are often presented in tables and graphs in scientific papers, where only selected centiles for some combination of norm-predictors are summarized. Therefore exact score evaluation for any combination of norm-predictors is usually impossible because it requires a detailed presentation of all estimated model parameters which are usually undisclosed. Furthermore, the fitted models, like those from the R gamlss package, may include individual data that are usually protected by law and consent, which prevent data sharing. Thus, this paper provides recommendations for publishing regression-based norms that allow precise score evaluation for any combination of the norm-predictors while protecting participant privacy. We outline specific requirements for such publications: a) the exact presentation of the underlying fitted regression model that contains the estimates of all model parameters and other information required for exact evaluation; b) computer sharable fit of the model that does not contain any sensitive information and can be used by those with programming skills to evaluate scores; and c) a web-based application that can be used by those without programming skills to use the results of the fitted model. To facilitate publication and utilization of such regression-based norms, we have developed and provided an open-source R package of tools for authors and users alike. Following our recommendations, any user can access the underlying models while data privacy is maintained. This approach ensures broad accessibility and practical application of norms, allowing other researchers to accurately interpret their individual data against such norms.