We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.AP

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Applications

Title: Specification analysis for technology use and teenager well-being: statistical validity and a Bayesian proposal

Abstract: A key issue in science is assessing robustness to data analysis choices, while avoiding selective reporting and providing valid inference. Specification Curve Analysis is a tool intended to prevent selective reporting. Alas, when used for inference it can create severe biases and false positives, due to wrongly adjusting for covariates, and mask important treatment effect heterogeneity. As our motivating application, it led an influential study to conclude there is no relevant association between technology use and teenager mental well-being. We discuss these issues and propose a strategy for valid inference. Bayesian Specification Curve Analysis (BSCA) uses Bayesian Model Averaging to incorporate covariates and heterogeneous effects across treatments, outcomes and sub-populations. BSCA gives significantly different insights into teenager well-being, revealing that the association with technology differs by device, gender and who assesses well-being (teenagers or their parents).
Subjects: Applications (stat.AP)
MSC classes: 62P15, 62P25
ACM classes: G.3
Cite as: arXiv:2201.05381 [stat.AP]
  (or arXiv:2201.05381v2 [stat.AP] for this version)

Submission history

From: David Rossell [view email]
[v1] Fri, 14 Jan 2022 10:33:22 GMT (460kb,D)
[v2] Fri, 6 May 2022 10:04:59 GMT (465kb,D)

Link back to: arXiv, form interface, contact.