Doing reliable research in comparative psychology: Challenges and proposals for improvement.

Abstract

Unlike some other areas of psychology that have experienced a 'reproducibility crisis', the extent to which research findings in comparative psychology are reliable is only just beginning to come under the spotlight. I outline what is known about where we as a field stand in terms of the reliability of our findings, and highlight some characteristic features of our research that give may cause for concern, focusing primarily on experimental comparative cognition. I then discuss ways that we as individual researchers and a wider community can take steps to improve our current practices (and in some cases already are), as well as highlighting the crucial role institutions and gatekeepers have to play in effecting change. By tackling potential issues head on, the field of comparative psychology can have more confidence that our research findings and the resultant claims we make about animal behaviour and cognition are reliable.

Pitfalls in research on ecological validity of novel executive function tests: A systematic review and a call to action.

“…The remaining articles (n = 13) did link EV with a test’s ability to predict functional outcomes; however, these remaining studies overwhelmingly relied on samples that were too small for the results to be replicable (for discussion of replicability, see Flora, 2020; Tackett & Miller, 2019; Tecwyn, 2021). Indeed, 69% ran all their analyses on fewer than 30 participants.…”

Editorial.

“…For a number of years, I have advocated for preregistered, prereviewed articles in our scholarly journals (Beran, 2018(Beran, , 2020, as have others (e.g., Farrar et al, 2020Farrar et al, , 2021Stevens, 2017;Vonk & Krause, 2018). I view preregistered articles as an important new tool to add to our more traditional submission types to help our field deal with the rising concern about failed replication (e.g., Tecwyn, 2021), the difficulty in recreating analytic plans, and so forth. I will strongly encourage and support authors in submitting preregistration articles for review, with in-principle acceptance of good proposed work, independent of how the results turn out when those strong studies are actually conducted.…”

Visual perception of emotion cues in dogs: a critical review of methodologies

“…The problem of highly controlled and “aseptic” laboratory studies is that they might find effects that are of little relevance in the “real world” where many more variables are interacting with the experimental variables of interest. This can lead to problems of replicability, which are a concern not only in this area, but in the wider area of psychology (Farrar et al 2021 ; Open Science Collaboration 2015 ; Tecwyn 2021 ). Furthermore, it also means research is focused on what we can measure in the laboratory rather than what might be ecologically important.…”