Open science practices in clinical psychology journals: An audit study

We conducted an audit of 60 clinical psychology journals, covering the first 2 quartiles by impact factor on Web of Science. We evaluated editorial policies in 5 domains crucial to reproducibility and transparency (prospective registration, data sharing, preprints, endorsement of reporting guidelines and conflict of interest [COI] disclosure). We examined implementation in a randomly selected cross-sectional sample of 201 articles published in 2017 in the "best practice" journals, defined as having explicit supportive policies in 4 out of 5 domains. Our findings showed that 15 journals cited prospective registration, 40 data sharing, 15 explicitly permitted preprints, 28 endorsed reporting guidelines, and 52 had mandatory policies for COI disclosure. Except for COI disclosure, few policies were mandatory: registration in 15 journals, data sharing in 1, and reporting guidelines for randomized trials in 18 and for meta-analyses in 15. Seventeen journals were identified as "best practice." An analysis of recent articles showed extremely low compliance for prospective registration (3% articles) and data sharing (2%). One preprint could be identified. Reporting guidelines were endorsed in 19% of the articles, though for most articles this domain was rated as nonapplicable. Only half of the articles included a COI disclosure. Desired open science policies should become clear and mandatory, and their enforcement streamlined by reducing the multiplicity of guidelines and templates. (PsycINFO Database Record (c) 2019 APA, all rights reserved).