1 / 5

Type I errors Purcell model

This study investigates the impact of correlations between variables M and T in Purcell's moderation model. When M and T are interrelated, especially through a mediator A, the standard specification can lead to a significant increase in Type I errors. Our simulation results demonstrate that increased false positives occur in various correlation scenarios. We conclude by emphasizing the importance of fitting a bivariate moderation model first, and adhering to it if moderation on the covariance between M and T is present. Otherwise, an extended univariate approach can be more powerful.

tacita
Télécharger la présentation

Type I errors Purcell model

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Type I errors Purcell model If M and T are correlated (eg via A), and M is correlated between twins as well, then the original specification of the Purcell moderation model can yield a large number of false positives Slides: S van der Sluis, VU University Amsterdam

  2. Results simulation study type I error M and T correlated via A Type I increased for Ba and Bc M and T correlated via C Type I increased for Ba and Bc M and T correlated via E Type I increased for Ba, Bc and Be Slides: S van der Sluis, VU University Amsterdam

  3. Extension means model Original model Extension means model Slides: S van der Sluis, VU University Amsterdam

  4. Unless… Unless the covariance between M and T is subject to moderation as well. Then the univariate parameterization should not be used (again increased false positive rate) Slides: S van der Sluis, VU University Amsterdam

  5. Conclusion • Always first fit the bivariate • moderation model. • 2. If moderation on covariance M and T is present: • stick to bivariate moderation model • 3. If moderation on covariance M and T is absent: • proceed with extended univariate parameterization (more powerful!) Slides: S van der Sluis, VU University Amsterdam

More Related