When size matters: advantages of weighted effect coding in observational studies

2

We are happy to host this guest blog post introducing two recently published Hints & Kinks* papers, written by the authors of the most cited IJPH article ‘The reliability of a two-item scale: Pearson, Cronbach, or Spearman-Brown?’

manfred

These two new papers are both availalbe to download for free and are introduced by Manfred te Grotenhuis. Manfred is an associate professor of quantitative data analysis at Radboud University. He does research and teaching in statistics, especially regression analysis. More information about this author can be found on this page.

 

 

 

When size matters:  advantages of weighted effect coding in observational studies

 

balance2

 

If your regression model contains a categorical predictor variable, you commonly test the significance of its categories against a preselected reference category. If all categories have (roughly) the same number of observations, you can also test all categories against the grand mean using effect (ANOVA) coding. In observational studies, however, the number of observations per category typically varies. This new Hints & Kinks shows how all categories can be tested against the sample mean. The paper explains the procedure, called weighted effect coding, using R, SPSS, and Stata on the accompanying website.
A follow-up Hints & Kinks expands the procedure to regression models that test interaction effects. The authors show that, within this framework, the weighted effect coded interaction displays the extra effect on top of the main effect found in a model without the interaction effect. This offers a promising new route to estimate interaction effects in observational data, where different category sizes often prevail. Software (R, SPSS, and Stata) to run these novel regression models is also available at the authors’ website.

All software + data can be found at: www.ru.nl/sociology/mt/wec/downloads.

You can also visit the project’s page on ResearchGate

UPDATE 16.11.16:  Weighted effect coded variables interacting with interval/ratio variables are now added to the R package ‘wec‘ and available in SPSS syntax on the accompanying website

UPDATE 08.11.16: Watch a video of the presentation of Manfred te Grotenhuis on weighted effect coding, held at the Swedish Institute for Social Research (SOFI), Stockholm University, November 4, 2016.

 

 

 

 

* Hints and Kinks are short methodological reports (1000 words max., no abstract) presenting topics relevant in survey research and surveillance. They report on experiences with techniques in a variety of areas and topics, such as writing questions, questionnaire design, survey implementation, or new and original ways to show results.

View the latest posts on the International Journal of Public Health homepage

2 Comments

By commenting, you’re agreeing to follow our community guidelines.

Your email address will not be published.