Examining Shady Paths to Research Success: Achievement Goals and the Engagement in Questionable Research Practices

• In Event: Motivation and Emotions of University Faculty: Predicting Teaching and Research Performance, Stress, and Well-Being

Abstract

Questionable research practices (QRPs) and their impact on the reliability of published results are a strongly debated topic in the scientific community (John et al., 2012). QRPs are strategies that violate scientific norms and aim on increasing the chance to publish on the possible cost of scientific accuracy (e.g. rounding down a p-value of .054 to p < .05). But what motivates scientists to engage in this malpractice? The easiest explanation would be the “rotten apple” hypothesis, assigning the blame for QRPs to dark aspects of researchers’ personality represented in the dark triad (psychopathy, Machiavellianism, and narcissism; Paulhus et al., 2002). However, an alternative is that engagement in QRPs can be explained by researchers’ achievement goals, that is, the way in which researchers strive for competence in the scientific domain (see Daumiller et al., 2016). We hypothesize that the strength of appearance approach goals (i.e., striving to appear competent in the scientific domain) positively influences engagement in QRPs. This seems likely because academic admiration depends on scientific output, and thus on the ability to produce statistically significant results. Furthermore, we expected that engagement in QRPs would be even stronger associated to appearance avoidance goals (i.e., striving to hide a potential lack of scientific competencies). In contrast, we assumed that learning goals would reduce engagement in QRPs regardless of whether learning goals are characterized by an approach (i.e., striving to enhance ones’ scientific competencies) or an avoidance goal valence (i.e., striving to avoid missing chances to enhance ones’ scientific competencies). This is because the endorsement of learning goals contradicts engagement in QRPs, which reduce the possibility to truly learn from acquired data.

We assessed achievement goals and the frequency of personal engagement in QRPs (measured in line with John et al., 2012) in a preregistered cross-sectional study. Data from 217 German researchers in the field of Psychology was analyzed using structural equation modelling. As predicted, we found positive effects of appearance approach goals on QRPs and negative effects of learning approach goals on QRPs. Contrary to our expectations, we found none of the predicted effects of either learning or appearance avoidance goals. Nevertheless, the direct effects of approach goals were robust for the introduction of the dark triad into the model (R²total = .30, ΔR²goals = .18; see figure 1 for the full model). Finally, exploratory regression analyses revealed a statistically significant moderation; that is, the more researchers favored publishing over scientific rigor, the stronger the association between appearance approach goals and the engagement in QRPs (see figure 2).

Our results provide first evidence that engagement in QRPs is at least partly due to achievement goal striving of researchers. This influence of personal motivation seems to be stronger than the direct influence of dark personality traits, which highlights the critical importance of the paths that researchers choose to strive for competence in science. It might be possible to reduce engagement in QRPs by fostering learning goals or by changing the incentive system in publishing (i.e. favoring scientific accuracy over p-values).

Authors

• Stefan Janke, University of Mannheim

• Martin Daumiller, University of Augsburg

• Selma Carolin Rudert, University of Basel