Suchergebnisse

We illustrate how standard psychometric inventories originally designed for assessing noncognitive human traits can be repurposed as diagnostic tools to evaluate analogous traits in large language models (LLMs). We start from the assumption that LLMs, inadvertently yet inevitably, acquire psychological traits (metaphorically speaking) from the vast text corpora on which they are trained. Such corpora contain sediments of the personalities, values, beliefs, and biases of the countless human authors of these texts, which LLMs learn through a complex training process. The traits that LLMs acquire in such a way can potentially influence their behavior, that is, their outputs in downstream tasks and applications in which they are employed, which in turn may have real-world consequences for individuals and social groups. By eliciting LLMs' responses to language-based psychometric inventories, we can bring their traits to light. Psychometric profiling enables researchers to study and compare LLMs in terms of noncognitive characteristics, thereby providing a window into the personalities, values, beliefs, and biases these models exhibit (or mimic). We discuss the history of similar ideas and outline possible psychometric approaches for LLMs. We demonstrate one promising approach, zero-shot classification, for several LLMs and psychometric inventories. We conclude by highlighting open challenges and future avenues of research for AI Psychometrics.

By Richard F. Antonak, University of New Hampshire, Durham, and Hanoch Livneh, Rhode Island College, Providence. With a Foreword by Harold E. Yuker. CONTENTS: Introduction; Fundamentals of Attitude Measurement, Direct Methods of Attitude Measurement; Multidimensional Scaling; Indirect Methods of Attitude Measurement; Reliability; Validity; Psychometric Characteristics of Attitude Scales; General Attitudes Toward People with Disabilities, Attitudes Toward Physical Disability Groups; Attitudes Toward Psychiatric Disability Groups; Attitudes Toward People Who are Mentally Retarded, Societal Respo

Os dados sobre nossas emoções, os chamados emotional data, constituem hoje uma valiosa commodity coletada e comercializada por plataformas de comunicação digital. Entre os maiores interessados em obtê-la estão corporações financeiras e políticas que, entre outros usos, baseiam suas decisões em informações sobre os afetos dos usuários das redes. Existem diferentes formas de se gerar emotional data, e uma delas é a análise de sentimentos. Este artigo aborda algumas características dessa ferramenta, investigando o seu funcionamento e os saberes psicométricos que a constituem. A análise de sentimentos é entendida não apenas como uma ferramenta de detecção de afetos, mas também de produção emocional, uma técnica que opera instrumentalizando as emoções para uma capitalização alheia ao indivíduo. É dessa maneira que é possível delineá-la — para além de um instrumento psicométrico — como um aparato psicopolítico. Neste sentido, conceitos como "sociedade de controle" (Deleuze, 1992), "sociedade confessional" (Bauman, 2012/2014), além da própria noção de "psicopolítica" (Han, 2014/2014b), são úteis para compreendermos aspectos da produção emocional assentes nas novas tecnologias da comunicação. Este artigo, portanto, pretende contribuir para o entendimento de um fator importante, mas ainda algo negligenciado nos estudos sobre big data e vigilância: o monitoramento e a produção de afetos como forma de controle subjetivo. ; The data about our affects, the so-called emotional data, constitute nowadays a valuable commodity, collected and marketed by digital communication platforms. Among the interested in obtaining it are financial and political corporations that base their decisions on information about network user's affects. There are different ways to generate emotional data, one of which is the sentiment analysis. This article addresses some characteristics of this tool, clarifying its operation and the psychometric knowledges that constitute it. Sentiment analysis is understood not only as a tool for detecting affects, but also for emotional production. It is in this sense that it is possible to outline it — beyond a psychometric instrument — as a psychopolitical apparatus, a technique that operates by instrumentalizing emotions for a capitalization beyond the individual. In this sense, concepts such as "control society" (Deleuze, 1992), "confessional society" (Bauman, 2012/2014), and the very notion of "psychopolitics" (Han, 2014/2014b), are useful to understand aspects of emotional production based on new communication technologies. This article, therefore, aims to contribute to the understanding of an important factor which is still somewhat neglected in studies on big data and surveillance: the monitoring and production of affects as a form of subjective control.

Os dados sobre nossas emoções, os chamados emotional data, constituem hoje uma valiosa commodity coletada e comercializada por plataformas de comunicação digital. Entre os maiores interessados em obtê-la estão corporações financeiras e políticas que, entre outros usos, baseiam suas decisões em informações sobre os afetos dos usuários das redes. Existem diferentes formas de se gerar emotional data, e uma delas é a análise de sentimentos. Este artigo aborda algumas características dessa ferramenta, investigando o seu funcionamento e os saberes psicométricos que a constituem. A análise de sentimentos é entendida não apenas como uma ferramenta de detecção de afetos, mas também de produção emocional, uma técnica que opera instrumentalizando as emoções para uma capitalização alheia ao indivíduo. É dessa maneira que é possível delineá-la — para além de um instrumento psicométrico — como um aparato psicopolítico. Neste sentido, conceitos como "sociedade de controle" (Deleuze, 1992), "sociedade confessional" (Bauman, 2012/2014), além da própria noção de "psicopolítica" (Han, 2014/2014b), são úteis para compreendermos aspectos da produção emocional assentes nas novas tecnologias da comunicação. Este artigo, portanto, pretende contribuir para o entendimento de um fator importante, mas ainda algo negligenciado nos estudos sobre big data e vigilância: o monitoramento e a produção de afetos como forma de controle subjetivo. ; The data about our affects, the so-called emotional data, constitute nowadays a valuable commodity, collected and marketed by digital communication platforms. Among the interested in obtaining it are financial and political corporations that base their decisions on information about network user's affects. There are different ways to generate emotional data, one of which is the sentiment analysis. This article addresses some characteristics of this tool, clarifying its operation and the psychometric knowledges that constitute it. Sentiment analysis is understood not only as a tool for detecting affects, but also for emotional production. It is in this sense that it is possible to outline it — beyond a psychometric instrument — as a psychopolitical apparatus, a technique that operates by instrumentalizing emotions for a capitalization beyond the individual. In this sense, concepts such as "control society" (Deleuze, 1992), "confessional society" (Bauman, 2012/2014), and the very notion of "psychopolitics" (Han, 2014/2014b), are useful to understand aspects of emotional production based on new communication technologies. This article, therefore, aims to contribute to the understanding of an important factor which is still somewhat neglected in studies on big data and surveillance: the monitoring and production of affects as a form of subjective control.

AbstractThe increasing interest in implementing social‐emotional learning (SEL) interventions within schools calls for more reliable and valid assessment tools to measure the effectiveness of SEL programs. We investigated the factor structure and psychometric properties of the Social Skills Improvement System – Social‐Emotional Learning Edition (SSIS SEL) with a group of ethnically and socio‐economically diverse grade 3 students in Ontario, Canada (n = 427). The SSIS SEL is an age‐normed measure used for the evaluation of interventions and prevention initiatives. Data collected included both teacher‐report and student self‐report measures. The factor structure proposed by the scale developers was not replicated using this smaller single grade sample. However, the study demonstrates evidence of good construct and convergent validity supporting the novel factor structure. Results from this study suggest that more research on the SSIS SEL's application in real‐world intervention research programs is needed.

Targeted social media advertising based on psychometric user profiling has emerged as an effective way of reaching individuals who are predisposed to accept and be persuaded by the advertising message. This article argues that in the case of political advertising, this may present a democratic and ethical challenge. Hypertargeting methods such as psychometrics can "crowd out" political communication with opposing views due to individual attention and time limitations, creating inequities in the access to information essential for voting decisions. The use of psychometrics also appears to have been used to spread both information and misinformation through social media in recent elections in the U.S. and Europe. This article is an applied ethics study of these methods in the context of democratic processes and compared to purely commercial situations. The ethical approach is based on the theoretical, contractarian work of John Rawls, which serves as a lens through which the author examines whether the rights of individuals, as Rawls attributes them, are violated by this practice. The article concludes that within a Rawlsian framework, use of psychometrics in commercial advertising on social media platforms, though not immune to criticism, is not necessarily unethical. In a democracy, however, the individual cannot abandon the consumption of political information, and since using psychometrics in political campaigning makes access to such information unequal, it violates Rawlsian ethics and should be regulated.

In den USA ist kürzlich ein Buch unter dem Titel "The Bell Curve" erschienen, das mit der Behauptung Aufsehen erregt, soziale Unterschiede, Kriminalität und Arbeitslosigkeit ließen sich durch verschiedene und niedrige Intelligenzquotienten erklären. Dabei sei Intelligenz weitgehend angeboren, und Schwarze seien nachweislich weniger intelligent als Weiße und Ostasiaten. Diese Renaissance eines psychometrisch begründeten Rassismus ist von den Autoren des Buches politisch beabsichtigt, was quellenkritisch nachgewiesen wird. In diesem Zuammenhang wird auf die Problematik interessengeleiteter finanzieller Wissenschaftsförderung aufmerksam gemacht. Die von den Autoren Herrnstein und Murray geforderten bildungspolitischen Maßnahmen werden ebenfalls geprüft. Es zeigt sich, daß eine Sozial- und Bildungspolitik, die auf fragwürdigen psychometrischen Daten aufbaut, sich wissenschaftlich nicht legitimieren läßt, und politische Interessen im Vordergrund stehen. (DIPF/Orig.) ; The US recently saw the publication of a book entitled "The Bell Curve" which made a great stir due to the authors' statement that social differences, criminality, and unemployment could be explained by differing and low intelligence quotients. Intelligence was largely inborn and blacks were demonstrably less intelligent than whites or East Asians. This renaissance of a psychometrically substantiated racism was politically calculated by the authors, as a critical study of the book's sources proves. In this context, the author draws attention to the problem of interest-guided financial sponsoring of science. Furthermore, the educational policy measures demanded by the book's authors, Herrnstein and Murray, are examined; it becomes apparent that their social and educational policy is based on questionable psychometric data, and can thus not be legitimized scientifically, and that political interests play a major role. (DIPF/Orig.)

Psychometrics is a branch of psychology concerned with the measurement of mental attributes, behavior, and performance, in addition to the design and analysis of tests and other instruments. The origins of this field are rooted in the explorations of 18th century scientists concerned with capturing phenomena in empirical ways. Less discussed is the use of tests and assessments to validate racialization, which thrusts persons racialized as Black into the early discourse on psychometrics. Scholars, scientists, and psychometricians racialized as Black have long engaged psychometrics providing two major contributions: infrastructure via personnel and training programs built by persons racialized as Black; and interdisciplinarity, which include disciplinary standards and knowledge production. This commentary names these important figures and describes their contributions to the field of psychometrics.

This study investigated the psychometric properties and aspects of the construct validity of the Rape Conformity Assessment (RCA). In addition, the test-retest reliability and internal consistency of Malamuth's Attraction to Sexual Aggression scale and Burt's Rape Myth Acceptance, Acceptance of Interpersonal Violence (AIV), and Adversarial Sexual Beliefs scales were investigated. One hundred twenty-six undergraduate males in three samples served as subjects. Findings suggest that the RCA may be useful as a disguised measure that addresses the social desirability problem in rape prevention and treatment outcome research. With the exception of the AI-V, the psychometric properties of the administered scales were adequate. The conformity manipulation revealed strong effects, which suggests that the potential to rape should be viewed not only as a trait but also as having situational components.

The Structured Trauma-Related Experiences and Symptoms Screener (STRESS) is a self-report instrument for youth of age 7–18 that inventories 25 adverse childhood experiences and potentially traumatic events and assesses symptoms of post-traumatic stress disorder using the revised criteria published in the Diagnostic and Statistical Manual for Mental Disorders, fifth edition (DSM-5). The STRESS can be administered by computer such that questions are read aloud and automatic scoring and feedback are provided. Data were collected on a sample of 229 children and adolescents of age 7–17 undergoing a forensic child abuse and neglect evaluation. The purpose of the current study was to examine preliminary psychometric characteristics of the computer-administered STRESS as well as its underlying factor structure in relation to the four-factor DSM-5 model. Results provide initial support for the use of the STRESS in assessing adverse and potentially traumatic experiences and traumatic stress in children and adolescents.