The Writing Style of Predatory Publishers

In 2010, librarian Jeffrey Beall started a list of journals that allegedly use predatory practices to recruit manuscripts for publication. Coined “Beall’s List,” this working catalogue highlights over two hundred open-access journals that may feign editorial processes, peerreview, or other procedures of a reputable publisher. Given the recent attention to scientific misconduct, an important question is whether there are methods to detect predatory publishers from authentic ones. In this study, we apply an automated language analysis technique from the social sciences to examine how predatory and authentic journals differ in their writing style in the About Us and Aim/Scope sections of their websites. Compared to authentic journals, predatory journals use more discrepancy terms (e.g., “should” “would”) and positive emotions (e.g., “exciting”) but fewer function words, including articles (e.g., “a” “the”), and prepositions (e.g., “before” “in”), quantifiers (e.g., “more” “less”), and terms related to causality (e.g., “therefore”). These results follow recent patterns in the deception literature, suggesting that language features may be useful when evaluating predatory versus authentic publishers. In addition to analyzing writing style, we examined meta-linguistic properties of predatory publishers (i.e., editorial statistics, website features, and contact information) from the same database of journals. Compared to authentic publishers, predatory publishers use more third-party email addresses, claim false impact factors, fake rapid peer review, and simulate academic expertise. This is the first study to investigate predatory publishing through an empirical social science lens and our results suggest that there are quantifiable linguistic and meta-linguistic indicators that can, to some degree, distinguish between predatory publishers and those journals that seek to publish honestly. Introduction In recent years, the rise of scientific misconduct has drawn attention to the “publish or perish” mentality consuming academia, which highlights a drive for researchers to publish early and often in their career. The pressure to publish regularly can tamper with the quality of research and the moral compass of academics across disciplines. When misconducts are discovered (i.e., data fabrication or falsification), publications are typically withdrawn from the journal via retractions. In the past decade, the number of retraction notices has increased tenfold and the problem is steadily becoming worse. Until recently, little academic work has examined scientific misconduct. High-profile cases of fraud in Social Psychology have popularized the issue, but the majority of the empirical literature has surveyed the prevalence of misconduct to understand its scope. Another form of deception in science has been overlooked, however: predatory publishing. This describes a phenomenon of open-access publishers feigning editorial processes, peer-review, or other procedures of a reputable publisher in order to attract high publication fees. Predatory publishing impacts the science enterprise because it questions whether academics can distinguish between real and fake research. Furthermore, if scientists base their own research on publications that have not been thoroughly vetted, this may lead to a significant decrease in the quality of research in circulation. P ge 24259.3 Predatory journals act as a revolving door for manuscripts and academics who want to publish quickly and effortlessly. Many predatory journals appear genuine and the journal title may match an authentic publisher. For example, the Journal of Cloud Computing is a publication by SpringerOpen (authentic) and IBIMA Publishing (predatory). This duplicity can mislead researchers and becomes problematic when scaled to larger domains of science. Given the difficulty academics may experience in distinguishing predatory from authentic journals, an important empirical question asks if there are differences in how these journals market themselves to academics. By using automated linguistic-analysis techniques from the social sciences we provide the first study measuring how predatory journals describe themselves differently on their websites relative to authentic publishers. To complete our investigation, we also incorporate a meta-linguistic analysis to understand how certain predatory traits are reflected online. Predatory Publishing Explained Jeffrey Beall, a librarian and Associate Professor at the University of Colorado Denver, first documented an issue with academic publishing in 2010 and created “Beall’s List.” This working document highlights over two hundred journals that fail to meet standards akin to authentic, established publishers. He uncovered that predatory publishers typically have six defining features. According to Beall’s taxonomy, predatory publishers: 1. “Publish papers already published in other venues/outlets without providing appropriate credits. 2. Use language claiming to be a ‘leading publisher’ even though the publisher may only be a startup or a novice organization. 3. Operate in a Western country chiefly for the purpose of functioning as a vanity press for scholars in a developing country. 4. Do minimal or no copyediting. 5. Publish papers that are not academic at all, e.g. essays by laypeople or obvious pseudoscience. 6. Have a ‘contact us’ page that only includes a web form, and the publisher hides or does not reveal its location.” Essentially, predatory publishers deceive academics by faking the practices and policies of top journals (e.g., peer review, editorial boards, impact factors). In doing so, predatory journals have the potential to degrade the quality of research in circulation and they reduce the likelihood of scholars publishing in reputable and prestigious journals. With scientific misconduct on the rise, it is important to understand how predatory publishers may contribute to concerns about integrity in science. In this study, we use publically available data (i.e., information on a publisher’s website) to compare language differences between predatory and authentic journal text from the About Us and Aim/Scope sections. Do predatory journals write and describe themselves differently than authentic journals? In order to answer this question, we first address relevant work in the deception field to examine the role that language plays in understanding truthful and dishonest discourse patterns. Deception and Writing Style Prior to 2000, the majority of deception research sought to uncover nonverbal cues, such as fidgeting, gaze aversion, or particular hand-movements, that may reveal deception in face-toP ge 24259.4 face interactions. Contrary to popular belief, there are no universal cues that reliably indicate deception, and the search for “Pinocchio’s nose” has met with little success. This realization has led many scholars to study deception from another angle: language and word patterns. With modern advances in computer science, the method of using computerized text-analysis to differentiate false from truthful speech has been widely applied across a range of contexts. In a seminal work from Newman, Pennebaker, Berry, and Richards (2003), automated linguistic analysis uncovered differences between false and truthful speech on abortion attitudes, attitudes towards friends, and a mock crime. Their work started a trend of researchers using computers to analyze writing style, particularly because of this method’s objectivity and the resources that were saved by not employing humans to judge word types. Writing style differences have been an interest for deception researchers in a variety of areas. Deceptive and truthful utterances have been compared across Computer-Mediated Communication technologies, statements made by American presidents, online dating profiles, and user-generated content including hotel reviews. These studies have revealed that psychological dynamics associated with deception can be revealed in language. For example, compared to honest language, deceptive language tends to feature an increase of negative emotion terms (e.g., hate, aggression, hurt) as a reflection of the guilty or anxiety associated with lying. Many of our social relationships are built on honesty, and deception has the potential to jeopardize interpersonal trust. Therefore, it becomes distressing to lie to a friend or colleague, and an increase of negative emotions reflects this apprehension. Deceptive speech also tends to feature fewer first-person singular pronouns (e.g., I, me) as a mechanism of psychological distancing. Liars typically distance themselves from deceit by using fewer “Iwords” and increasing the number of social references (e.g., he, she, they) in order to deflect attention from the self. It is personally and socially damaging to be called a liar; therefore, removing the self from a lie is ideal for face-saving and relationship-saving strategies. Finally, lies are usually less detailed than truths because it is difficult to give specifics about information that is fabricated. The number of quantifiers (e.g., more, less) is one example of language specificity, and lies often contain fewer quantifiers than truths. The Influence of Context on Deceptive Writing Style Despite the aforementioned findings, deception patterns are not identical across research domains. As circumstances change, the psychological dynamics and associated patterns of language use should change as well. That is, deception cues are not universal because of important shifts in context along at least three proposed dimensions: psychological dynamics, goals, and genre conventions. In order to successfully understand how people lie in different scenarios, the psychological impact of the deception (i.e., emotional and cognitive involvement) must be considered. An understanding of the pragmatic goals for engaging in deception may reveal the purpose of the lie and ex