Can generative AI infer thinking style from language? Evaluating the utility of AI as a psychological text analysis tool.

Generative AI, short for Generative Artificial Intelligence, a class of artificial intelligence systems, is not currently the choice technology for text analysis, but prior work suggests it may have some utility to assess dynamics like emotion. The current work builds upon this empirical foundation to consider how analytic thinking scores from a large language model chatbot, ChatGPT, were linked to analytic thinking scores from dictionary-based tools like Linguistic Inquiry and Word Count (LIWC). Using over 16,000 texts from four samples and tested against three prompts and two large language models (GPT-3.5, GPT-4), the evidence suggests there were small associations between ChatGPT and LIWC analytic thinking scores (meta-analytic effect sizes: .058 < rs < .304; ps < .001). When given the formula to calculate the LIWC analytic thinking index, ChatGPT performed incorrect mathematical operations in 22% of the cases, suggesting basic word and number processing may be unreliable with large language models. Researchers should be cautious when using AI for text analysis.

Political ideology shapes risk and benefit judgments of COVID-19 vaccines.

In April 2021, the use of the Johnson & Johnson COVID-19 vaccine was paused to investigate whether it had caused serious blood clots to a small number of women (six out of 6.8 million Americans who had been administered that vaccine). As these events were unfolding, we surveyed a sample of Americans (N = 625) to assess their reactions to this news, whether they supported the pausing of the vaccine, and potential psychological factors underlying their decision. In addition, we employed automated text analyses as a supporting method to more classical quantitative measures. Results showed that political ideology influenced the support for the pausing of the vaccine; liberals were more likely to oppose it than conservatives. In addition, the effect of political ideology was mediated by the difference between perceived benefit and risk and the language style used to produce reasons in support (or against) the decision to pause the vaccine. Liberals perceived the benefit of vaccines higher than the risk, used a more analytic language style when stating their reasons, and had a more positive attitude toward the vaccine. We discuss the implications of our findings considering vaccine hesitancy and risk perception during the COVID-19 pandemic.

Cross-checking journalistic fact-checkers: The role of sampling and scaling in interpreting false and misleading statements.

Professional fact-checkers and fact-checking organizations provide a critical public service. Skeptics of modern media, however, often question the accuracy and objectivity of fact-checkers. The current study assessed agreement among two independent fact-checkers, The Washington Post and PolitiFact, regarding the false and misleading statements of then President Donald J. Trump. Differences in statement selection and deceptiveness scaling were investigated. The Washington Post checked PolitiFact fact-checks 77.4% of the time (22.6% selection disagreement). Moderate agreement was observed for deceptiveness scaling. Nearly complete agreement was observed for bottom-line attributed veracity. Additional cross-checking with other sources (Snopes, FactCheck.org), original sources, and with fact-checking for the first 100 days of President Joe Biden's administration were inconsistent with potential ideology effects. Our evidence suggests fact-checking is a difficult enterprise, there is considerable variability between fact-checkers in the raw number of statements that are checked, and finally, selection and scaling account for apparent discrepancies among fact-checkers.

Contextual considerations for deception production and detection in forensic interviews.

Most deception scholars agree that deception production and deception detection effects often display mixed results across settings. For example, some liars use more emotion than truth-tellers when discussing fake opinions on abortion, but not when communicating fake distress. Similarly, verbal and nonverbal cues are often inconsistent predictors to assist in deception detection, leading to mixed accuracies and detection rates. Why are lie production and detection effects typically inconsistent? In this piece, we argue that aspects of the context are often unconsidered in how lies are produced and detected. Greater theory-building related to contextual constraints of deception are therefore required. We reintroduce and extend the Contextual Organization of Language and Deception (COLD) model, a framework that outlines how psychological dynamics, pragmatic goals, and genre conventions are aspects of the context that moderate the relationship between deception and communication behavior such as language. We extend this foundation by proposing three additional aspects of the context - individual differences, situational opportunities for deception, and interpersonal characteristics - for the COLD model that can specifically inform and potentially improve forensic interviewing. We conclude with a forward-looking perspective for deception researchers and practitioners related to the need for more theoretical explication of deception and its detection related to the context.

Instrumental goal activation increases online petition support across languages.

Research on processing fluency and instrumental goal activation suggests people often perceive complex information positively when effort in a task is valued. The current article evaluates this idea in five online petition samples (total N = 1,047,655 petitions and over 200 million words), assessing how the linguistic fluency of a petition associates with support. Consistent with prior work, petitions with lower rates of lexical fluency (fewer common words) associated with more signatures and an increased probability of petitions making a concrete change than those with higher rates of lexical fluency (more common words). Exploratory results suggest other forms of linguistic complexity also associated with petition support: petitions with more analytic writing (e.g., more formal and complex writing patterns) and with less structural fluency (less readable writing) received more signatures than those with less analytic writing and more structural fluency. Controlling for the political leaning of the petition writers as inferred by their language patterns revealed consistent effects. Crucially, the lexical fluency results were maintained across eight languages as well. Various types of linguistic complexity are therefore instrumental to get people to support online causes. Contributions to fluency theory and the psychology of giving are discussed. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Gender and ethnicity bias in medicine: a text analysis of 1.8 million critical care records.

Gender and ethnicity biases are pervasive across many societal domains including politics, employment, and medicine. Such biases will facilitate inequalities until they are revealed and mitigated at scale. To this end, over 1.8 million caregiver notes (502 million words) from a large US hospital were evaluated with natural language processing techniques in search of gender and ethnicity bias indicators. Consistent with nonlinguistic evidence of bias in medicine, physicians focused more on the emotions of women compared to men and focused more on the scientific and bodily diagnoses of men compared to women. Content patterns were relatively consistent across genders. Physicians also attended to fewer emotions for Black/African and Asian patients compared to White patients, and physicians demonstrated the greatest need to work through diagnoses for Black/African women compared to other patients. Content disparities were clearer across ethnicities, as physicians focused less on the pain of Black/African and Asian patients compared to White patients in their critical care notes. This research provides evidence of gender and ethnicity biases in medicine as communicated by physicians in the field and requires the critical examination of institutions that perpetuate bias in social systems.

Psychological trauma and emotional upheaval as revealed in academic writing: The case of COVID-19.

The current paper used a preregistered set of language dimensions to indicate how scientists psychologically managed the COVID-19 pandemic and its effects. Study 1 evaluated over 1.8 million preprints from arXiv.org and assessed how papers written during the COVID-19 pandemic reflected patterns of psychological trauma and emotional upheaval compared to those written before the pandemic. The data suggest papers written during the pandemic contained more affect and more cognitive processing terms to indicate writers working through a crisis than papers written before the pandemic. Study 2 (N = 74,744 published PLoS One papers) observed consistent emotion results, though cognitive processing patterns were inconsistent. Papers written specifically about COVID-19 contained more emotion than those not written about COVID-19. Finally, Study 3 (N = 361,189 published papers) replicated the Study 2 emotion results across more diverse journals and observed papers written during the pandemic contained a greater rate of cognitive processing terms, but a lower rate of analytic thinking, than papers written before the pandemic. These data suggest emotional upheavals are associated with psychological correlates reflected in the language of scientists at scale. Implications for psychology of language research and trauma are discussed.

Examining the Examiners: How Medical Death Investigators Describe Suicidal, Homicidal, and Accidental Death.

This study describes differences in medicolegal death investigators' written descriptions for people who died by homicide, suicide, or accident. We evaluated 17 years of death descriptions from a midsized metropolitan midwestern county in the United States to assess how death investigators psychologically respond to different manners of death (N = 10,408 cases). Automated text analyses suggest investigators describe accidental deaths with more immediacy relative to homicides, while they also described suicidal deaths in less emotional terms than homicides as well. These data suggest medicolegal death investigators have different psychological reactions to circumstances and manners of death as indicated by their professional writing. Future research may surface context-specific psychological reactions to vicarious trauma that could inform the design or personalization of workplace-coping interventions.

Why we dehumanize illegal immigrants: A US mixed-methods study.

Dehumanization is a topic of significant interest for academia and society at large. Empirical studies often have people rate the evolved nature of outgroups and prior work suggests immigrants are common victims of less-than-human treatment. Despite existing work that suggests who dehumanizes particular outgroups and who is often dehumanized, the extant literature knows less about why people dehumanize outgroups such as immigrants. The current work takes up this opportunity by examining why people dehumanize immigrants said to be illegal and how measurement format affects dehumanization ratings. Participants (N = 672) dehumanized such immigrants more if their ratings were made on a slider versus clicking images of hominids, an effect most pronounced for Republicans. Dehumanization was negatively associated with warmth toward illegal immigrants and the perceived unhappiness felt by illegal immigrants from U.S. immigration policies. Finally, most dehumanization is not entirely blatant but instead, captured by virtuous violence and affect as well, suggesting the many ways that dehumanization can manifest as predicted by theory. This work offers a mechanistic account for why people dehumanize immigrants and addresses how survey measurement artifacts (e.g., clicking on images of hominids vs. using a slider) affect dehumanization rates. We discuss how these data extend dehumanization theory and inform empirical research.

Ability-related political polarization in the COVID-19 pandemic.

In two large-scale longitudinal datasets (combined N = 5761), we investigated ability-related political polarization in responses to the COVID-19 pandemic. We observed more polarization with greater ability in emotional responses, risk perceptions, and product-purchase intentions across five waves of data collection with a diverse, convenience sample from February 2020 through July 2020 (Study 1, N = 1267). Specifically, more liberal participants had more negative emotional responses and greater risk perceptions of COVID-19 than conservative participants. Compared to conservatives, liberal participants also interpreted quantitative information as indicating higher COVID-19 risk and sought COVID-related news more from liberal than conservative news media. Of key importance, we also compared verbal and numeric cognitive abilities for their independent capacity to predict greater polarization. Although measures of numeric ability, such as objective numeracy, are often used to index ability-related polarization, ideological differences were more pronounced among those higher in verbal ability specifically. Similar results emerged in secondary analysis of risk perceptions in a nationally representative longitudinal dataset (Study 2, N = 4494; emotions and purchase intentions were not included in this dataset). We further confirmed verbal-ability-related polarization findings on non-COVID policy attitudes (i.e., weapons bans and Medicare-for-all) measured cross-sectionally. The present Study 2 documented ability-related polarization emerging over time for the first time (rather than simply measuring polarization in existing beliefs). Both studies demonstrated verbal ability measures as the most robust predictors of ability-related polarization. Together, these results suggest that polarization may be a function of the amount and/or application of verbal knowledge rather than selective application of quantitative reasoning skills.

Virtual reality and the psychology of climate change.

Researchers and practitioners have used virtual reality (VR) as a tool to understand attitudes and behaviors around climate change for decades. As VR has become more immersive, mainstream, and commercially available, it has also become a medium for education about climate issues, a way to indirectly expose users to novel stimuli, and a tool to tell stories about antienvironmental activity. This review explicates the relationship between VR and climate change from a psychological perspective and offers recommendations to make virtual experiences engaging, available, and impactful for users. Climate change is perhaps the most urgent global issue of our lifetime with irreversible consequences. It therefore requires innovative experiential approaches to teach its effects and modify attitudes in support of proenvironmental actions.

The predictive utility of word familiarity for online engagements and funding.

Metacognitive frameworks such as processing fluency often suggest people respond more favorably to simple and common language versus complex and technical language. It is easier for people to process information that is simple and nontechnical compared to complex information, therefore leading to more engagement with targets. In two studies covering 12 field samples (total n = 1,064,533), we establish and replicate this simpler-is-better phenomenon by demonstrating people engage more with nontechnical language when giving their time and attention (e.g., simple online language tends to receive more social engagements). However, people respond to complex language when giving their money (e.g., complex language within charitable giving campaigns and grant abstracts tend to receive more money). This evidence suggests people engage with the heuristic of complex language differently depending on a time or money target. These results underscore language as a lens into social and psychological processes and computational methods to measure text patterns at scale.

Dehumanization During the COVID-19 Pandemic.

Communities often unite during a crisis, though some cope by ascribing blame or stigmas to those who might be linked to distressing life events. In a preregistered two-wave survey, we evaluated the dehumanization of Asians and Asian Americans during the COVID-19 pandemic. Our first wave (March 26-April 2, 2020; N = 917) revealed dehumanization was prevalent, between 6.1% and 39% of our sample depending on measurement. Compared to non-dehumanizers, people who dehumanized also perceived the virus as less risky to human health and caused less severe consequences for infected people. They were more likely to be ideologically Conservative and believe in conspiracy theories about the virus. We largely replicated the results 1 month later in our second wave (May 6-May 13, 2020; N = 723). Together, many Americans dehumanize Asians and Asian Americans during the COVID-19 pandemic with related perceptions that the virus is less problematic. Implications and applications for dehumanization theory are discussed.

Virtuous violence from the war room to death row.

How likely is it that someone would approve of using a nuclear weapon to kill millions of enemy civilians in the hope of ending a ground war that threatens thousands of American troops? Ask them how they feel about prosecuting immigrants, banning abortion, supporting the death penalty, and protecting gun rights and you will know. This is the finding from two national surveys of Democrats and Republicans that measured support for punitive regulations and policies across these four seemingly unrelated issues, and a fifth, using nuclear weapons against enemy civilians (in survey 1) or approving of disproportionate killing with conventional weapons (in survey 2). Those who support these various policies that threaten harm to many people tend to believe that the victims are blameworthy and it is ethical to take actions or policies that might harm them. This lends support to the provocative notion of "virtuous violence" put forth by Fiske and Rai [A. P. Fiske, T. S. Rai, Virtuous Violence: Hurting and Killing to Create, Sustain, End, and Honor Social Relationships (2014)], who assert that people commit violence because they believe it is the morally right thing to do. The common thread of punitiveness underlying and connecting these issues needs to be recognized, understood, and confronted by any society that professes to value fundamental human rights and wishes to prevent important decisions from being affected by irrelevant and harmful sociocultural and political biases.

Social, psychological, and demographic characteristics of dehumanization toward immigrants.

This study extends the current body of work on dehumanization by evaluating the social, psychological, and demographic correlates of blatant disregard for immigrants. Participants (n = 468) were randomly assigned to read a scenario where 1) an immigrant or 2) an immigrant and their child were caught illegally crossing the southern border of the United States, and then rated how long they should spend in jail if convicted. Participants reported that they would sentence the immigrant to more jail time than the immigrant and child. Those who sent immigrants to jail for more time also viewed them as socially distant and less human, described immigration in impersonal terms, and endorsed other social harms unrelated to immigration (e.g., the death penalty for convicted murderers). Crucially, endorsed social harms accounted for explained variance beyond simply holding conservative views. We position these data within the current literature on dehumanization theory and immigration issues.

Psychological and physiological effects of applying self-control to the mobile phone.

This preregistered study examined the psychological and physiological consequences of exercising self-control with the mobile phone. A total of 125 participants were randomly assigned to sit in an unadorned room for six minutes and either (a) use their mobile phone, (b) sit alone with no phone, or (c) sit with their device but resist using it. Consistent with prior work, participants self-reported more concentration difficulty and more mind wandering with no device present compared to using the phone. Resisting the phone led to greater perceived concentration abilities than sitting without the device (not having external stimulation). Failing to replicate prior work, however, participants without external stimulation did not rate the experience as less enjoyable or more boring than having something to do. We also observed that skin conductance data were consistent across conditions for the first three-minutes of the experiment, after which participants who resisted the phone were less aroused than those who were without the phone. We discuss how the findings contribute to our understanding of exercising self-control with mobile media and how psychological consequences, such as increased mind wandering and focusing challenges, relate to periods of idleness or free thinking.

Immersive Virtual Reality Field Trips Facilitate Learning About Climate Change.

Across four studies, two controlled lab experiments and two field studies, we tested the efficacy of immersive Virtual Reality (VR) as an education medium for teaching the consequences of climate change, particularly ocean acidification. Over 270 participants from four different learning settings experienced an immersive underwater world designed to show the process and effects of rising sea water acidity. In all of our investigations, after experiencing immersive VR people demonstrated knowledge gains or inquisitiveness about climate science and in some cases, displayed more positive attitudes toward the environment after comparing pre- and post-test assessments. The analyses also revealed a potential post-hoc mechanism for the learning effects, as the more that people explored the spatial learning environment, the more they demonstrated a change in knowledge about ocean acidification. This work is unique by showing distinct learning gains or an interest in learning across a variety of participants (high school, college students, adults), measures (learning gain scores, tracking data about movement in the virtual world, qualitative responses from classroom teachers), and content (multiple versions varying in length and content about climate change were tested). Our findings explicate the opportunity to use immersive VR for environmental education and to drive information-seeking about important social issues such as climate change.

Linguistic traces of a scientific fraud: the case of Diederik Stapel.

When scientists report false data, does their writing style reflect their deception? In this study, we investigated the linguistic patterns of fraudulent (N  =  24; 170,008 words) and genuine publications (N  =  25; 189,705 words) first-authored by social psychologist Diederik Stapel. The analysis revealed that Stapel's fraudulent papers contained linguistic changes in science-related discourse dimensions, including more terms pertaining to methods, investigation, and certainty than his genuine papers. His writing style also matched patterns in other deceptive language, including fewer adjectives in fraudulent publications relative to genuine publications. Using differences in language dimensions we were able to classify Stapel's publications with above chance accuracy. Beyond these discourse dimensions, Stapel included fewer co-authors when reporting fake data than genuine data, although other evidentiary claims (e.g., number of references and experiments) did not differ across the two article types. This research supports recent findings that language cues vary systematically with deception, and that deception can be revealed in fraudulent scientific discourse.

An unusual cutaneous manifestation of Crohn's disease.

A 61-year-old man with a 12-year history of quiescent Crohn's disease on mesalamine presented to his gastroenterologist in April 2009, complaining of abdominal cramping, diarrhea, and a 25-lb weight loss over 6 weeks. He did not respond to prednisone 50 mg and 6-mercaptopurine 100 mg daily. Abdominal computed tomography findings revealed diffuse submucosal edema consistent with extensive colitis. Colonoscopy demonstrated diffuse inflammation with erythema, friability, and shallow ulcerations in the rectum and colon. Biopsies were consistent with Crohn's colitis. He was admitted for infliximab infusion for his unremitting diarrhea. Five days before admission, the patient noted mild swelling and redness of the left lower eyelid, which progressed to involve the right lower eyelid with frank pus draining from both eyes. He had no visual impairment or eye pain. Two days before admission, an ophthalmologist prescribed a steroid eyedrop with no relief. He also complained of seropurulent painful skin lesions on his face and scalp, which spread to involve his upper trunk and proximal arms. On admission to the hospital, dermatology, ophthalmology, and infectious disease consultations were obtained to rule out disseminated infection before initiation of infliximab therapy. The patient was afebrile and hemodynamically stable. His oral mucosa was normal. He had prominent bilateral lower eyelid edema, erythema, and superficial erosions with hemorrhagic crusting and frank green purulent drainage from both eyes, with crusting along the lower lash line and bilateral sclera injection (Figure 1). On his scalp, face, trunk, and proximal extremities, he had 25 to 30 erythematous, 4- to 8-mm papulopustules with narrow red halos, some with central necrosis and crusting (Figure 2). Cultures from the purulent ocular drainage and pustules on the trunk and arms were all negative for bacteria, virus, and fungi. Gram stain from the eye drainage showed polymorphonuclear leukocytes without organisms. Tissue cultures were negative for bacterial, fungal, and mycobacterial infection. Skin biopsy taken from the central upper back demonstrated subcorneal pustules with areas of eroded epidermis and collections of neutrophils in the superficial dermis (Figure 3). Special stains were negative for organisms. He received infliximab infusion 5 mg/kg for a total dose of 420 mg over 2 hours. Within 48 hours of infusion, there was notable decrease in size of lesions, in addition to reduction of purulent drainage from both eyes. The patient was discharged home following infliximab infusion. His skin lesions resolved during a period of 2 weeks, leaving small pink atrophic scars. He received his second infusion of infliximab 2 weeks after discharge with continued improvement in his gastrointestinal symptoms.