Scientist as Subject: How Researcher Behaviors Influence Psychological Knowledge

Abstract

Interacting with the published literature ("knowledge consumption") and publishing new scientific findings ("knowledge production") are two key moments in the scientist's search for truth, and bias in either of these can distort what is known about an area of research. This dissertation details three studies conducted on researchers in psychology that together provide evidence of scientists' behaviors influencing these key moments of knowledge production and knowledge consumption. Methods:Psychologists were recruited to participate in each study (N = 215 and N = 587).Studies used custom web tools and social network methods to collect unique datasets on psychologists' social networks and how they approach the scientific literature. The analytic approach differed based on each study. For studies on knowledge consumption, Gini coefficients and measures of unpredictability were calculated to better understand the dynamics of the published literature. For studies on knowledge production, the generalized network scale up method was used to estimate the size of the population of iii current users of questionable research practices, and regression was used to better understand the relationship between attitudes and stigma against certain psychologists. Results:Chapter 1: Overview and IntroductionHistorians don't want to write a history of historians. They are quite happy to plunge endlessly into limitless historical detail. But they themselves don't want to be counted as part of the limitless historical detail. They don't want to be part of the historical order. It's as if doctors didn't want to fall ill and die. Charles Péguy, L'Argent, suiteFrench sociologist Pierre Bourdieu starts his 1988 book, Homo Academicus, with the above quote from poet and essayist Charles Péguy, commenting on the preference of the researcher to stand apart from what they research. Bourdieu's theoretical investigation of the academic world at a time of unrest and change (the 1968 University of Paris protests) presents to the reader the academy as an object for study. In doing so, Bourdieu brings the researcher into full display and argues that the authority and objectivity central to an academic's success is not inherent to the individual, but the result of the academic's position in the power structures of academia (Bourdieu, 1988).In writing his book, Bourdieu aimed to "exoticize the domestic", asking scholars to critically engage with the academic world they inhabit and to question what drives their research questions, methods, and conclusions. Was objective curiosity the sole driver of inquiry, or was research shaped by the influences of academic power and conformity? In order to promote radical change in academic standards and research, Bourdieu asked his peers to honestly reflect on their position of power, the production and consumption of knowledge, and their role in the validation of that knowledge (Forte, 2015).

Credibility of Preprints: An interdisciplinary Survey of Researchers

“…Peer-review likely functions as this type of heuristic cue, and so displaying information such as download or view counts, endorsements, or community comments on preprints could serve a similar purpose. Previous work by Fox (2019) found that displaying download count information about papers alongside abstracts altered download behaviors. Though downloads are not a direct measure of credibility, they are moderately correlated with citations (Yan & Gerstein, 2011;Priem, Piwowar, & Hemminger, 2012) and past research has used them as in indication of user trust and satisfaction (Nicholas et al, 2008;Tenopir et al, 2014).…”

Credibility of preprints: an interdisciplinary survey of researchers

“…Peer-review may function as this type of heuristic cue, and so displaying information such as download or view counts, endorsements, or community comments on preprints could serve a similar purpose. Previous work found that displaying download count information about papers alongside abstracts can alter download behaviours [ 25 ]. Though downloads are not a direct measure of credibility, they are moderately correlated with citations [ 26 , 27 ] and past research has used them as an indication of user trust and satisfaction [ 28 , 29 ].…”