Spelling suggestions: "subject:"causal inference"" "subject:"causal cnference""
81 |
A influência da eleição de mulheres na participação política feminina: uma análise no cenário brasileiroHeimann, Natália 17 February 2016 (has links)
Submitted by Natália Heimann (natalia.heimann27@gmail.com) on 2016-03-04T01:15:10Z
No. of bitstreams: 1
Dissertação - Natália Heimann.pdf: 1350748 bytes, checksum: 06d600e20839497fee9017ab5b537ed4 (MD5) / Rejected by Renata de Souza Nascimento (renata.souza@fgv.br), reason: Natália, bom dia
Devido às normas da ABNT, será necessário realizar os seguintes ajustes para que possamos aceitar seu trabalho junto à biblioteca:
Retirar a acentuação do nome Getúlio
Centralizar o título Agradecimentos.
Após alterações, submeter o arquivo novamente.
att on 2016-03-04T11:04:11Z (GMT) / Submitted by Natália Heimann (natalia.heimann27@gmail.com) on 2016-03-05T01:59:25Z
No. of bitstreams: 1
Dissertação - Natália Heimann.pdf: 1351109 bytes, checksum: eabd31fde56a86273e764be7a5a37fa8 (MD5) / Approved for entry into archive by Renata de Souza Nascimento (renata.souza@fgv.br) on 2016-03-07T16:36:26Z (GMT) No. of bitstreams: 1
Dissertação - Natália Heimann.pdf: 1351109 bytes, checksum: eabd31fde56a86273e764be7a5a37fa8 (MD5) / Made available in DSpace on 2016-03-07T16:40:19Z (GMT). No. of bitstreams: 1
Dissertação - Natália Heimann.pdf: 1351109 bytes, checksum: eabd31fde56a86273e764be7a5a37fa8 (MD5)
Previous issue date: 2016-02-17 / Women participation in politics is a persistent subject found in studies and research on gender inequality across multiple social spheres. With a still incipient literature, the current studies of the effect of women´s electoral victory over female political participation in subsequent elections do not focus on the Brazilian case. Through data gathered from the Brazilian Supreme Electoral Court, this study aims to measure the influence of women´s municipal electoral victory over female political filiation in subsequent elections. The methodological framework walks through a regression discontinuity analysis (RDD), whose functionality is to test structural shifts caused by the election of women. The results are dubious and the lack of discernible causal effects fails to validate the main hypothesis. / A participação política das mulheres é tema recorrente nos estudos sobre a desigualdade de gênero em diversas esferas da sociedade. Com uma literatura ainda incipiente, o estudo sobre o efeito que a eleição de mulheres tem sobre o aumento da participação política feminina em eleições subsequentes ainda não tem vertentes com foco no caso brasileiro. O objetivo deste trabalho é, mediante estudo dos dados do Tribunal Superior Eleitoral (TSE), mensurar a influência que a eleição de mulheres para o cargo de prefeito tem sobre a filiação de novas mulheres aos partidos políticos em pleitos subsequentes. O quadro metodológico se desenvolve ao redor das regressões descontínuas (RDD na abreviação em inglês), cuja funcionalidade é testar descontinuidades estruturais que seriam causadas pela eleição de prefeitas. Dentre os resultados, encontramos relações causais dúbias e a ausência de robustez nas análises estatísticas não nos permite tirar conclusões que corroborem a hipótese testada no trabalho.
|
82 |
Econometría de evaluación de impactoGarcía Núñez, Luis 10 April 2018 (has links)
In recent years the program evaluation methods have become very popular in applied microeconomics. However, the variety of these methods responds to specific problems, which are normally determined by the data available and the impact the researcher tries to measure. This paper summarizes the main methods in the current literature, emphasizing the assumptions under which the average treatment effect and the average treatment effect on the treated are identified. Additionally, after each section I briefly present some applications of these methods. This document is a didactic presentation for advanced students in economics and applied researchers who wish to learn the basics of these techniques / En años recientes los métodos de evaluación de impacto se han difundido ampliamente en la investigaciónmicroeconómica aplicada. Sin embargo, la variedad de métodos responde a problemas particulares y específicos los cuales están determinados normalmente por los datos disponibles y el impacto que se busca medir. El presente documento resume las principales corrientes disponibles en la literatura actual, poniendo énfasis en los supuestos bajo los cuales el efecto tratamiento promedio y el efecto tratamiento promedio sobre los tratados se encuentran identificados. Adicionalmente se presentan algunos ejemplos de aplicaciones prácticas de estos métodos. Se busca hacer una presentación didáctica que pueda ser útil a estudiantes avanzados y a investigadores aplicados que busquen conocer los principios básicos de estas técnicas.
|
83 |
Quantification du biais de sélection en sécurité routière : apport de l’inférence causale / Causal inference to quantify selection bias in road traffic safetyDufournet, Marine 01 December 2017 (has links)
Les principaux facteurs de l'insécurité routière sont connus, et l'enjeu réside aujourd'hui dans la mesure de l'effet d'un facteur, et la hiérarchisation de l'ensemble des causes intervenant dans la survenue de l'accident. Toutefois, les données disponibles concernent généralement que des accidentés. En l'absence de non-accidentés, l'épidémiologiste du risque routier se heurte à une sélection extrême. Une des solutions classiques est d'utiliser des analyses en responsabilité, et de mesurer l'effet causal d'un facteur sur le risque d'être responsable d'un accident. Néanmoins, la validité des analyses en responsabilité repose sur l'hypothèse, discutable, que les non-responsables sont représentatifs des circulants. L'objectif de cette thèse est donc de déterminer si les données disponibles d'accidentés permettent de fournir, via les analyses en responsabilité, des estimations des effets causaux sans biais, et notamment sans un biais de sélection résiduel. Nous montrons dans cette thèse que, dès lors que l'inclusion dépend de la gravité de l'accident, et que le facteur étudié a un impact sur la vitesse, il est impossible d'estimer l'effet causal du facteur sur le risque d'être responsable de l'accident grave sans un biais de sélection résiduel. Ce résultat est tout d'abord démontré de manière formelle, grâce à l'utilisation des modèles causaux structuraux. Ces modèles sont fondés sur une structure graphique, le DAG, qui représente les différentes relations entre les variables. Ce DAG permet la description des variables réellement observées, mais également des variables contrefactuelles, variables observables dans un monde contrefactuel où l'on aurait fixé l'exposition à une certaine valeur. L'effet causal étant défini à partir de ces variables contrefactuelles partiellement observées, c'est la structure du DAG qui permet de déterminer si l'effet causal peut être estimé en fonction des variables observées. Or, la structure du DAG conduisant à la survenue d'un accident grave ne permet pas d'exprimer l'effet causal du facteur étudié sur la responsabilité de l'accident grave en fonction des distributions observées sur les accidentés graves. Conditionner les estimations sur les accidentés graves correspond à ajuster sur une variable du DAG appelée « collider », et ainsi à introduire un biais dit de collision. En générant un modèle relativement simple, nous donnons à nos résultats théoriques une illustration numérique. En effet, lorsque les données ne dépendent pas de la gravité de l'accident, ou que le facteur étudié n'a pas d'effet sur la vitesse, la mesure estimable à partir des analyses en responsabilité est une mesure sans biais de l'effet causal, sous certaines hypothèses de prévalences faibles. Lorsque l'inclusion dépend de la gravité de l'accident, il existe un biais et ce biais induit par les analyses en responsabilité est d'autant plus grand que l'intensité de la relation entre le facteur et la vitesse, et celle entre la vitesse et l'accident est grand. Les schémas d'étude présentés permettent d'approcher des situations où le facteur étudié serait l'alcool ou le cannabis. Dans le cas de l'alcool, il apparait que sous le modèle simple considéré, la mesure d'association estimable serait une sous-estimation de l'effet causal. En revanche, dans le cas du cannabis, la mesure d'association correspondrait à une sur-estimation de l'effet causal. D'autre part, les outils de l'inférence causale nous ont permis de fournir une description formelle de la validité externe et interne, ainsi qu'une description formelle de la mesure d'association estimable via les analyses en responsabilité. Cette question de la validité interne d'une mesure se pose dans d'autres champs d'application que la sécurité routière. Elle se pose notamment dans le cas du paradoxe de l'obésité [etc...] / Many factors associated with the risk and severity of road accidents are now widely considered as causal : alcohol, speed, usage of a mobile phone... Therefore, questions asked by decision-makers now mostly concern the magnitude of their causal effects, as well as the burden of deaths or victims attributable to these various causes of accident. One particularity of road safety epidemiology is that available data generally describe drivers and vehicles involved in road accidents only, or even severe road accidents only. This extreme selection precludes the estimation of causal effects. To circumvent this absence of « control » population of non-crash involved drivers, it is common to use responsibility analysis and to assess the causal effect of a given factor on the risk of being responsible for an accident among involved drivers. The underlying assumption is that non-responsible drivers represent a random sample of the general driving population that was « selected » to crash by circumstances beyond their control and therefore have the same risk factor profile as other drivers on the road at the same time. However, this randomness assumption is questionable. The objective of this thesis is to determine whether available data in road safety allow us to assess causal effects on responsibility without a residual selection bias. We show that a good approximation of causal effect of a given factor on the risk of being responsible is possible only if the inclusion into the dataset does not depend on the severity of the accident, or if the given factor has no effect on speed. This result is shown by using the Structural Causal Model (SCM) framework. The SCM framework is based on a causal graph : the DAG (directed acyclic graph), which represents the relationships among variables. The DAG allows the description of what we observe in the actual world, but also what we would have observed in counterfactual worlds, if we could have intervened and forced the exposure to be set to a given level. Causal effects are then defined by using counterfactual variables, and it is the DAG’s structure which determines whether causal effects are identifiable, or recoverable, and estimable from the distribution of observed variables. However, the assumptions embedded in the DAG which describes the occurence of a severe accident does not ensure that a causal odds ratios is expressible in terms of the observable distribution. Conditioning the estimations on involved drivers in a severe crash correspond to conditioning on a variable in the DAG called « collider », and to create a « collider bias ». We present numerical results to illustrate our theoretical arguments and the magnitude of the bias between the estimable association measure and some causal effects. Under the simple generative model considered, we show that, when the inclusion depends on the severity of the accident, the bias between the estimable association measure and causal effect is larger than the relation between the exposure and speed, or speed and the occurrence of a severe accident is strong. Moreover, the presented designs allow us to describe some situations where the exposure could be alcohol or cannabis intoxication. In the case of alcohol, where alcohol and speed are positively correlated, the estimable associational effect underestimates the causal effect. In the case of cannabis, where cannabis and speed are negatively correlated, the estimable associational effect overestimates the causal effect. On the other hand, we provide a formal definition of internal and external validity, and a counterfactual interpretation of the estimable quantity in the presence of selection bias, when causal effects are not recoverable. This formal interpretation of the estimable quantity in the presence of selection bias is not only useful in the context of responsibility analyses. It is for instance useful to explain the obesity paradox
|
84 |
Analyse de la prise en charge des patients traumatisés sévères dans le contexte français : processus de triage et processus de soin / Analysis of Severe Trauma Patients Management : Triage and Care ProcessHamada, Sophie Rym 20 December 2019 (has links)
La traumatologie est un problème de santé publique au troisième rang des années de vie perdues ajustées sur l’incapacité en France. L’investissement sanitaire et le volume de recherche qu’elle génère sont en deçà de ce que représente son impact sociétal. L’objet de ce travail de recherche était de plonger au cœur du parcours du patient traumatisé sévère pour en cibler trois problématiques clefs et tenter de répondre aux interrogations qu’elles génèrent.Les données utilisées provenaient essentiellement d’un observatoire de traumatologie lourde hospitalier (Traumabase®), régional et national, qui collige un ensemble de variables épidémiologiques, cliniques, paracliniques, et thérapeutiques des patients traumatisés sévères admis en centre de traumatologie.Le premier projet a ciblé l’orientation initiale (triage) des patients traumatisés sévères suite à un accident de la circulation au sein de la région Île de France et son effet sur la mortalité. Les patients initialement mal triés, transférés secondairement dans les centres de traumatologie régionaux, ne présentaient pas un pronostic plus sombre que les patients qui étaient transportés directement. Le système de soin dans son ensemble permettait de leur assurer un devenir équivalent. Une analyse en population réalisée par un chainage probabiliste des données avec les fiches d’accident de l’observatoire national de la sécurité routière a permis d’approcher le taux de sous triage conduisant au décès dans la région (0,15%) et d’objectiver que 60% des décès survenaient avant toute admission hospitalière.Le second projet visait l’optimisation de la jonction entre l’équipe médicalisée préhospitalière et l’équipe intrahospitalière. Il s’est attelé à développer un outil de prédiction de la sévérité des patients hémorragiques pour permettre l’anticipation de l’admission des patients les plus graves. Cet outil, le Red Flag, avait pour cahier des charges d’être simple et pragmatique, et de ne pas nécessiter de dispositif externe pour l’utiliser. Il a identifié cinq caractéristiques (shock index>1, pression artérielle moyenne <70mmHg, hémoglobine capillaire < 13g/dL, bassin instable et intubation), dont la présence de deux ou plus d’entre-elles permettait d’activer l’alerte pour l’hôpital receveur. Cet outil devra être évalué en prospectif pour confirmer ses performances et évaluer son impact sur l’organisation et le devenir des patients.Le troisième projet de recherche ciblait plus spécifiquement une des thérapeutiques de la coagulopathie aigue du traumatisé sévère en choc hémorragique. Il a tenté de quantifier l’impact de l’administration de concentré de fibrinogène à la phase précoce du choc hémorragique traumatique (6 premières heures) sur la mortalité toutes causes confondues des 24 premières heures par une approche d’inférence causale (score de propension et méthode d’estimation double robuste). Il n’a pas été retrouvé d’effet significatif sur la mortalité, un manque de puissance pouvant être responsable de ce résultat (différence de risque observée : -0,031, Intervalle de confiance 95% [-0,084 ; 0,021]).Ainsi l’ensemble de ces 3 projets de recherche ont permis de répondre à des problématiques ciblées du parcours du patient traumatisé sévère, générant par la même de nouvelles perspectives d’analyse pour mieux circonscrire les réponses de terrain. / In France, the third most frequent cause of disability adjusted life years lost is trauma, an observation that makes trauma a public health challenge. However, investment in trauma care and specific research fails to meet this challenge and to acknowledge the associated societal and economic impact.The purpose of this research was to explore the core of the pathway of a major trauma patient and bring to light key issues and question and to find answers. The data used in this research were mainly extracted from a regional and national trauma registry, the Traumabase®. The registry collects epidemiological, clinical, paraclinical and therapeutic variables for patients with severe trauma admitted to participating trauma centres. The first project focused on the effects of triage on patients with severe trauma following a road traffic accident in the Ile de France region. Patients who were initially under triaged and then transferred to regional trauma centres did not have a worse prognosis than patients who were transported directly. The emergency medical system as a whole ensured that they would have an equivalent outcome. A population analysis carried out by a probabilistic data chainage using the accident records of the National Road Safety Observatory made it possible to approach the undertriage rate leading to death in the region (0.15%) and to reveal that 60% of deaths occurred before any hospital admission. The second project developed a pragmatic pre-alert tool based on simple, clinical prehospital criteria to predict acute hemorrhage in trauma patients. This tool is meant to increase the performance of the receiving hospital trauma team of these critically sick patients and activate a specific hemorrhage pathway. The study identified five variables (shock index>1, mean blood pressure <70mmHg, capillary hemoglobin <13g/dL, unstable pelvis and intubation). If two or more variables were present, the tool identified patient with acute hemorrhage and the corresponding pathway should be activated. This tool requires prospective validation and assessment of its impact on care provision and patient outcome.The third research project focused on a therapeutic component of trauma induced coagulopathy. The study attempted to quantify the effect of fibrinogen concentrate administration at the early phase of traumatic hemorrhagic shock (first 6 hours) on 24 hours all-cause mortality using a causal inference approach (propensity score and double robust estimator). The research did not demonstrate any impact on mortality (observed risk difference: -0.031, 95% confidence interval [-0.084; 0.021]); a lack of power might be responsible for this result.
|
85 |
Effects of MIFID II on Stock Trade Volumes of Nasdaq Stockholm / MIFID II- Effekter på Nasdaq Stockholms Handlade AktievolymerElling, Eva January 2019 (has links)
Introducing new financial legislation to financial markets require caution to achieve the intended outcome. This thesis aims to investigate whether or not the newly installed revised Markets in Financial Instruments Directive- the MIFID II regulation - temporally influenced the trading stock volume levels of Nasdaq Stockholm during its introduction to the Swedish stock market. A first approach of a generalized Negative Binomial model is carried out on aggregated data, followed by an individual Fixed Effects model in an attempt to eliminate omitted variable bias caused by missing unobserved variables for the individual stocks. The aggregated data is attained by taking the equally weighted average of the trading volume and adjusting for seasonality through Seasonal and Trend decomposition using Loess in combination with a regression model with ARIMA errors to mitigate calendar effects. Due to robustness of the aggregated data, the Negative Binomial model manage to capture significant effects of the regulation on the Small Cap. segment, even though clusters of the data show signs of divergent reactions to MIFID II. Since the Fixed Effects model operate on non-aggregated TSCS data and because of the varying effects on each stock the Fixed Effect model fails in its attempt to do the same. / Implementation av nya finansiella regelverk på finansmarknaden kräver aktsamhet för att uppnå de tilltänka målen. Det här arbetet undersöker huruvida MIFID II regleringen orsakade en temporär medelvärdesskiftning av de handlade aktievolymerna på Nasdaq Stockholm under regelverkets introduktion på den svenska marknaden. Först testas en generaliserad Negative Binomial regression applicerat på aggregerad data, därefter en individuell Fixed Effects modell för att försöka eliminera fel på grund av saknade, okända variabler. Det aggrigerade datasettet erhålls genom att ta genomsnittet av handelsvolymerna och justera dessa för sässongsmässiga mönster med metoden STL i kombination med regression med ARIMA residualer för att även ta hänsyn till kalender relaterade effekter. Eftersom den aggrigerade datan är robust lyckas the Negative Binomial regressionen fånga signifikanta effekter av regleringen för Small Cap. segmentet trots att datat uppvisar tecken på att subgrupper inom segmentet reagerat väldigt olika på den nya regleringen. Eftersom Fixed Effects modellen är applicerad på icke-aggrigerad TSCS data och pågrund av den varierande effekten på de individuella aktierna lyckas inte denna modell med detta.
|
86 |
Gated Single Assignment Form Partnered with Value-Based Statistical Fault Localization for Numerical Java ProgramsTraben, Oliver 26 May 2023 (has links)
No description available.
|
87 |
Causal Inference on Tactical Simulations using Bayesian Structure LearningLagerkvist Blomqvist, Karl January 2022 (has links)
This thesis explores the possibility of using Bayesian Structure Learning and Do-Calculus to perform causal inference on data from tactical combat simulations provided by Saab. A four-step approach is considered whose first step is to find a Bayesian Network from the data using Bayesian Structure Learning and Probability Distribution Fitting. These Bayesian Networks describe a set of conditional independencies ambiguously. This ambiguity gives rise to a set of feasible Structural Causal Models that describes feasible causal relationships in the data. The approach then continues in its second step by selecting at least one of these Structural Causal Models that can be utilized for performing causal inference using Do-Calculus and Probabilistic Inference in the approach’s third and fourth steps respectively. The thesis concludes that there exist several difficulties with the approach that together with a lack of a methodology for error estimation reduces the method’s reliability. The recommendation is thus to consider the possibility of performing randomized controlled experiments using the tactical simulator before continuing the development of this approach. / Det här examensarbetet utforskar möjligheten att använda Bayesiansk Strukturinlärning och Do-Calculus för att utföra Kausal Inferens på data från taktiska stridsimuleringar framtagna av Saab. En fyrastegsmetod beaktas vars första steg är att hitta ett Bayesiansk Nätverk genom användandet av Bayesiansk Strukturinlärning och Sannolikhetsfördelnings-anpassning. Dessa Bayesianska Nätverk beskriver en mängd betingade oberoendet i datamängden på ett icke-entydligt sett. Denna icke-entydlighet ger upphov till en mängd av möjliga Strukturella Kausala Modeller som beskriver möjliga kausala strukturer i datamängden. Metodens andra steg fortsätter med att välja minst en av dessa Strukturella Kausala Modeller som kan användas för att åstakomma Kausal Inferens med hjälp av Do-Calculus och Stokastisk Inferens i metodens tredje respektive fjärde steg. Slutsatsen från examensarbetet är att det finns ett flertal svårigheter med metoden som tillsamans med en avsaknad av en feluppskattningsmetodik minskar metodens tillförlitlighet. Rekommendationen är därför att undersöka möjligheten att genomföra kontrollerade slumpmässiga experiment innan metodiken vidareutvecklas.
|
88 |
Family Behavior and Children’s Wellbeing: Statistical Modeling and Measurement IssuesRodríguez Sánchez, Alejandra 30 June 2023 (has links)
In dieser Dissertation gehe ich auf verschiedene statistische Modellierungs- und Messprobleme ein, die eine kausale Interpretation der in der Literatur zu Familiensoziologie und sozialer Ungleichheit gefundenen Zusammenhängen erschweren. Erstens legt die Lebensverlaufsforschung nahe, dass das Problem der Verzerrung durch Selektion in der Literatur über die Abwesenheit von Vätern komplexer sein könnte als angenommen. Durch die Korrektur von dynamischen Verzerrungen wird die Schätzung des kausalen Effektes der Abwesenheit des Vaters auf das Wohlergehen der Kinder reduziert. Zweitens wird angenommen, dass familiäre Instabilität in der Kindheit das Wohlbefinden der Kinder negativ beeinflusst. Allerdings könnten zeitabhängige konfundierende Faktoren, die durch vergangene Episoden familiärer Instabilität beeinflusst werden und sich auf die künftige Stabilität der Familie auswirken, einen Teil der angenommenen negativen Auswirkungen erklären. Ich zeige, dass eine dynamische Version der Selektionshypothese eine wesentliche Rolle bei der Entkräftung der Hypothese der familiären Instabilität spielt. Drittens deuten die Forschungsergebnisse darauf hin, dass die soziale Stratifizierung bei den Sprachkenntnissen von Vorschulkindern durch Eingriffe in den Erziehungsstil von Eltern mit wenig Ressourcen erheblich verringert werden könnten. Mit Hilfe einer kausalen Mediationsanalyse zeige ich, dass die elterliche Erziehung nur etwa ein Drittel des Gesamteffekts des sozioökonomischen Status auf die frühen Sprachfähigkeiten mediieren. Viertens wird die Messung kognitiver Fähigkeiten durch verschiedene Merkmale standardisierter Beurteilungen erschwert. Diese Probleme haben wichtige Auswirkungen auf die Quantifizierung sozialer Ungleichheit bei unbeobachtbaren Variablen und auf die Forschung zu kausalen Effekten. Die Dissertation schließt mit einem Plädoyer zur rigoroseren Anwendung von Methoden der kausalen Inferenz in Familiensoziologie und Forschung zu sozialer Ungleichheit. / In this dissertation, I consider various statistical modeling and measurement issues that complicate the causal attributions made about those associations in the literature in family sociology and social inequality. First, life course informed research suggests that the problem of selection bias in the father absence literature may be more complex than currently thought. After adjusting for dynamic biases, estimates of father absence's effect on children's wellbeing are reduced. Second, family instability experienced during childhood is said to negatively affect children's wellbeing. However, time-dependent confounders affected by past episodes of family instability and affecting future family stability might explain away part of the negative impact. I show that a dynamic version of the selection hypothesis counters the family instability hypothesis, and the effects of cumulative family instability are small and not consistent with the family instability hypothesis. Third, research suggest that socioeconomic status gaps in language skills among preschoolers could be substantially reduced by intervening on the parenting styles, practices, and parental investments of low-resource parents. Employing interventional causal mediation analysis, however, I show parenting mediates around one third of the total effect of SES on early language skills. Fourth, the measurement of cognitive abilities is complicated by various features of standardized assessments. Those problems have important implications for the quantification of social inequality in unobservable variables and for causal inference research because test scores capture non-random noise. The dissertation concludes by making a plea for furthering causal inference thinking in family sociology, social inequality, social mobility, and family demography research.
|
89 |
Essays on using machine learning for causal inferenceJacob, Daniel 01 March 2022 (has links)
Um Daten am effektivsten zu nutzen, muss die moderne Ökonometrie ihren Werkzeugkasten an Modellen erweitern und neu denken. Das Feld, in dem diese Transformation am besten beobachtet werden kann, ist die kausale Inferenz.
Diese Dissertation verfolgt die Absicht Probleme zu untersuchen, Lösungen zu präsentieren und neue Methoden zu entwickeln Machine Learning zu benutzen, um kausale Parameter zu schätzen. Dafür werden in der Dissertation zuerst verschiedene neuartige Methoden, welche als Ziel haben heterogene Treatment Effekte zu messen, eingeordnet. Im zweiten Schritt werden, basierend auf diesen Methoden, Richtlinien für ihre Anwendung in der Praxis aufgestellt. Der Parameter von Interesse ist der „conditional average treatment effect“ (CATE). Es kann gezeigt werden, dass ein Vergleich mehrerer Methoden gegenüber der Verwendung einer einzelnen Methode vorzuziehen ist. Ein spezieller Fokus liegt dabei auf dem Aufteilen und Gewichten der Stichprobe, um den Verlust in Effizienz wettzumachen. Ein unzulängliches Kontrollieren für die Variation durch verschiedene Teilstichproben führt zu großen Unterschieden in der Präzision der geschätzten Parameter. Wird der CATE durch Bilden von Quantilen in Gruppen unterteilt, führt dies zu robusteren Ergebnissen in Bezug auf die Varianz.
Diese Dissertation entwickelt und untersucht nicht nur Methoden für die Schätzung der Heterogenität in Treatment Effekten, sondern auch für das Identifizieren von richtigen Störvariablen. Hierzu schlägt diese Dissertation sowohl die „outcome-adaptive random forest“ Methode vor, welche automatisiert Variablen klassifiziert, als auch „supervised randomization“ für eine kosteneffiziente Selektion der Zielgruppe. Einblicke in wichtige Variablen und solche, welche keine Störung verursachen, ist besonders in der Evaluierung
von Politikmaßnahmen aber auch im medizinischen Sektor wichtig, insbesondere dann, wenn kein randomisiertes Experiment möglich ist. / To use data effectively, modern econometricians need to expand and rethink their toolbox. One field where such a transformation has already started is causal inference. This thesis aims to explore further issues, provide solutions, and develop new methods on how machine learning can be used to estimate causal parameters. I categorize novel methods to estimate heterogeneous treatment effects and provide a practitioner’s guide for implementation. The parameter of interest is the conditional average treatment effect (CATE). It can be shown that an ensemble of methods is preferable to relying on one method. A special focus, with respect to the CATE, is set on the comparison of such methods and the role of sample splitting and cross-fitting to restore efficiency. Huge differences in the estimated parameter accuracy can occur if the sampling uncertainty is not correctly accounted for. One feature of the CATE is a coarser representation through quantiles. Estimating groups of the CATE leads to more robust estimates with respect to the sampling uncertainty and the resulting high variance.
This thesis not only develops and explores methods to estimate treatment effect heterogeneity but also to identify confounding variables as well as observations that should receive treatment. For these two tasks, this thesis proposes the outcome-adaptive random forest for automatic variable selection, as well as supervised randomization for a cost-efficient selection of the target group. Insights into important variables and those that are not true confounders are very helpful for policy evaluation and in the medical sector when randomized control trials are not possible.
|
90 |
The Indirect Effects of Mediation: A Dynamic Model of Mediation and ConflictSchricker, Ezra 31 October 2016 (has links)
No description available.
|
Page generated in 0.061 seconds