A Personal Model of Trumpery: Linguistic Deception Detection in a Real-World High-Stakes Setting

2021 ◽  
pp. 095679762110159
Author(s):  
Sophie Van Der Zee ◽  
Ronald Poppe ◽  
Alice Havrileck ◽  
Aurélien Baillon
Keyword(s):  
Real World ◽  
Language Use ◽  
Deception Detection ◽  
Quantitative Model ◽  
Added Value ◽  
High Stakes ◽  
Linguistic Differences ◽  
Out Of Sample ◽  
Out Of Sample Prediction ◽  
Linguistic Models

Language use differs between truthful and deceptive statements, but not all differences are consistent across people and contexts, complicating the identification of deceit in individuals. By relying on fact-checked tweets, we showed in three studies (Study 1: 469 tweets; Study 2: 484 tweets; Study 3: 24 models) how well personalized linguistic deception detection performs by developing the first deception model tailored to an individual: the 45th U.S. president. First, we found substantial linguistic differences between factually correct and factually incorrect tweets. We developed a quantitative model and achieved 73% overall accuracy. Second, we tested out-of-sample prediction and achieved 74% overall accuracy. Third, we compared our personalized model with linguistic models previously reported in the literature. Our model outperformed existing models by 5 percentage points, demonstrating the added value of personalized linguistic analysis in real-world settings. Our results indicate that factually incorrect tweets by the U.S. president are not random mistakes of the sender.

An empirical evaluation of interpersonal deception theory in a real-world, high-stakes environment

2020 ◽  
Vol 10 (3) ◽  
pp. 185-199
Author(s):  
Joseph York Thomas ◽  
David P. Biros
Keyword(s):  
Real World ◽  
Deception Detection ◽  
Empirical Evaluation ◽  
Future Research ◽  
High Stakes ◽  
Capture Process ◽  
Content Type ◽  
Interpersonal Deception Theory ◽  
Deceptive Behavior ◽  
Speech Cues

Purpose The study of deception and the theories, which have been developed have relied heavily on laboratory experiments in controlled environments, using American college students participating in mock scenarios. The purpose of this paper is to validate previous deception detection research in a real-world, high stakes environment where the unit of analysis is the question–response pair. Design/methodology/approach The study used previously confirmed linguistic and paralinguistic speech cues and the constructs of deception in an attempt to validate a leading deception theory, interpersonal deception theory (IDT). A combination of descriptive and predictive analysis was conducted to best understand the relationship between speech cues and changes in the subjects’ behavior. Findings The result validates IDT with mixed results on individual measures and their constructs. However, there is clear evidence across the 711 question-response pairs that not only was it possible to differentiate truth from deceptive behavior but also patterns of behavior can be seen over time. Research limitations/implications Because of the real-world nature of the study, it is difficult to generalize the results to a larger population. However, one implication for future research is the development of methods to capture, process and prepare raw speech into data ready for analysis. Originality/value This paper attempts to fill the gap between the controlled mock scenarios and the harsh reality of real-world deception.

scholarly journals Assessing social preferences in reimbursement negotiations for new Pharmaceuticals in Oncology: an experimental design to analyse willingness to pay and willingness to accept

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Dominik J. Wettstein ◽  
Stefan Boes
Keyword(s):  
Willingness To Pay ◽  
Real World ◽  
Cognitive Biases ◽  
Social Preferences ◽  
Standard Of Care ◽  
Added Value ◽  
Amazon Mechanical Turk ◽  
Patient Organisation ◽  
Patient Benefit

Abstract Background Price negotiations for specialty pharmaceuticals take place in a complex market setting. The determination of the added value of new treatments and the related societal willingness to pay are of increasing importance in policy reform debates. From a behavioural economics perspective, potential cognitive biases and other-regarding concerns affecting outcomes of reimbursement negotiations are of interest. An experimental setting to investigate social preferences in reimbursement negotiations for novel, oncology pharmaceuticals was used. Of interest were differences in social preferences caused by incremental changes of the patient outcome. Methods An online experiment was conducted in two separate runs (n = 202, n = 404) on the Amazon Mechanical Turk (MTurk) platform. Populations were split into two (run one) and four (run two) equally sized treatment groups for hypothetical reimbursement decisions. Participants were randomly assigned to the role of a public price regulator for pharmaceuticals (buyer) or a representative of a pharmaceutical company (seller). In run two, role groups were further split into two different price magnitude framings (“real world” vs unconverted “real payoff” prices). Decisions had real monetary effects on other participants (in the role of premium payers or investors) and via charitable donations to a patient organisation (patient benefit). Results 56 (run one) and 59 (run two) percent of participants stated strictly monotone preferences for incremental patient benefit. The mean incremental cost-effectiveness ratio (ICER) against standard of care (SoC) was higher than the initial ICER of the SoC against no care. Regulators stated lower reservation prices in the “real world” prices group compared to their colleagues in the unconverted payoff group. No price group showed any reluctance to trade. Overall, regulators rated the relevance of the patient for their decision higher and the relevance of their own role lower compared to sellers. Conclusions The price magnitude of current oncology treatments affects stated preferences for incremental survival, and assigned responsibilities lead to different opinions on the relevance of affected stakeholders. The design is useful to further assess effects of reimbursement negotiations on societal outcomes like affordability (cost) or availability (access) of new pharmaceuticals and test behavioural policy interventions.

scholarly journals Real-world evidence: perspectives on challenges, value, and alignment of regulatory and national health technology assessment data collection requirements

2021 ◽  
Vol 37 ◽  
Author(s):  
Hannah Sievers ◽  
Angelika Joos ◽  
Mickaël Hiligsmann
Keyword(s):  
Health Technology Assessment ◽  
Technology Assessment ◽  
Real World ◽  
Health Technology ◽  
Practical Experience ◽  
Added Value ◽  
Real World Data ◽  
Conflicting Demands ◽  
Real World Evidence ◽  
Semistructured Interviews

Abstract Objective This study aims to assess stakeholder perceptions on the challenges and value of real-world evidence (RWE) post approval, the differences in regulatory and health technology assessment (HTA) real-world data (RWD) collection requirements under the German regulation for more safety in drug supply (GSAV), and future alignment opportunities to create a complementary framework for postapproval RWE requirements. Methods Eleven semistructured interviews were conducted purposively with pharmaceutical industry experts, regulatory authorities, health technology assessment bodies (HTAbs), and academia. The interview questions focused on the role of RWE post approval, the added value and challenges of RWE, the most important requirements for RWD collection, experience with registries as a source of RWD, perceptions on the GSAV law, RWE requirements in other countries, and the differences between regulatory and HTA requirements and alignment opportunities. The interviews were recorded, transcribed, and translated for coding in Nvivo to summarize the findings. Results All experts agree that RWE could close evidence gaps by showing the actual value of medicines in patients under real-world conditions. However, experts acknowledged certain challenges such as: (i) heterogeneous perspectives and differences in outcome measures for RWE generation and (ii) missing practical experience with RWD collected through mandatory registries within the German benefit assessment due to an unclear implementation of the GSAV. Conclusions This study revealed that all stakeholder groups recognize the added value of RWE but experience conflicting demands for RWD collection. Harmonizing requirements can be achieved through common postlicensing evidence generation (PLEG) plans and joint scientific advice to address uncertainties regarding evidence needs and to optimize drug development.

scholarly journals Offline Algorithms in Low-Frequency Trading

Queue ◽  
2020 ◽  
Vol 18 (6) ◽  
pp. 37-51
Author(s):  
Terence Kelly
Keyword(s):  
Real Time ◽  
Real World ◽  
High Frequency ◽  
Low Frequency ◽  
Combinatorial Auctions ◽  
Hard Problem ◽  
Np Hard ◽  
High Stakes ◽  
Drill Bits ◽  
Np Hard Problem

Expectations run high for software that makes real-world decisions, particularly when money hangs in the balance. This third episode of the Drill Bits column shows how well-designed software can effectively create wealth by optimizing gains from trade in combinatorial auctions. We'll unveil a deep connection between auctions and a classic textbook problem, we'll see that clearing an auction resembles a high-stakes mutant Tetris, we'll learn to stop worrying and love an NP-hard problem that's far from intractable in practice, and we'll contrast the deliberative business of combinatorial auctions with the near-real-time hustle of high-frequency trading. The example software that accompanies this installment of Drill Bits implements two algorithms that clear combinatorial auctions.

scholarly journals An Empirical Comparison of Bias Reduction Methods on Real-World Problems in High-Stakes Policy Settings

2021 ◽  
Vol 23 (1) ◽  
pp. 69-85
Author(s):  
Hemank Lamba ◽  
Kit T. Rodolfa ◽  
Rayid Ghani
Keyword(s):  
Machine Learning ◽  
Real World ◽  
Empirical Work ◽  
Bias Reduction ◽  
Research Community ◽  
High Stakes ◽  
Social Good ◽  
Social Service Delivery ◽  
Reduction Methods ◽  
Applications Of Machine Learning

Applications of machine learning (ML) to high-stakes policy settings - such as education, criminal justice, healthcare, and social service delivery - have grown rapidly in recent years, sparking important conversations about how to ensure fair outcomes from these systems. The machine learning research community has responded to this challenge with a wide array of proposed fairness-enhancing strategies for ML models, but despite the large number of methods that have been developed, little empirical work exists evaluating these methods in real-world settings. Here, we seek to fill this research gap by investigating the performance of several methods that operate at different points in the ML pipeline across four real-world public policy and social good problems. Across these problems, we find a wide degree of variability and inconsistency in the ability of many of these methods to improve model fairness, but postprocessing by choosing group-specific score thresholds consistently removes disparities, with important implications for both the ML research community and practitioners deploying machine learning to inform consequential policy decisions.

Predicting the trend of stock index based on feature engineering and CatBoost model

2021 ◽  
pp. 2150027
Author(s):  
Renzhe Xu ◽  
Yudong Chen ◽  
Tenglong Xiao ◽  
Jingli Wang ◽  
Xiong Wang
Keyword(s):  
Time Scale ◽  
Stock Index ◽  
Feature Engineering ◽  
Multiple Time ◽  
Multiple Time Scale ◽  
Back Testing ◽  
Price Series ◽  
Out Of Sample ◽  
Out Of Sample Prediction ◽  
Sharp Ratio

As an important tool to measure the current situation of the whole stock market, the stock index has always been the focus of researchers, especially for its prediction. This paper uses trend types, which are received by clustering price series under multiple time scale, combined with the day-of-the-week effect to construct a categorical feature combination. Based on the historical data of six kinds of Chinese stock indexes, the CatBoost model is used for training and predicting. Experimental results show that the out-of-sample prediction accuracy is 0.55, and the long–short trading strategy can obtain average annualized return of 34.43%, which is a great improvement compared with other classical classification algorithms. Under the rolling back-testing, the model can always obtain stable returns in each period of time from 2012 to 2020. Among them, the SSESC’s long–short strategy has the best performance with an annualized return of 40.85% and a sharp ratio of 1.53. Therefore, the trend information on multiple time-scale features based on feature engineering can be learned by the CatBoost model well, which has a guiding effect on predicting stock index trends.

scholarly journals Test de Français Laval-Montreal: Does It Measure What It Should Measure?

2016 ◽  
Vol 32 ◽  
pp. 87
Author(s):  
Romain Schmitt ◽  
Shahrzad Saif
Keyword(s):  
Preservice Teachers ◽  
Language Use ◽  
Target Language ◽  
French Language ◽  
Task Characteristics ◽  
Language Abilities ◽  
High Stakes ◽  
Formation Initiale ◽  
Teacher Training Programs ◽  
Language Needs

This article reports on a study conducted as part of a larger investigation of the predictive validity of the Test de Français Laval-Montreal (TFLM), a high-stakes French language test used for admission and placement purposes for Teacher- Training Programs (TTPs) in major francophone universities in Canada (Schmitt, 2015). The objective of this study is to examine the validity of TFLM tasks for measuring language abilities required by tasks common to the Target Language Use (TLU; Bachman & Palmer, 2010) domains in which preservice teachers are expected to function. Adopting Messick’s conception of construct validity (1989) and Bachman & Palmer’s Framework of Task Characteristics (2010), the study features a comprehensive task analysis detailing the characteristics of TFLM tasks in contrast to those of three major TLU academic and instructional contexts linked to the test. The results of the study are discussed in terms of the standards of validity (Messick, 1996) and qualities of usefulness (Bachman & Palmer, 1996). Findings suggest that TFLM tasks and constructs do not represent those of the TLU contexts and do not address the language needs of preservice teachers as identified by the Ministère de l’Éducation, du Loisir et du Sport (MELS). The implications for the consequential aspect of TFLM validity and the potential nega- tive consequences of TFLM use as an admission test are discussed. Cet article présente une partie d’une étude plus complète sur la validité prédictive du Test de Français Laval-Montréal (TFLM), test de langue française à enjeux critiques utilisé comme test d’admission et de placement dans les programmes de formation initiale en enseignement d’importantes universités francophones au Canada (Schmitt, 2015). Le but de ce e étude est d’analyser la validité des tâches du TFLM à des fins d’évaluation des compétences linguistiques exigées dans les tâches communes aux domaines d’utilisation de la langue cible dans lesquels les enseignants en formation doivent fonctionner (Target Language Use (TLU); Bachman & Palmer, 2010). Basée sur la conception de la validité conceptuelle de Messick (1989) et le cadre d’analyse des caractéristiques des tâches de Bachman & Palmer (2010), l’étude compare de manière détaillée les tâches du TFLM à celles de trois contextes académiques et pédagogiques d’emploi de la langue cible. Les résultats de cette analyse sont évalués en termes de validité (Messick, 1996) et des qualités des tests (Bachman & Palmer, 1996). Les résultats indiquent que les tâches du TFLM et les construits qu’il est sensé évaluer ne correspondent pas à ceux des contextes d’emploi de la langue cible et ne répondent pas aux besoins des ensei- gnants en formation tels qu’identi és par le Ministère de l’Éducation, du Loisir et du Sport (MELS). La validité du TFLM, les conséquences ainsi que les aspects potentiellement négatifs de son utilisation comme test d’admission sont discutés. 

Customizing persuasive messages; the value of operative measures

2018 ◽  
Vol 35 (2) ◽  
pp. 208-217 ◽  
Author(s):  
Maurits Kaptein
Keyword(s):  
Interactive Media ◽  
Database Systems ◽  
Customer Behavior ◽  
Customer Relationship ◽  
Persuasive Messages ◽  
Psychological Traits ◽  
Content Type ◽  
Out Of Sample ◽  
Out Of Sample Prediction ◽  
Operative Measures

Purpose This paper aims to examine whether estimates of psychological traits obtained using meta-judgmental measures (as commonly present in customer relationship management database systems) or operative measures are most useful in predicting customer behavior. Design/methodology/approach Using an online experiment (N = 283), the study collects meta-judgmental and operative measures of customers. Subsequently, it compares the out-of-sample prediction error of responses to persuasive messages. Findings The study shows that operative measures – derived directly from measures of customer behavior – are more informative than meta-judgmental measures. Practical implications Using interactive media, it is possible to actively elicit operative measures. This study shows that practitioners seeking to customize their marketing communication should focus on obtaining such psychographic observations. Originality/value While currently both meta-judgmental measures and operative measures are used for customization in interactive marketing, this study directly compares their utility for the prediction of future responses to persuasive messages.

scholarly journals Reasoning with linguistic preferences using NPN logic

2010 ◽  
Vol 7 (3) ◽  
pp. 511-528 ◽  
Author(s):  
Goran Devedzic ◽  
Danijela Milosevic ◽  
Lozica Ivanovic ◽  
Dragan Adamovic ◽  
Miodrag Manic
Keyword(s):  
Side Effects ◽  
Decision Analysis ◽  
Decision Maker ◽  
Real World ◽  
Cognitive Maps ◽  
Fuzzy Cognitive Maps ◽  
Basic Notion ◽  
The Real ◽  
Alternative Framework ◽  
Linguistic Models

Negative-positive-neutral logic provides an alternative framework for fuzzy cognitive maps development and decision analysis. This paper reviews basic notion of NPN logic and NPN relations and proposes adaptive approach to causality weights assessment. It employs linguistic models of causality weights activated by measurement-based fuzzy cognitive maps? concepts values. These models allow for quasi-dynamical adaptation to the change of concepts values, providing deeper understanding of possible side effects. Since in the real-world environments almost every decision has its consequences, presenting very valuable portion of information upon which we also make our decisions, the knowledge about the side effects enables more reliable decision analysis and directs actions of decision maker.

Sign in / Sign up

Export Citation Format

Share Document