The role of norms in legitimising and preventing questionable research practices

In recent years, there has been an increasing interest in the methods, practices, conventions
and norms that inform how scientific knowledge is produced, especially in the social and
behavioural sciences. Most recently, the field of metascience has focused on questionable
research practices (QRPs), which are suboptimal research practices located between sound
science and outright misconduct. P-hacking, inappropriate authorship and salami-slicing are
examples. The causes of QRPs remain unclear beyond the oft-rehearsed reasons of pressure
to publish and secure extramural funding. The importance of elucidating the causal and
preventative factors should not be understated. QRPs are considered to be causal factors in
the replication crisis (Begley & Ioannidis, 2015), while their high prevalence suggests they
are more threatening to the quality of science than outright misconduct, which is rare by
comparison. Thus, this project seeks to contribute to a growing literature examining the
factors underlying QRPs, with the broad aim of increasing the veracity and trustworthiness
of social scientific knowledge.
The proposed project asks how research norms influence research integrity. It seeks to
examine the normative frameworks of appropriate research practice, primarily in the social
sciences, and asks how research norms influence attitudes towards research and research
behaviours. Simply put, the project examines the extent to which QRPs are legitimated by
norms and expectations of research communities. The project draws upon the fields of
metascience, the sociology of science, and research on norms and normative change from
social psychology and behavioural economics.
The project adopts a novel approach, using some preliminary focus groups, a large-n survey
and a small number of survey-based experiments. A scientist's normative subscriptions to
discrete research practices are knowable through measuring whether engagement in those
practices elicits disapproval. Finding extensive disapproval for particular QRPs within an
epistemic community would suggest that the QRPs are normatively undesirable. Thus, the
project adopts a large-n survey of European social scientists, seeking to measure, inter alia,
the extent of disapproval with a set of research norms and practices. The sample will be
recruited from an enhanced version of Web of Science (WoS), maintained at Leiden
University. Analysis of data will encompass descriptive statistics, ANOVA, regression, factor
analysis and other latent variable techniques, using R. The set of norms and practices
utilised will be informed by preliminary focus groups and an extensive review of metascience
literature exploring QRPs. Finally, to address whether a scientist's perceptions of research
norms influence his or her own attitude towards QRPs, one to three randomised surveybased
experiments will be conducted. An indicative example will utilise a between-groups,
two-condition design, using a sample of participants from the prior survey. In the
experimental condition, participants will be informed of the distribution of measured
research norms. In the control condition this information will be withheld. Through
comparison with the baseline data acquired from the prior survey, it can be identified
whether exposure to the true norm changes a scientist's attitude towards that QRP. Three
papers will be produced and submitted for publication.
The proposed project will produce actionable knowledge that can aid in the enhancement of
research integrity. The output will be utilised to inform guidelines, tools and
recommendations, such as those produced by the 'Standard Operating Procedures for
Research Integrity' project, of which the University of Essex is a collaborator. In turn, this will
support research funders and research performing institutions in developing guidelines and
policies for improving research integrity.