Rxivist logo

Discernment of Mediator and Outcome Measurement in the PACE trial

By Ewan Carr, Silia Vitoratou, Trudie Chalder, Kimberley A Goldsmith

Posted 28 Jan 2021
medRxiv DOI: 10.1101/2021.01.25.21250436

Background: When measuring latent traits, such as those used in psychology and psychiatry, it can be unclear whether the instruments used are measuring different concepts. This issue is particularly important in context of mediation analysis, since for a sound mediation hypothesis the mediator and outcome should be distinct. We sought to assess the extent of measurement overlap between mediators and outcomes in the PACE trial (n=640). Methods: Potential measurement overlap was assessed using generalised linear latent variable models where confirmatory factor models quantified the extent to which the addition of cross-loading items resulted in significant improvements in model fit. Results: Out of 26 mediator-outcome pairs considered, only six showed evidence of cross-loading items, supporting the suggestion that mediator and outcome constructs in the PACE trial were conceptually distinct. Conclusions: This study highlights the importance of assessing measurement overlap in mediation analyses with latent traits to ensure mediator and outcome instruments are distinct.

Download data

  • Downloaded 221 times
  • Download rankings, all-time:
    • Site-wide: 142,645
    • In psychiatry and clinical psychology: 800
  • Year to date:
    • Site-wide: 33,425
  • Since beginning of last month:
    • Site-wide: 79,736

Altmetric data


Downloads over time

Distribution of downloads per paper, site-wide


PanLingua

News