Causal Explanations Research Articles

Categorical and Continuous Features in Counterfactual Explanations of AI Systems

Recently, eXplainable AI (XAI) research has focused on the use of counterfactual explanations to address interpretability, algorithmic recourse, and bias in AI system decision-making. The developers of these algorithms claim they meet user requirements in generating counterfactual explanations with “plausible”, “actionable” or “causally important” features. However, few of these claims have been tested in controlled psychological studies. Hence, we know very little about which aspects of counterfactual explanations really help users understand the decisions of AI systems. Nor do we know whether counterfactual explanations are an advance on more traditional causal explanations that have a longer history in AI (e.g., in expert systems). Accordingly, we carried out three user studies to (i) test a fundamental distinction in feature-types, between categorical and continuous features, and (ii) compare the relative effectiveness of counterfactual and causal explanations. The studies used a simulated, automated decision-making app that determined safe driving limits after drinking alcohol, based on predicted blood alcohol content, where users’ responses were measured objectively (using predictive accuracy) and subjectively (using satisfaction and trust judgments). Study 1 (N = 127) showed that users understand explanations referring to categorical features more readily than those referring to continuous features. It also discovered a dissociation between objective and subjective measures: counterfactual explanations elicited higher accuracy than no-explanation controls but elicited no more accuracy than causal explanations, yet counterfactual explanations elicited greater satisfaction and trust than causal explanations. In Study 2 (N = 136) we transformed the continuous features of presented items to be categorical (i.e., binary) and found that these converted features led to highly accurate responding. Study 3 (N = 211) explicitly compared matched items involving either mixed features (i.e., a mix of categorical and continuous features) or categorical features (i.e., categorical and categorically-transformed continuous features), and found that users were more accurate when categorically-transformed features were used instead of continuous ones. It also replicated the dissociation between objective and subjective effects of explanations. The findings delineate important boundary conditions for current and future counterfactual explanation methods in XAI.

Dyslexia in the twenty-first century: a commentary on the IDA definition of dyslexia

In offering a commentary upon the IDA definition, we address its main components in turn. While each is technically accurate, we argue that, when taken together, the definition, or more accurately, the use to which it is often put, becomes problematic. We outline different current conceptions of dyslexia and conclude that the operationalisation of the definition for diagnostic purposes often results in scientifically questionable diagnoses and inadvertently leads to significant educational inequity. We propose a simpler definition that describes the primary difficulty, avoids reference to causal explanation, unexpectedness, and secondary outcomes, and redirects practitioner and policymaker focus to the importance of addressing and meeting the needs of all struggling readers.

A Descriptive Diagnosis or a Causal Explanation? Accuracy of Depictions of Depression on Authoritative Health Organization Websites

Plain Language SummaryMost psychiatric diagnoses are descriptive: They describe states of mental distress and dysfunction but do not in themselves contain causal explanations. Nonetheless, diagnoses in psychiatry are commonly talked about as if they are concrete entities that explain the symptoms they describe. In this study, we examined whether health organizations themselves contribute to this logical fallacy. We searched for popular websites managed by leading mental health organizations, and evaluated whether they discussed the diagnosis of depression accurately as a description, or inaccurately as a cause for depressive symptoms. We found that the majority of websites presented depression as a cause, instead of a description of symptoms. We discuss the potential harmful consequences of inaccurately presenting descriptive psychiatric diagnoses as causes.

IF YOU COULD READ MY MIND: ON THE HISTORY OF MIND AND OTHER MATTERS

ABSTRACTIn The Primacy of Method in Historical Research: Philosophy of History and the Perspective of Meaning, Jonas Ahlskog presents a critical and lucid engagement with contemporary philosophies of history and makes a sustained case for a return to the ideas of history and social science as developed by R. G. Collingwood and Peter Winch. What philosophy needs again is, first, a recognition of the “primacy of method”—that is, the insight that what one knows about reality depends on how one knows it. Second, philosophers need to take “the duality of method” seriously again and to recognize that the modes of explanation in the human sciences and the natural sciences are categorically different from each other—especially now that this difference has been blurred in recent debates about the Anthropocene. Ahlskog's book is thus also a contribution to the classical debate about causal explanation versus meaningful understanding. On closer analysis, however, Ahlskog's “untimely meditations” on “historical method” suffer from an insufficient engagement with counterarguments. A first line of critique challenges the idea that human action cannot be explained causally. A second line of critique challenges the idea that all aspects of human action can be “understood,” because the unintended aspects and consequences of individual actions cannot. These require causal explanation. A third line of critique concerns Ahlskog's denial of the fundamental plurality of ideas of history and the social sciences. Squeezing this plurality into one philosophical mold comes at a price. Unintentionally, Ahlskog's “untimely meditations” also show that much.

The Plague: Modern life

AbstractThe social structures and thought patterns of the modern world are the fruits of the Enlightenment, which begins by eliminating final causal explanations in favour of purely material and efficient causes. The development and great technical success of Enlightenment procedures has, however, produced a cultural blindness about the good. Camus's novel shows us this cultural blindness through characters who themselves suffer from it; for modern man, it is almost a natural evil—we are born into it. Camus' hope must have been that the vision of the characters' incapacity might help break us from the same incapacity. What the cured life looks like Camus does not say; in the world of our plague, it appears as either odd or mystery.

CUBE: Causal Intervention-Based Counterfactual Explanation for Prediction Models

CUBE: Causal Intervention-Based Counterfactual Explanation for Prediction Models

Lessons learned from Norway: a values-based formulation of inclusive education

ABSTRACT Given the numerous conceptual approaches to understanding inclusive education, there is an obvious risk of fragmentation and stagnation in the field. In response, this paper aims to contribute to advancement by going beyond previous work and developing a holistic formulation of inclusive education. Its starting point is that the persistent challenge of offering equal opportunities and full participation for all students in regular schools appears to have complex causal explanations on multiple levels, including the ideological/political, practical, and individual/subjective. This holistic formulation draws on analyses of a) the connections between inclusive education, student diversity and the concept of special educational needs, b) some problematic experiences from Norway, and c) the current state of the field. In this formulation, three core values of inclusion— welcoming communities, combating discriminatory attitudes, and education for all – as articulated in the Salamanca Declaration, are linked to the institutional practice principles of participation, human diversity, and differentiation. The discussion explores these values and their alignment with the principles of inclusive practice and students’ inclusionary outcome in school. Furthermore, the interplay among these values and their corresponding principles of practice principle is illustrated, emphasising the need for awareness when prioritising one over another, both in research and practice.

Can data science achieve the ideal of evidence-based decision-making in environmental regulation?

Data science, including big data analytics and machine-learning techniques, is increasingly utilized in environmental governance. This study explores the perspectives of environmental experts on the limitations of data science in evidence-based environmental regulations. Based on semi-structured interviews with 20 environmental experts, this study examines three key limitations: the lack of local specificity in big data, the uncertainty of the environment as an open system, and the opacity of machine learning. The main findings of the study are as follows: First, the lack of spatial and temporal specificity in big data poses challenges in using them as regulatory evidence, requiring revalidation through traditional assessment methods. Second, while machine learning performs well in closed virtual environments, it is susceptible to the fluid and uncertain nature of the environment in an open world. Third, the lack of causal explanation owing to the inscrutability of deep learning limits its utility for regulatory evidence. The study concludes that while the applicability of data science to evidence-based decision-making in environmental regulation is limited, it can still be utilized in data-driven environmental governance without the need for region-specific information or a high burden of decision-making responsibility. This includes tasks such as strategic environmental impact assessments, policy resource allocation, or environmental policy prioritization.

Recent Advances in Targeting Transition Metals (Copper, Iron, and Zinc) in Alzheimer's Disease.

Changes in the transition metal homeostasis in the brain are closely linked with Alzheimer's disease (AD), including intraneuronal iron accumulation and extracellular copper and zinc pooling in the amyloid plague. The brain copper, zinc, and iron surplus are commonly acknowledged characteristics of AD, despite disagreements among some. This has led to the theory that oxidative stress resulting from abnormal homeostasis of these transition metals may be a causative explanation behind AD. In the nervous system, the interaction of metals with proteins appears to be an essential variable in the development or suppression of neurodegeneration. Chelation treatment may be an option for treating neurodegeneration induced by transition metal ion dyshomeostasis. Some clinicians even recommend using chelating agents as an adjunct therapy for AD. The current review also looks at the therapeutic strategies that have been attempted, primarily with metal-chelating drugs. Metal buildup in the nervous system, as reported in the AD, could be the result of compensatory mechanisms designed to improve metal availability for physiological functions.

THE EXPLANATIONS ONE NEEDS FOR THE EXPLANATIONS ONE GIVES—THE NECESSITY OF EXPLAINABLE AI (XAI) FOR CAUSAL EXPLANATIONS OF AI-RELATED HARM:DECONSTRUCTING THE ‘REFUGE OF IGNORANCE’ IN THE EU’S AI LIABILITY REGULATION

This paper examines how explanations related to the adverse outcomes of Artificial Intelligence (AI) contribute to the development of causal evidentiary explanations in disputes surrounding AI liability. The study employs a dual approach: first, it analyzes the emerging global caselaw in the field of AI liability, seeking to discern prevailing trends regarding the evidence and explanations considered essential for the fair resolution of disputes. Against the backdrop of those trends, the paper evaluates the upcoming legislation in the European Union (EU) concerning AI liability, namely the AI Liability Directive (AILD) and Revised Product Liability Directive (R-PLD). The objective is to ascertain whether the systems of evidence and procedural rights outlined in this legislation, particularly the right to request the disclosure of evidence, enable litigants to adequately understand the causality underlying AI-related harms. Moreover, the paper seeks o determine if litigants can effectively express their views before dispute-resolution authorities based on that understanding. An examination of the AILD and R-PLD reveals that their evidence systems primarily support ad hoc explanations, allowing litigants and courts to assess the extent of the defendants' compliance with the standards enshrined in regulatory instruments, such as the AI Act. However, the paper contends that, beyond ad hoc explanations, achieving fair resolution in AI liability disputes necessitates post-hoc explanations. These should be directed at unveiling the functionalities of AI systems and the rationale behind harmful automated decisions. The paper thus suggests that ‘full’ explainable AI (XAI) that is, both ad hoc and post hoc, is necessary so that the constitutional requirements associated with the right to a fair trial (access to courts, equality of arms, contradictory debate) can be effectively met. Keywords: AI, Causation, Explainability, Fair Trial, Procedural Fairness, Equality of Arms, Effective Participation, AI liability, Product Liability, AI Act, AI Liability Directive, Product Liability Directive

The Dopamine Hypothesis as a Causal Explanation of Schizophrenia

Schizophrenia, a severe psychiatric disorder, profoundly impacts individuals, families, and society, characterized by symptoms such as delusions, hallucinations, disorganized thinking, speech, and motor behavior, as outlined in the DSM-5 criteria. Coined by Eugen Bleuler in 1911 to emphasize fragmented cognition, the term "schizophrenia" replaced Emile Kraepelin's "dementia praecox"(1908) reflecting its chronic nature, with an estimated prevalence of 1% in the general population and a significant heritability rate of around 79%. The dopamine hypothesis, central to schizophrenia research, suggests heightened dopaminergic transmission as a primary factor in its development, supported by the efficacy of antipsychotic drugs targeting dopamine receptors. However, recent studies have revealed complexities beyond dopamine dysfunction, including the glutamate hypothesis, which proposes deficits in glutamate activity as an alternative explanation. This essay critically evaluates the dopamine hypothesis within the broader biopsychosocial framework, emphasizing the interaction of biological, psychological, and social factors in schizophrenia etiology. Despite its foundational role, the dopamine hypothesis has limitations in fully clarifying the multifaceted nature of schizophrenia, highlighting the need for comprehensive approaches integrating diverse perspectives and methodologies to enhance understanding of this complex disorder. Keywords: Dopamine Hypothesis, Schizophrenia, Mental Illness, Psychotic Symptoms, GENE x ENVIROMENT Interaction

The Explanatory Power of Discourse Analysis

This paper discusses the explanatory power of discourse analysis, an approach to typically considered one of the main qualitative methods for data analysis. Although discourse analysis is typically not used to develop explanatory models, some have claimed that discourse models can causally explain social phenomena, such as institutional change. I analyze a case of institutional change to discuss and provide arguments for two claims. First, discourse analysis cannot generate causal explanations of social phenomena. Secondly, discourse analysis can generate non-causal explanations, specifically constitutive explanations of social phenomena. I conclude the paper by discussing possible implications for social research.

Genoeconomics: Decoding the genetic basis of entrepreneurial success

This paper explores the emerging intersection of genomics, neuroscience, and entrepreneurship. Specifically, it examines if and how genetic factors and neurobiological processes influence entrepreneurial behaviors and success. The key questions addressed are: 1) Is there a genetic basis for the entrepreneurial mindset? 2) Can genetic markers and neuroimaging data predict entrepreneurial outcomes? 3) What challenges arise when combining genetic, neuroscientific, and business data? 4) How can entrepreneurs utilize insights from genomics and neuroscience? The paper argues that synchronizing molecular genetics with behavioral data can offer new causal explanations for entrepreneurial cognition under the umbrella of "genoeconomics." However, there are methodological difficulties in analyzing genetically informative data. It cautions against genetic determinism while highlighting opportunities from mapping genetic entrepreneurial proclivities. Entrepreneurship is an intricate phenotype. Are entrepreneurs born or made or developed or adopted? What makes an entrepreneur think outside the box and pursuing avant-garde paths to attain entrepreneurial goals? Alvin Toffler mentioned of various ‘Waves’ viz. Agrarian, Industrial and so on. Economics and Business Management, today, is experiencing inestimable waves viz. Heterodox Wave, Genetic Wave, Molecular Wave, Cellular Wave, Hereditary Wave, [‘Infoplosion’ or ‘Info-Tectonics’ Wave, Capability Approach Wave, Institutional Wave, Behavioral / Experimental Wave [Cognitive and Emotional dimensions / Develops and uses experiments typically with human subjects], Evolutionary Wave, Cognitive Wave, Information Wave , Artificial [Synthetic and / or Fabricated] Intelligence Wave and Neuro Wave, to list a few, from a rational perspective. Of central interest in this paper is Managerial Business - Economics Wave. It deals with, with complex, deep-rooted problems, gears and technique to dissect market demand, assess expenses, establish pricing strategies, assess risks, and appreciate competitive dynamics. Is entrepreneurship a genetic trait? Of recent origin is emergence of incorporating genetic and biological markers into entrepreneurship. Issue that confronts is, is there a need to sync molecular genetics and entrepreneurial business data to genomic entrepreneurship; under umbrella of ‘Genoeconomics’? Does this mean that entrepreneurship is in genes? Do genes determine entrepreneurial success? Can genetics or molecular genetics and entrepreneurial data predict response to multifarious behavioral intercession? Can genetics calculate retort to intricate behavioral interference? Are genetic markers of interest for entrepreneurship research? What challenges occur when analyzing genetically informative facts? How, if at all, should entrepreneurs use and combine molecular genetics and business data? What challenges arise when analyzing genetically informative data? The moot issue is, can entrepreneurship be genetic? Can an entrepreneur be ‘genetically-inclined’? One debatable question is why do entrepreneurial parents have entrepreneurial children? What kinds of opportunities will genetic mapping of entrepreneur offer business entities? Do the big Five Personality Traits [Openness, Conscientiousness, Extraversion, Agree-Ableness and Neuroticism] play a role? Do genetic factors influence tendency to acquire skills and develop attributes relevant to entrepreneurship? Is it crucial to emphasize that a genetic predisposition to entrepreneurship should definitely, unequivocally, not be confounded with genetic determinism? This paper aims to seek a conceptualistic conformation to the above issues

Understanding the impact of a residential housing program for people living with HIV/AIDS: results from a realist evaluation

BackgroundIn Canada, the Public Health Agency surveillance of new HIV cases has demonstrated annual increasing rates since 2020. The rates of new HIV cases are highest in the province of...

The Campbellian construct validity: some comments

ABSTRACT The purpose of this paper is to elaborate and comment on some themes related to construct validity, construct choice, and construct respecification in the Campbellian validity system. It is argued that Reichardt’s [(2008, November). An alternative to the Campbellian conceptualization of validity. (Paper presentation). Evaluation 2008 Conference, Denver, CO, United States] assertion that construct validity and external validity are equivalent in the Campbellian system is not correct. The illustrated restricted kind of construct respecification given by Shadish, Cook, and Campbell [(2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin] can be interpreted as part of causal explanation. Their illustrations of general respecifications, on the other hand, should not be considered construct respecifications, but be regarded instead as tentative basic explanations of study results. Choice of research problem or hypothesis – as well as of the included constructs – should be based on practical or theoretical needs in the research field.

Thermodynamic rationale for transformation-induced dislocations in shape memory alloys

Shape Memory Alloys (SMAs) undergo stress-induced martensitic phase-transformation affording a “superelastic” behavior with functional applications. Such behavior is undermined by the formation and accumulation of irreversible residual strains in each cycle. These residual strains arise from transformation-induced dislocation-emission and current understanding has postulated the microstructural role of emitted dislocations to accommodate lattice-mismatch while also observing a preference to occur during reverse-transformation. This study develops a thermodynamic framework to offer a causal explanation for dislocation-emission from Gibbs’ free energy considerations. Superelastic stress-strain curves for a reversible pathway without emitted-dislocations and for an irreversible pathway with emitted-dislocations are derived. The role of emitted-strain in relaxing the transformation-strain of the martensitic-inclusion and in accruing residual strain is proposed. It is shown that both pathways obey the first law of thermodynamics but it is the second law of thermodynamics that dictates the path preference. It is shown that the irreversible path achieves the critical condition for spontaneous reverse-transformation at a higher stress-level than the reversible path. Thus, the irreversible pathway initiates earlier during unloading and is thermodynamically selected during reverse-transformation. The driving-forces associated with the irreversible pathway are analyzed to establish the cause of emitted-dislocations and why it is thermodynamically preferred despite offering a higher lattice-friction barrier. Consequently, a new approach to target fatigue-resistant SMA properties is offered, focusing on the interplay of the individual driving-forces coming from the elastic strain-energy, work-interaction, and lattice-friction, as revealed by the thermodynamic framework.

Supporting Emotional Intelligence, Productivity and Team Goals while Handling Software Requirements Changes

Background: Research shows that emotional intelligence (EI) should be used alongside cognitive intelligence during requirements change (RC) handling in Software Engineering (SE), especially in agile settings. Objective: We wanted to study the role of EI in-depth during RC handling. Method: We conducted a mixed-methods study (an interview study followed by a survey study) with 124 software practitioners. Findings: We found the causal condition, intervening condition and causes lead to key direct consequences of regulating own emotions, managing relationships, and extended consequences of sustaining productivity, setting and sustaining team goals. We found several strategies of supporting EI during RC handling. Further, we found strong correlations between six strategies and one being aware of own emotions, regulating own emotions, sustaining team productivity, and setting and sustaining team goals. Conclusion: Empathising with others and tracking commitments and decisions as a team are key strategies that have strong correlations between managing emotions, between sustaining team productivity, and between setting and sustaining team goals. To the best of our knowledge, the framework we present in this paper is the first theoretical framework on EI in SE research. We provide recommendations for software practitioners to consider during RC handling.

The Impact of Offsetting Internal Attributions on Investor Judgments

While prior research has primarily focused on how self-serving versus incentive-incompatible attributions impact investors’ judgments, we focus on how combining self-serving or incentive-incompatible attributions with offsetting internal attributions impact investors’ judgments. Offsetting internal attributions are causal explanations attributed to internal actions that impact earnings positively (negatively) in an overall negative (positive) news environment. Our first experiment demonstrates that when a firm has overall negative news, highlighting an offsetting, silver-lining attribution is viewed as more credible/forthcoming when it is accompanied by an overall incentive-incompatible versus self-serving attribution. Therefore, highlighting the silver-lining offset improves investors’ earnings estimates only when the management has also been highlighting incentive-incompatible versus self-serving attributions. Our second experiment finds that highlighting an offsetting, blemish attribution together with self-serving attributions increases credibility/forthcomingness assessments more than highlighting self-serving attributions alone. Our study extends research on management attributions and strategic disclosures and has practical implications.

Analysis of undergraduate chemistry students’ responses to substitution reaction mechanisms: a road to mastery

Abstract This study analyzed third-year undergraduate Chemistry major students’ drawings and written explanations of substitution reactions. Seventy (70) students were purposively selected for this study. The main data collection instrument was a diagnostic test and students’ responses were analyzed using deductive coding. The study aimed to unearth students’ conceptual understanding and difficulties on substitution reactions to provide significant insights into improving teaching strategies and learning outcomes. The findings revealed that: 1. Students were more familiar with SN2 reaction mechanisms and could answer questions on SN2 reaction mechanisms better than SN1 reaction mechanisms; 2. Students’ use of ‘chemical vocabulary’ did not translate into an understanding of electron movement and causal mechanistic explanation; 3. About 97 % of the students who gave a correct/partially correct description provided a description of what was happening in the reaction without any further explanation of why the reaction occurred; 4. Students had a slightly better understanding of drawing the correct mechanisms than providing accurate explanations. This study recommends that, in teaching organic reaction mechanisms, instructors should emphasize on electron-pushing formalisms and explain how and why reactions occur to encourage mechanistic thinking in students. Also, students should be given ample practice in organic reaction mechanisms to improve mastery.

Causal inference in ethnographic research: Refining explanations with abductive logic, strength of evidence assessments, and graphical models.

In their classic accounts, anthropological ethnographers developed causal arguments for how specific sociocultural structures and processes shaped human thought, behavior, and experience in particular settings. Despite this history, many contemporary ethnographers avoid establishing in their work direct causal relationships between key variables in the way that, for example, quantitative research relying on experimental or longitudinal data might. As a result, ethnographers in anthropology and other fields have not advanced understandings of how to derive causal explanations from their data, which contrasts with a vibrant "causal revolution" unfolding in the broader social and behavioral sciences. Given this gap in understanding, we aim in the current article to clarify the potential ethnography has for illuminating causal processes related to the cultural influence on human knowledge and practice. We do so by drawing on our ongoing mixed methods ethnographic study of games, play, and avatar identities. In our ethnographic illustrations, we clarify points often left unsaid in both classic anthropological ethnographies and in more contemporary interdisciplinary theorizing on qualitative research methodologies. More specifically, we argue that for ethnographic studies to illuminate causal processes, it is helpful, first, to state the implicit strengths and logic of ethnography and, second, to connect ethnographic practice more fully to now well-developed interdisciplinary approaches to causal inference. In relation to the first point, we highlight the abductive inferential logic of ethnography. Regarding the second point, we connect the ethnographic logic of abduction to what Judea Pearl has called the ladder of causality, where moving from association to intervention to what he calls counterfactual reasoning produces stronger evidence for causal processes. Further, we show how graphical modeling approaches to causal explanation can help ethnographers clarify their thinking. Overall, we offer an alternative vision of ethnography, which contrasts, but nevertheless remains consistent with, currently more dominant interpretive approaches.