Flawed Decision-making Research Articles

COVID-19 death toll predictions show that triggering counterfactual thinking deteriorates judgmental performance

BackgroundEffective communication during a health crisis is critical as it directly influences psychological and behavioral responses that will shape the further progression of the crisis. Past research has suggested that one type of cognitive mechanism that is likely to be affected by the framing of public health messages relates to counterfactual thinking.MethodsBased on 6731 incentivized daily forecasts collected over 377 days (from April 2020–May 2021), we investigate the role of triggering counterfactual thinking when interpreting public information regarding the daily US death toll from COVID-19.ResultsHere we show that individuals who engaged in thinking about “interventions that could have led to an alternative evolution of the death toll” prior to making forecasts exhibit greater judgmental bias in their predictions compared to the control group. Specifically, subjects in the treatment group tend to generate upward counterfactuals and underestimate the death toll, potentially due to anchoring on more favorable scenarios and insensitivity to trend changes. Interestingly, this behavior is also observed among individuals who had recovered from COVID-19 (or someone in their close social circle).ConclusionsOur findings underscore the importance of using debiasing strategies and neutral communication during health crises to mitigate the generation of upward counterfactuals, thus reducing the likelihood of systematic misperceptions and flawed decision-making.

Evaluating Internal Control Mechanisms in Malaysian Public Sector Initiatives: Insights from the Public Accounts Committee (PAC) Reports

This study examines internal control challenges in Malaysian government projects, drawing on Public Accounts Committee (PAC) reports from 2019 to 2022. The focus on this period stems from the shift in PAC chairmanship from government to opposition, which enhanced perceived independence. However, the literature suggests that the PAC’s success depends more on collaborative decision-making than on the chairman’s political affiliation. Analyzing 29 government agencies across various sectors revealed recurring weaknesses in internal controls. The study employed the Committee of Sponsoring Organizations of the Treadway Commission’s (COSO) Internal Control-Integrated Framework (ICIF) to systematically identify these weaknesses. A content analysis method was used to examine the themes, which were grouped into five ICIF components: control environment, risk assessment, control activities, information and communication, and monitoring. Weaknesses in the control environment include conflicts of interest, compromised ethical values, inadequate governance, and a lack of accountability and transparency. Risk assessment deficiencies were found in financial risk management, procurement and acquisition risks, geopolitical and strategic risks, intellectual property and technology risks, and project management risks. Control activity weaknesses involved poor resource and process management, non-compliance, and flawed decision-making. Information and communication issues included poor documentation, inadequate system integration and data sharing, limited stakeholder engagement, weak communication protocols, and poor information management. Monitoring deficiencies were noted in regulatory compliance, contract management, financial oversight, monitoring systems, and disbursement tracking. Addressing these weaknesses in internal controls is crucial for minimizing the risks of fraud, waste, and loss of public trust

Optimizing data engineering for AI: improving data quality and preparation for machine learning application

Data engineering has become a cornerstone in modern artificial intelligence (AI) and machine learning (ML) initiatives, playing a critical role in transforming raw data into actionable insights. Despite significant progress in algorithmic development and computational power, the effectiveness of AI models is still highly dependent on the quality of their input data. This study delves into a comprehensive exploration of data engineering practices, focusing on strategies to optimize data quality and data preparation processes for machine learning applications. We begin by recognizing that AI systems, regardless of their level of sophistication, are only as robust as the data used to train them. Therefore, datasets contaminated by inconsistencies, missing values, redundancy, or lack of structural integrity can significantly degrade both model accuracy and performance, leading to flawed decision-making.In this extensive work, we argue that robust data engineering pipelines, characterized by rigorous data ingestion, cleaning, transformation, and feature engineering processes, are vital to the success of modern AI systems. Through an in-depth review of current literature, we identify common challenges faced during data preparation, such as the integration of heterogeneous data sources, handling of large-scale streaming data, and ensuring real-time system responsiveness. Furthermore, we explore traditional approaches, including Extract-Transform-Load (ETL) techniques, along with more contemporary methods like ELT (Extract-Load-Transform) and streaming pipelines that cater to the dynamic needs of big data environments.The study’s methodological framework encompasses a multi-stage process in which we adopt both qualitative and quantitative measures to evaluate data pipeline designs. We synthesize findings from scholarly research, industry best practices, and real-world implementations to formulate a set of standards for measuring data readiness, including timeliness, accuracy, completeness, consistency, and integrity. These metrics serve as foundational benchmarks to ascertain where conventional pipelines fall short and where novel optimization techniques can be introduced. Finally, we present results from experimental validations that reveal how improved data engineering methodologies do not merely enhance the predictive strength of machine learning models but also optimize computational efficiency by reducing training times and resource utilization.By demonstrating measurable benefits—including cleaner datasets, lower error rates, and higher model performance—this paper underscores the significance of placing data engineering and data quality at the forefront of AI development. The conclusion consolidates these insights and addresses the broader implications for future work, emphasizing the need for continued innovation in data pipeline optimization, governance, and standardization. Implementing robust data engineering practices can have transformative effects on various domains, ranging from healthcare and finance to e-commerce and manufacturing, where data-driven insights are increasingly shaping strategic decision-making. It is our hope that this comprehensive examination stimulates ongoing research and facilitates the adoption of best practices across the global AI and ML community.

Incorporating Climate Adaptation into a Forest Management Plan: A Case Study on the Research and Teaching Forest of Michigan Technological University

Abstract We incorporated climate adaptation into a forest management plan for Michigan Technological University’s Ford Forest, a 2,000-ha property in Michigan’s Upper Peninsula used for education, research, and timber revenues. Our process was an opportunity to test the existing climate adaptation literature, pulling from multiple sources to meet the diverse needs of the institution at the time the plan was created. We present outcomes as well as lessons learned summarized in a series of broad takeaway messages. (1) Climate adaptation is a means to an end. The end must be defined and is made up of critical values and an adapted state. (2) Given the instability inherent in climate change, achieving adaptation means adjusting forest attributes in response to change but also preparing for unforeseen outcomes via adaptive management. (3) Decisions surrounding goals affect the entire process. Considering climate change–driven constraints when setting goals will improve outcomes. (4) Flawed decision-making is a risk associated with certain organizational contexts and affects the identification of goals, vulnerabilities, and adaptation options. (5) Climate adaptation actions cannot be evaluated in the short term for efficacy because climate change is ongoing. However, the appropriateness of a plan in responding to anticipated change can be evaluated. Study Implications: This work informs strategic planning for climate change adaptation in forests. We created a climate-informed forest management plan using adaptation frameworks for a university-owned forest in Upper Michigan at the confluence of the northern hardwoods and boreal forest ecotypes. We offer assessment of our plan outcomes and insights into how our decision-making context affected them. This case study expands on the ongoing scientific conversation, incorporating concepts from management science, on how best to adapt natural systems to climate change to protect human values derived from ecosystem services.

Offender-Centric Policing in Cases of Rape

The article explores the idea of ‘offender-centric’ policing in cases of rape, with its focus on suspect and offender admissions and behaviours. It features discussion of 11 cases, illustrating offender-centric pathways to charge or conviction, the challenges facing complainants, suspects and police officers, along with missed opportunities to focus on a suspect’s behaviour. The importance of victim care and support is discussed, and it is argued that victim care should accompany an offender-centric approach to rape investigation. It is also argued that there are potential dangers with offender-centric tactics, specifically, that without due care it may become a self-confirming investigative tool influenced by confirmation bias which may lead to flawed decision-making. The article concludes by arguing that offender-centric policing has benefits in those cases with suspects who engage in predatory behaviour, have a history of previously undisclosed sexual offending and domestic violence and other problematic behaviours. It also has value in focusing the attention of investigators on what steps were taken by a suspect to ascertain the complainant’s consent. While the offender-centric approach cannot address all investigative challenges in rape cases, it is a useful addition to existing strategies.

Contextual information and cognitive bias in the forensic investigation of fatal fires: Do these incidents present an increased risk of flawed decision-making?

In sudden or unexpected deaths, there is a need to identify whether or not the death resulted from a criminal act. Adding evidentiary problems associated with fire and arson to this further complicates the investigation. A multi-agency approach with an open exchange of information is recommended, though the use of unrestrictive contextual information is problematic. Extraneous contextual information may introduce bias into the investigative process, risking flawed decision-making, and fatal fires may be particularly vulnerable to this. The authors examine the effect of cognitive bias on fire investigation and argue that, in fatal fire investigations, the sharing of contextual information should involve the use of strict information management strategies to ensure that forensic experts only receive task-relevant contextual information.

Gamble with Your Head and Not Your Heart: A Conceptual Model for How Thinking-Style Promotes Irrational Gambling Beliefs.

Dual process theory suggests that people use two processing systems to filter information and form judgments that direct a course of action: an intuitive and an analytic system. While the intuitive system is necessary for efficient and effective daily functioning, reliance on fast, intuitive thinking when gambling is likely to result in biased or flawed decision-making. Those who gamble tend to endorse an array of fallacious or irrational beliefs that contribute to risky decision-making and excessive gambling. This paper argues that gambling beliefs may be developed and reinforced through underlying cognitive mechanisms described by dual process theory. More specifically, gamblers tend to apply assumptions and theories developed based on their understanding of the natural world to artificial gambling contexts where such rules do not apply. As a result, gamblers develop biased interpretations and understandings for how gambling works, which tend to align with personal schemas, experiences and gambling motivations. These beliefs are used in future gambling contexts to inform decision-making. Gamblers are often unlikely or unwilling to reflect on the veracity of beliefs as they are often used to justify gambling behaviours. Educating gamblers on how they make decisions and encouraging them to think more analytically may help to reduce the strength with which erroneous beliefs about gambling are endorsed, resulting in safer gambling decisions.

Damages, the Crown-Soldier Relationship, and the Canadian Forces Grievance Process

A recent Reference initiated before the Federal Court (but not yet decided) concerning a grievance by an officer of the Canadian Forces (CF) offers an opportunity to re-examine the unique legal dimensions of the Crown-soldier relationship. This relationship remains grounded in a Victorian perspective – some might say ‘antiquated’ – in which members of the armed forces are viewed as being in a ‘unilateral commitment in return for which the Queen assumes no obligations’, thus limiting recourse for perceived wrongs on the part of the Crown. One of the principal failings of much of the case law concerning this ‘employment’ relationship is that it tends to state what the relationship is not, rather than what it is. Using the issues raised by the Reference, this paper proposes to examine how the Crown-soldier relationship may be viewed in the context of modern public law, focusing on the circumstances under which the Crown might be liable for damages or compensation where members of the armed forces are subject to flawed decision-making. One of the principal conclusions drawn by the author is that the Federal Court is unlikely to signal a paradigm shift in this relationship and, barring legislative reform, the courts will take an incrementalist approach to modernization.

Investigating decision-making mechanisms and biases in Dutch criminal investigation teams by using a serious game

Investigating decision-making mechanisms and biases in Dutch criminal investigation teams by using a serious game In this article we examine by means of a serious game how ten teams of police leaders from major criminal investigation teams from five regional forces in the Netherlands, during criminal investigation, deal with tunnel vision and other potential causes of flawed decision-making, described according to Naturalistic Decision-Making models. Findings show that in the serious game, the danger of tunnel vision was widely acknowledged and that a great deal of energy was wasted as a result. In addition, the teams proved susceptible to other types of decision-making pitfalls. For example, the teams searched predominantly for confirmatory evidence, unconsciously used ingrained process-related rules of thumb, and there was evidence of a form of ‘information impulsion fallacy’. The present research is an elaboration on existing literature in that it attempts to shed light on decision-making practices during criminal investigations. The study shows that a serious game can be a useful tool to uncover decision-making behaviour.

Aggregating spatially explicit criteria: avoiding spatial compensation

Since the shift from safety-oriented planning towards risk-based flood management planning, both hydrology/hydraulic research as well as operations research have made remarkable progress. Unfortunately, in the transfer of information between both areas of expertise, valuable spatial information is being lost, which may lead to flawed decision-making. Spatial aggregation of positive or negative criteria score across cells as is customary holds a major pitfall: spatial compensation at the river basin scale. If one cell's benefits outweigh another cell's detriments, these detriments are lost in the spatial aggregation process. However, not all variables are commensurable. A decision-maker might reasonably object to an alternative that generates additional damage, even if they are amply compensated elsewhere. Current aggregation procedures have more of a tendency to veil these problems then to deal with them. Moreover, valuable information defined as spatial equity is also often lost in the decision matrix. In this paper, classical criteria are changed and new criteria are selected to inhibit the abundant spatial compensation, add the variable of spatial equity and provide decision-makers with non-ensconced and therefore more accurate results. The shortcomings and error-proneness of the established frugal summation procedures are outlined by intuitive simplified examples. These examples also serve to illustrate and test the logical performance of the proposed methodology. To test their applicability, they are used in an actual test case, selected for the problem's intricacy.

Withdrawing and withholding life-sustaining treatment

There has been considerable criticism recently over the use of end-of-life pathways and do-not-resuscitate orders with vulnerable, incapable patients, often without discussion. This criticism has led to press and judicial scrutiny of the lawfulness of decisions to withdraw or withhold life-sustaining treatment. In this article, the author reviews the law and argues that flawed decision-making by doctors and nurses, rather than the actual end-of-life pathway or do-not-resuscitate order, is where the lawfulness of withdrawing and withholding treatment must be challenged.

“Cost-effectiveness” Estimates Result in Flawed Decisionmaking in Listing Drugs for Reimbursement

Facing financial pressures, the provinces and territories have chosen to use "cost-effectiveness" for making decisions about drug listings. This study examines the scientific basis for the procedures used to determine cost-effectiveness in 5 Canadian provinces. Questionnaires were mailed to key provincial informants asking about the respondent's expertise and role, the administrative and scientific basis for decision-making, organizational structures and other factors in the scientific evaluation and decision-making process, and the transparency of the process. There were also questions about the data required and received and their importance, the place of cost-effectiveness and other economic impact evaluations, the data sources for them, and the use of follow-up monitoring to evaluate the decisions made. Information required by the provinces for decision-making about cost-effectiveness is not available to them at the time of their decisions about listing new medications. The primary sources of data on both efficacy and cost-effectiveness are pharmaceutical companies. Efficacy information is generated in a scientifically rigorous manner, whereas the effectiveness and cost data are estimates potentially subject to biases and evaluated by judgement (expert opinion) alone. Moreover, there is no collaboration in the assessment process between provinces. The outcomes are large differences between provinces in the decisions made and, hence, in the pharmaceuticals accessible to residents. Current methods for making decisions about provincial drug listings are based on inadequate data, and the lack of consistency in the provinces' decisions suggest they may be scientifically flawed. We recommend establishing a single national scientific review committee, with re-evaluation of each drug's cost-effectiveness after a suitable period of monitored use.

Fatigue in anesthesia: implications and strategies for patient and provider safety.

HEALTHCARE delivery takes place 24 h a day, 7 days a week, and is colloquially termed a “24/7” operation. Anesthesia providers are required to deliver critical around-the-clock care to a variety of patients. This parallels the situation in many other domains that provide such services, e.g., transportation, law enforcement, communications, fire fighting, technology, manufacturing, and the military. Even “convenience” industries (e.g., gas stations and grocery stores) now provide uninterrupted access. These continuous operational demands present unique physiologic challenges to the humans who are called on to provide safe operations within these systems. Human physiologic design dictates circadian patterns of alertness and performance and includes a vital need for sleep. Human requirements for sleep and a stable circadian clock can be, and often are, in direct opposition to the societal demand for continuous operations. Recently, patient safety has taken center stage in health care. The Institute of Medicine’s report “To Err Is Human: Building a Safer Health System,” revealed that medical errors contribute to many hospital deaths and serious adverse events. The response to this report was widespread and included the Quality Interagency Coordination Task Force’s response to the President of the United States, “Doing What Counts for Patient Safety: Federal Actions to Reduce Medical Errors and Their Impact.” 2 This report listed more than 100 action items to be undertaken by federal agencies to improve quality and reduce medical errors. One action promised by the Agency for Healthcare Research and Quality was “the development and dissemination of evidence-based, best safety practices to provider organizations.” In addition to the multiple recommendations to improve patient safety, the report from the Agency for Healthcare Research and Quality included a review chapter on sleep, fatigue,# and medical errors.** There is evidence that the issue of fatigue in health care is coming to prominence on a national level. In April 2001, Public Citizen (a consumer and health advocacy group) and a consortium of interested parties petitioned the Occupational Safety and Health Administration to implement new regulations on resident work hours (table 1). The primary intent of the regulations is to provide more humane working conditions, which the petitioners declare will result in a better standard of care for all patients. Also, the Patient and Physician Safety and Protection Act of 2001, which would limit resident physician work hours, was introduced in Congress. Recently, the Accreditation Council on Graduate Medical Education, the accrediting organization for residency training programs in the United States, has approved common program requirements for resident duty and rest hours that will take effect in July 2003.†† This article is accompanied by an Editorial View. Please see: Lydic R: Fact and fantasy about sleep and anesthesiology. ANESTHESIOLOGY 2002; 97:1050–1.

3 Human factors in the operating room: interpersonal determinants of safety, efficiency and morale

The majority of accidents in technical professions have human error as a causal element. More critically, these errors tend to involve interpersonal issues: communications, leadership, conflict, flawed decision-making, etc. Aviation has responded by initiating formal instructions in the interpersonal aspects of human factors through programmes that are known as Crew Resource Management (CRM) training. Research into the interpersonal activities in the operating room has demonstrated similar problems in communication, conflict and situational awareness. To define the issues more sharply, an input-process-outcome model of operating room team performance was adapted from one developed in aviation. The model defines areas such as attitudes, organizational culture and group processes that are amenable to improvement through training and organizational interventions. It has also led to the development of formal human factors training programmes that, like that in aviation, involve simulation. Team-Oriented Medical Simulation (TOMS) at the University of Basel/Kantonsspital is a training programme that includes a full operating room simulator. Anaesthetists, surgeons, nurses and orderlies conduct scheduled surgery on an instrumented mannequin that allows both anaesthetic and laparoscopic simulation. Initial results from simulations show highly favourable reactions from participants. Human factors and quality assurance programmes need to be data-driven to have a positive impact. Survey and systematic observational methodologies have been used to define areas that may benefit from training. Results from three hospitals showed highly significant differences in attitudes regarding the interpersonal aspects of surgery between organizations and among subgroups (anaesthetists, surgeons, nurses) within organizations. We conclude that integrated quality assurance and human factors programmes have the potential to increase safety, efficiency and job satisfaction among medical personnel.

Groupthink and Integrative Complexity in British Foreign Policy-Making: the Munich Case

Walker, S. G. and Watson, G. L. Groupthink and Integrative Complexity in British Foreign Policy-Making: the Munich Case. Cooperation and Conflict, XXIV, 1989, 199-212. We focus in this paper upon the groupthink construct as a partial explanation for the flawed decision-making by British leaders in managing the Munich crisis with Germany. In order to assess its explanatory power, we follow the precedent established in previous case studies, which explore the link between groupthink and the levels of integrative complexity exhibited by decision-makers. The hypothesis linking the two phenomena is that in cases where groupthink is present, the level of integrative complexity will be significantly lower than in cases where it is absent. The Munich case is a particularly appropriate case for replicating the test of this hypothesis because, unlike other groupthink cases, it did not lead to unsuccessful military interventions or escalations. Our results indicate that: (1) groupthink existed in the Munich case; however, it did not persist throughout the entire crisis. (2) Its presence and absence within the crisis fluctuated consistently with the expectations associated with the theory which accounts for its dynamics. (3) The integrative complexity exhibited by Prime Minister Chamberlain and Foreign Secretary Halifax varied as predicted by the hypothesis linking it with groupthink.