In 2022, the Journal of Statistics and Data Science Education (JSDSE) instituted augmented requirements for authors to post deidentified data and code underlying their papers. These changes were prompted by an increased focus on reproducibility and open science (NASEM 2019). A recent review of data availability practices noted that "such policies help increase the reproducibility of the published literature, as well as make a larger body of data available for reuse and re-analysis" (PLOS ONE, 2024). JSDSE values accessibility as it endeavors to share knowledge that can improve educational approaches to teaching statistics and data science. Because institution, environment, and students differ across readers of the journal, it is especially important to facilitate the transfer of a journal article's findings to new contexts. This process may require digging into more of the details, including the deidentified data and code. Our goal is to provide our readers and authors with a review of why the requirements for code and data sharing were instituted, summarize ongoing trends and developments in open science, discuss options for data and code sharing, and share advice for authors.
{"title":"Guidelines and Best Practices to Share Deidentified Data and Code","authors":"Nicholas J. Horton, Sara Stoudt","doi":"arxiv-2405.18232","DOIUrl":"https://doi.org/arxiv-2405.18232","url":null,"abstract":"In 2022, the Journal of Statistics and Data Science Education (JSDSE)\u0000instituted augmented requirements for authors to post deidentified data and\u0000code underlying their papers. These changes were prompted by an increased focus\u0000on reproducibility and open science (NASEM 2019). A recent review of data\u0000availability practices noted that \"such policies help increase the\u0000reproducibility of the published literature, as well as make a larger body of\u0000data available for reuse and re-analysis\" (PLOS ONE, 2024). JSDSE values\u0000accessibility as it endeavors to share knowledge that can improve educational\u0000approaches to teaching statistics and data science. Because institution,\u0000environment, and students differ across readers of the journal, it is\u0000especially important to facilitate the transfer of a journal article's findings\u0000to new contexts. This process may require digging into more of the details,\u0000including the deidentified data and code. Our goal is to provide our readers\u0000and authors with a review of why the requirements for code and data sharing\u0000were instituted, summarize ongoing trends and developments in open science,\u0000discuss options for data and code sharing, and share advice for authors.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141171713","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
It is often asserted that to control for the effects of confounders, one should include the confounding variables of concern in a statistical model as a covariate. Conversely, it is also asserted that control can only be concluded by design, where the results from an analysis can only be interpreted as evidence of an effect because the design controlled for the cause. To suggest otherwise is said to be a fallacy of cum hoc ergo propter hoc. Obviously, these two assertions create a conundrum: How can the effect of confounder be controlled for with analysis instead of by design without committing cum hoc ergo propter hoc? The present manuscript answers this conundrum.
人们通常认为,要控制混杂因素的影响,就应在统计模型中将相关的混杂变量作为一个变量。反之,也有人断言,只有通过设计才能得出控制的结论,即由于设计控制了原因,分析结果只能被解释为效果的证据。反之,则是 "既成事实"(cum hoc ergo propter hoc)的谬误。很明显,这两个论断造成了一个难题:如何通过分析而不是设计来控制混杂因素的影响,而又不犯兼有因果关系的谬误?本手稿回答了这一难题。
{"title":"The Epistemology behind Covariate Adjustment","authors":"Grayson L. Baird, Stephen L. Bieber","doi":"arxiv-2405.17224","DOIUrl":"https://doi.org/arxiv-2405.17224","url":null,"abstract":"It is often asserted that to control for the effects of confounders, one\u0000should include the confounding variables of concern in a statistical model as a\u0000covariate. Conversely, it is also asserted that control can only be concluded\u0000by design, where the results from an analysis can only be interpreted as\u0000evidence of an effect because the design controlled for the cause. To suggest\u0000otherwise is said to be a fallacy of cum hoc ergo propter hoc. Obviously, these\u0000two assertions create a conundrum: How can the effect of confounder be\u0000controlled for with analysis instead of by design without committing cum hoc\u0000ergo propter hoc? The present manuscript answers this conundrum.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141173143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The 2021 Nobel Prize in Economics recognized a theory of causal inference, which deserves more attention from philosophers. To that end, I develop a dialectic that extends the Lewis-Stalnaker debate on a logical principle called Conditional Excluded Middle (CEM). I first play the good cop for CEM, and give a new argument for it: a Quine-Putnam indispensability argument based on the Nobel-Prize winning theory. But then I switch sides and play the bad cop: I undermine that argument with a new theory of causal inference that preserves the success of the original theory but dispenses with CEM.
2021 年诺贝尔经济学奖认可了一种因果推理理论,该理论值得哲学家们更多关注。为此,我提出了一个辩证法,扩展了刘易斯-斯塔尔纳克关于一个名为 "条件排除中间"(CEM)的逻辑原则的辩论。我首先为 CEM 扮演了一个好警察的角色,并为它提供了一个新的论证:一个基于诺贝尔奖获奖理论的奎因-普特南不可或缺性论证。但随后,我又换了一边,扮演了坏警察的角色:我用一种新的因果推理理论来破坏这一论证,这种理论保留了原有理论的成功之处,但却摒弃了CEM。
{"title":"The Logic of Counterfactuals and the Epistemology of Causal Inference","authors":"Hanti Lin","doi":"arxiv-2405.11284","DOIUrl":"https://doi.org/arxiv-2405.11284","url":null,"abstract":"The 2021 Nobel Prize in Economics recognized a theory of causal inference,\u0000which deserves more attention from philosophers. To that end, I develop a\u0000dialectic that extends the Lewis-Stalnaker debate on a logical principle called\u0000Conditional Excluded Middle (CEM). I first play the good cop for CEM, and give\u0000a new argument for it: a Quine-Putnam indispensability argument based on the\u0000Nobel-Prize winning theory. But then I switch sides and play the bad cop: I\u0000undermine that argument with a new theory of causal inference that preserves\u0000the success of the original theory but dispenses with CEM.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141147813","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Benjamin Williams, Erin M. Schliep, Bailey Fosdick, Ryan Elmore
Team and player evaluation in professional sport is extremely important given the financial implications of success/failure. It is especially critical to identify and retain elite shooters in the National Basketball Association (NBA), one of the premier basketball leagues worldwide because the ultimate goal of the game is to score more points than one's opponent. To this end we propose two novel basketball metrics: "expected points" for team-based comparisons and "expected points above average (EPAA)" as a player-evaluation tool. Both metrics leverage posterior samples from Bayesian hierarchical modeling framework to cluster teams and players based on their shooting propensities and abilities. We illustrate the concepts for the top 100 shot takers over the last decade and offer our metric as an additional metric for evaluating players.
{"title":"Expected Points Above Average: A Novel NBA Player Metric Based on Bayesian Hierarchical Modeling","authors":"Benjamin Williams, Erin M. Schliep, Bailey Fosdick, Ryan Elmore","doi":"arxiv-2405.10453","DOIUrl":"https://doi.org/arxiv-2405.10453","url":null,"abstract":"Team and player evaluation in professional sport is extremely important given\u0000the financial implications of success/failure. It is especially critical to\u0000identify and retain elite shooters in the National Basketball Association\u0000(NBA), one of the premier basketball leagues worldwide because the ultimate\u0000goal of the game is to score more points than one's opponent. To this end we\u0000propose two novel basketball metrics: \"expected points\" for team-based\u0000comparisons and \"expected points above average (EPAA)\" as a player-evaluation\u0000tool. Both metrics leverage posterior samples from Bayesian hierarchical\u0000modeling framework to cluster teams and players based on their shooting\u0000propensities and abilities. We illustrate the concepts for the top 100 shot\u0000takers over the last decade and offer our metric as an additional metric for\u0000evaluating players.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141147846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Despite their cost, randomized controlled trials (RCTs) are widely regarded as gold-standard evidence in disciplines ranging from social science to medicine. In recent decades, researchers have increasingly sought to reduce the resource burden of repeated RCTs with factorial designs that simultaneously test multiple hypotheses, e.g. experiments that evaluate the effects of many medications or products simultaneously. Here I show that when multiple interventions are randomized in experiments, the effect any single intervention would have outside the experimental setting is not identified absent heroic assumptions, even if otherwise perfectly realistic conditions are achieved. This happens because single-treatment effects involve a counterfactual world with a single focal intervention, allowing other variables to take their natural values (which may be confounded or modified by the focal intervention). In contrast, observational studies and factorial experiments provide information about potential-outcome distributions with zero and multiple interventions, respectively. In this paper, I formalize sufficient conditions for the identifiability of those isolated quantities. I show that researchers who rely on this type of design have to justify either linearity of functional forms or -- in the nonparametric case -- specify with Directed Acyclic Graphs how variables are related in the real world. Finally, I develop nonparametric sharp bounds -- i.e., maximally informative best-/worst-case estimates consistent with limited RCT data -- that show when extrapolations about effect signs are empirically justified. These new results are illustrated with simulated data.
{"title":"Identification of Single-Treatment Effects in Factorial Experiments","authors":"Guilherme Duarte","doi":"arxiv-2405.09797","DOIUrl":"https://doi.org/arxiv-2405.09797","url":null,"abstract":"Despite their cost, randomized controlled trials (RCTs) are widely regarded\u0000as gold-standard evidence in disciplines ranging from social science to\u0000medicine. In recent decades, researchers have increasingly sought to reduce the\u0000resource burden of repeated RCTs with factorial designs that simultaneously\u0000test multiple hypotheses, e.g. experiments that evaluate the effects of many\u0000medications or products simultaneously. Here I show that when multiple\u0000interventions are randomized in experiments, the effect any single intervention\u0000would have outside the experimental setting is not identified absent heroic\u0000assumptions, even if otherwise perfectly realistic conditions are achieved.\u0000This happens because single-treatment effects involve a counterfactual world\u0000with a single focal intervention, allowing other variables to take their\u0000natural values (which may be confounded or modified by the focal intervention).\u0000In contrast, observational studies and factorial experiments provide\u0000information about potential-outcome distributions with zero and multiple\u0000interventions, respectively. In this paper, I formalize sufficient conditions\u0000for the identifiability of those isolated quantities. I show that researchers\u0000who rely on this type of design have to justify either linearity of functional\u0000forms or -- in the nonparametric case -- specify with Directed Acyclic Graphs\u0000how variables are related in the real world. Finally, I develop nonparametric\u0000sharp bounds -- i.e., maximally informative best-/worst-case estimates\u0000consistent with limited RCT data -- that show when extrapolations about effect\u0000signs are empirically justified. These new results are illustrated with\u0000simulated data.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141062584","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Carlos del-Castillo-Negrete, Rylan Spence, Troy Butler, Clint Dawson
We present a novel method for generating sequential parameter estimates and quantifying epistemic uncertainty in dynamical systems within a data-consistent (DC) framework. The DC framework differs from traditional Bayesian approaches due to the incorporation of the push-forward of an initial density, which performs selective regularization in parameter directions not informed by the data in the resulting updated density. This extends a previous study that included the linear Gaussian theory within the DC framework and introduced the maximal updated density (MUD) estimate as an alternative to both least squares and maximum a posterior (MAP) estimates. In this work, we introduce algorithms for operational settings of MUD estimation in real or near-real time where spatio-temporal datasets arrive in packets to provide updated estimates of parameters and identify potential parameter drift. Computational diagnostics within the DC framework prove critical for evaluating (1) the quality of the DC update and MUD estimate and (2) the detection of parameter value drift. The algorithms are applied to estimate (1) wind drag parameters in a high-fidelity storm surge model, (2) thermal diffusivity field for a heat conductivity problem, and (3) changing infection and incubation rates of an epidemiological model.
我们提出了一种在数据一致(DC)框架内生成序列参数估计并量化动态系统中认识不确定性的新方法。数据一致性框架不同于传统的贝叶斯方法,因为它结合了初始密度的前推,在参数方向上进行选择性正则化,而在更新后的密度中,数据并未提供相关信息。这项研究扩展了之前的研究,将线性高斯理论纳入了 DC 框架,并引入了最大更新密度(MUD)估计,作为最小二乘法和最大后验(MAP)估计的替代方法。在这项工作中,我们介绍了 MUD 估计的实际或接近实时的操作设置算法,在这种情况下,空间-时间数据包以数据包的形式到达,以提供参数的更新估计并识别潜在的参数漂移。DC 框架内的计算诊断对于评估 (1) DC 更新和 MUD 估计的质量以及 (2) 参数值漂移的检测至关重要。这些算法被应用于估算:(1) 高保真风暴潮模型中的风阻参数;(2) 热传导问题中的热扩散场;(3) 流行病学模型中不断变化的感染率和潜伏率。
{"title":"Sequential Maximal Updated Density Parameter Estimation for Dynamical Systems with Parameter Drift","authors":"Carlos del-Castillo-Negrete, Rylan Spence, Troy Butler, Clint Dawson","doi":"arxiv-2405.08307","DOIUrl":"https://doi.org/arxiv-2405.08307","url":null,"abstract":"We present a novel method for generating sequential parameter estimates and\u0000quantifying epistemic uncertainty in dynamical systems within a data-consistent\u0000(DC) framework. The DC framework differs from traditional Bayesian approaches\u0000due to the incorporation of the push-forward of an initial density, which\u0000performs selective regularization in parameter directions not informed by the\u0000data in the resulting updated density. This extends a previous study that\u0000included the linear Gaussian theory within the DC framework and introduced the\u0000maximal updated density (MUD) estimate as an alternative to both least squares\u0000and maximum a posterior (MAP) estimates. In this work, we introduce algorithms\u0000for operational settings of MUD estimation in real or near-real time where\u0000spatio-temporal datasets arrive in packets to provide updated estimates of\u0000parameters and identify potential parameter drift. Computational diagnostics\u0000within the DC framework prove critical for evaluating (1) the quality of the DC\u0000update and MUD estimate and (2) the detection of parameter value drift. The\u0000algorithms are applied to estimate (1) wind drag parameters in a high-fidelity\u0000storm surge model, (2) thermal diffusivity field for a heat conductivity\u0000problem, and (3) changing infection and incubation rates of an epidemiological\u0000model.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140941397","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We explore whether the human ratings of open ended responses can be explained with non-content related features, and if such effects vary across different mathematics-related items. When scoring is rigorously defined and rooted in a measurement framework, educators intend that the features of a response which are indicative of the respondent's level of ability are contributing to scores. However, we find that features such as response length, a grammar score of the response, and a metric relating to key phrase frequency are significant predictors for response ratings. Although our findings are not causally conclusive, they may propel us to be more critical of he way in which we assess open ended responses, especially in high stakes scenarios. Educators take great care to provide unbiased, consistent ratings, but it may be that extraneous features unrelated to those which were intended to be rated are being evaluated.
{"title":"Predicting Short Response Ratings with Non-Content Related Features: A Hierarchical Modeling Approach","authors":"Aubrey Condor","doi":"arxiv-2405.08574","DOIUrl":"https://doi.org/arxiv-2405.08574","url":null,"abstract":"We explore whether the human ratings of open ended responses can be explained\u0000with non-content related features, and if such effects vary across different\u0000mathematics-related items. When scoring is rigorously defined and rooted in a\u0000measurement framework, educators intend that the features of a response which\u0000are indicative of the respondent's level of ability are contributing to scores.\u0000However, we find that features such as response length, a grammar score of the\u0000response, and a metric relating to key phrase frequency are significant\u0000predictors for response ratings. Although our findings are not causally\u0000conclusive, they may propel us to be more critical of he way in which we assess\u0000open ended responses, especially in high stakes scenarios. Educators take great\u0000care to provide unbiased, consistent ratings, but it may be that extraneous\u0000features unrelated to those which were intended to be rated are being\u0000evaluated.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140941398","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Instrumental variables (IV) are a commonly used tool to estimate causal effects from non-randomized data. A prototype of an IV is a randomized trial with non-compliance where the randomized treatment assignment serves as an IV for the non-ignorable treatment received. Under a monotonicity assumption, a valid IV non-parametrically identifies the average treatment effect among a non-identifiable complier subgroup, whose generalizability is often under debate. In many studies, there could exist multiple versions of an IV, for instance, different nudges to take the same treatment in different study sites in a multi-center clinical trial. These different versions of an IV may result in different compliance rates and offer a unique opportunity to study IV estimates' generalizability. In this article, we introduce a novel nested IV assumption and study identification of the average treatment effect among two latent subgroups: always-compliers and switchers, who are defined based on the joint potential treatment received under two versions of a binary IV. We derive the efficient influence function for the SWitcher Average Treatment Effect (SWATE) and propose efficient estimators. We then propose formal statistical tests of the generalizability of IV estimates based on comparing the conditional average treatment effect among the always-compliers and that among the switchers under the nested IV framework. We apply the proposed framework and method to the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial and study the causal effect of colorectal cancer screening and its generalizability.
工具变量(IV)是从非随机数据中估计因果效应的常用工具。IV 的一个原型是具有非遵从性的随机试验,其中随机治疗分配可作为所接受的不可忽略的治疗的 IV。在单调性假设下,有效的 IV 可以非参数地识别不可识别的违规者亚群中的平均治疗效果,其普遍性往往受到争议。在许多研究中,可能存在多个版本的 IV,例如,在一项多中心临床试验中,不同的研究地点对采取相同治疗方法的不同劝告。这些不同版本的静脉注射可能会导致不同的依从率,为研究静脉注射估计值的可推广性提供了一个独特的机会。在本文中,我们引入了一个新颖的嵌套 IV 假设,并研究了在两类人群中平均治疗效果的识别问题:始终遵从者和转换者,这两类人群是根据二元 IV 的两个版本下共同接受的潜在治疗来定义的。我们推导出转换者平均治疗效果(SWATE)的有效影响函数,并提出了有效的估计值。然后,我们在比较嵌套 IV 框架下始终遵守者和转换者之间的条件平均治疗效果的基础上,对 IV 估计值的可推广性提出了正式的统计检验。我们将提出的框架和方法应用于前列腺癌、肺癌、结直肠癌和卵巢癌(PLCO)筛查试验,研究结直肠癌筛查的因果效应及其可推广性。
{"title":"Nested Instrumental Variables Design: Switcher Average Treatment Effect, Identification, Efficient Estimation and Generalizability","authors":"Rui Wang, Ying-Qi Zhao, Oliver Dukes, Bo Zhang","doi":"arxiv-2405.07102","DOIUrl":"https://doi.org/arxiv-2405.07102","url":null,"abstract":"Instrumental variables (IV) are a commonly used tool to estimate causal\u0000effects from non-randomized data. A prototype of an IV is a randomized trial\u0000with non-compliance where the randomized treatment assignment serves as an IV\u0000for the non-ignorable treatment received. Under a monotonicity assumption, a\u0000valid IV non-parametrically identifies the average treatment effect among a\u0000non-identifiable complier subgroup, whose generalizability is often under\u0000debate. In many studies, there could exist multiple versions of an IV, for\u0000instance, different nudges to take the same treatment in different study sites\u0000in a multi-center clinical trial. These different versions of an IV may result\u0000in different compliance rates and offer a unique opportunity to study IV\u0000estimates' generalizability. In this article, we introduce a novel nested IV\u0000assumption and study identification of the average treatment effect among two\u0000latent subgroups: always-compliers and switchers, who are defined based on the\u0000joint potential treatment received under two versions of a binary IV. We derive\u0000the efficient influence function for the SWitcher Average Treatment Effect\u0000(SWATE) and propose efficient estimators. We then propose formal statistical\u0000tests of the generalizability of IV estimates based on comparing the\u0000conditional average treatment effect among the always-compliers and that among\u0000the switchers under the nested IV framework. We apply the proposed framework\u0000and method to the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer\u0000Screening Trial and study the causal effect of colorectal cancer screening and\u0000its generalizability.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140941495","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Economic policy sciences are constantly investigating the quality of well-being of broad sections of the population in order to describe the current interdependence between unequal living conditions, low levels of education and a lack of integration into society. Such studies are often carried out in the form of surveys, e.g. as part of the EU-SILC program. If the survey is designed at national or international level, the results of the study are often used as a reference by a broad range of public institutions. However, the sampling strategy per se may not capture enough information to provide an accurate representation of all population strata. Problems might arise from rare, or hard-to-sample, populations and the conclusion of the study may be compromised or unrealistic. We propose here a two-phase methodology to identify rare, poorly sampled populations and then resample the hard-to-sample strata. We focused our attention on the 2019 EU-SILC section concerning the Italian region of Liguria. Methods based on dispersion indices or deep learning were used to detect rare populations. A multi-frame survey was proposed as the sampling design. The results showed that factors such as citizenship, material deprivation and large families are still fundamental characteristics that are difficult to capture.
{"title":"Strategies for Rare Population Detection and Sampling: A Methodological Approach in Liguria","authors":"G. Lancia, E. Riccomagno","doi":"arxiv-2405.01342","DOIUrl":"https://doi.org/arxiv-2405.01342","url":null,"abstract":"Economic policy sciences are constantly investigating the quality of\u0000well-being of broad sections of the population in order to describe the current\u0000interdependence between unequal living conditions, low levels of education and\u0000a lack of integration into society. Such studies are often carried out in the\u0000form of surveys, e.g. as part of the EU-SILC program. If the survey is designed\u0000at national or international level, the results of the study are often used as\u0000a reference by a broad range of public institutions. However, the sampling\u0000strategy per se may not capture enough information to provide an accurate\u0000representation of all population strata. Problems might arise from rare, or\u0000hard-to-sample, populations and the conclusion of the study may be compromised\u0000or unrealistic. We propose here a two-phase methodology to identify rare,\u0000poorly sampled populations and then resample the hard-to-sample strata. We\u0000focused our attention on the 2019 EU-SILC section concerning the Italian region\u0000of Liguria. Methods based on dispersion indices or deep learning were used to\u0000detect rare populations. A multi-frame survey was proposed as the sampling\u0000design. The results showed that factors such as citizenship, material\u0000deprivation and large families are still fundamental characteristics that are\u0000difficult to capture.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140828627","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The Monty Hall problem is notorious for its deceptive simplicity. Although today it is widely used as a provocative thought experiment to introduce Bayesian thinking to students of probability, in the not so distant past it was rejected by established mathematicians. This essay provides some historical background to the problem and explains why it is considered so counter-intuitive to many. It is argued that the main barrier to understanding the problem is the back-grounding of the concept of dependence in probability theory as it is commonly taught. To demonstrate this, a Bayesian solution is provided and augmented with a probabilistic graphical model (PGM) inspired by the work of Pearl (1988, 1998). Although the Bayesian approach produces the correct answer, without a representation of the dependency structure of events implied by the problem, the salient fact that motivates the problem's solution remains hidden.
{"title":"What's So Hard about the Monty Hall Problem?","authors":"Rafael C. Alvarado","doi":"arxiv-2405.00884","DOIUrl":"https://doi.org/arxiv-2405.00884","url":null,"abstract":"The Monty Hall problem is notorious for its deceptive simplicity. Although\u0000today it is widely used as a provocative thought experiment to introduce\u0000Bayesian thinking to students of probability, in the not so distant past it was\u0000rejected by established mathematicians. This essay provides some historical\u0000background to the problem and explains why it is considered so\u0000counter-intuitive to many. It is argued that the main barrier to understanding\u0000the problem is the back-grounding of the concept of dependence in probability\u0000theory as it is commonly taught. To demonstrate this, a Bayesian solution is\u0000provided and augmented with a probabilistic graphical model (PGM) inspired by\u0000the work of Pearl (1988, 1998). Although the Bayesian approach produces the\u0000correct answer, without a representation of the dependency structure of events\u0000implied by the problem, the salient fact that motivates the problem's solution\u0000remains hidden.","PeriodicalId":501323,"journal":{"name":"arXiv - STAT - Other Statistics","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140828628","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}