No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Statistical Models and Shoe Leather
Abstract
Regression models have been used in the social sciences at least since 1899, when Yule published a paper on the causes of pauperism. Regression models are now used to make causal arguments in a wide variety of applications, and it is perhaps time to evaluate the results. No definitive answers can be given, but this paper takes a rather negative view. Snow's work on cholera is presented as a success story for scientific reasoning based on nonexperimental data. Failure stories are also discussed, and comparisons may provide some insight. In particular, this paper suggests that statistical technique can seldom be an adequate substitute for good design, relevant data, and testing predictions against reality in a variety of settings.
... While the distinction between predictive and causal statements has contributed to holding the truce among different quantitative research traditions (Freedman, 1991;Watts, 2014), unease is rising as scholars are increasingly using machine learning (ML) algorithms to analyze social phenomena (Bail, 2017;Lazer et al., 2020;Molina & Garip, 2019;Nelson, 2020;Shmueli, 2010;Turco & Zuckerman, 2017;Verhagen, 2022;Watts, 2017). ML is the study of how algorithms can learn from data (e.g., past social events) with no or little human guidance, thereby predicting new data instances (e.g., future social events) (Hastie et al., 2009). ...
... The data modeling culture (DMC) refers roughly to practices aiming to conduct model validation, and thus, statistical inference on one or several quantities of interest-distributions, model parameters, and alike. In the context of social sciences, such inferences often refer to defining a procedure that estimates a true quantity , to minimize the error | (Freedman, 1991). This true quantity is assumed to exist independently of the statistical model. ...
... This culture is the modus operandi of many strands in engineering, computer science, industry, and policy (Sanders 2019). As AMC procedures do not align with the hypothetico-deductive scientific method, this culture lacks subscribers from causally oriented social science research (Freedman, 1991;Molina & Garip, 2019). 3 The key difference between DMC and AMC is that the former is process oriented (i.e., modeling the generative process of the data), while the latter is performance oriented (i.e., building an emulator to match the predictive performance of a social system as closely as possible). ...
... The data reveal that gender affects the admit ratio, but they do not tell us HOW this happens. To study the causal chain, we must follow the real-world process closely; in Freedman's (1991) terminology, we must expend shoe leather to learn about causes. As we follow the admissions process step-by-step, we will see that females are mostly applying to HUM while males are applying to ENG. ...
... There are many possible hypothesis and explanations which could be explored and tested here. For a discussion of how causes are discovered in real life examples, see Freedman (1991) on Statistical Models and Shoe Leather. ...
... It is evident from the above given discussion that causal relationships in real life can actually be very complex. The hard work involved in the process of searching for causes has been described by Freedman (1991) in several real-world examples. Because econometricians are not taught to think about causes, most of the regression relationships we write down are spurious. ...
Bitter fighting among Christian factions and immoral behavior among Church leaders led to a transition to secular thought in Europe (see Zaman (2018) for details). One of the consequences of rejection of religion was the rejection of all unobservables. Empiricists like David Hume rejected all knowledge which was not based on observations and logic. He famously stated that: ““If we take in our hand any volume; of divinity or school metaphysics, for instance; let us ask, Does it contain any abstract reasoning concerning quantity or number? No. Does it contain any experimental reasoning concerning matter of fact and existence? No. Commit it then to the flames: for it can contain nothing but sophistry and illusion.” David Hume further realized that causality was not observable. This means that it is observable that event Y happened after event X, but it is not observable that Y happened due to X. The underlying mechanisms which connect X to Y are not observable. Current Article discusses the impact of changing causal structures on relationships and results of econometric analysis. it shows that conventional econometric analysis is devoid of causal chains which makes it impossible to get realistic results.
... This approach, which is compatible with data analysis and inference as "principled argument" (Abelson, 1995), requires data from multiple vantage points. Without the use of experimental designs, which cannot be effectively applied in the CVI context (Papachristos, 2025), and fixed, variable-driven measurement we are left with a host of approaches that require more "shoe leather" (Freedman, 1991). ...
... Jeong and Rothenhäusler (2025) extend stability analysis to account for distributional uncertainty, modeling violations of assumptions such as ignorability as small perturbations to the data-generating distribution. Using different estimation strategies is also commonly recommended to corroborate a causal hypothesis (Freedman, 1991;Rosenbaum, 2010;Karmakar et al., 2019). Our method aligns with and extends these perspectives by aiming to stabilize causal inference amidst the uncertainty about which adjustment set is valid, and offering a principled way to reconcile inference using multiple plausible adjustment sets. ...
In observational causal inference, it is common to encounter multiple adjustment sets that appear equally plausible. It is often untestable which of these adjustment sets are valid to adjust for (i.e., satisfies ignorability). This discrepancy can pose practical challenges as it is typically unclear how to reconcile multiple, possibly conflicting estimates of the average treatment effect (ATE). A naive approach is to report the whole range (convex hull of the union) of the resulting confidence intervals. However, the width of this interval might not shrink to zero in large samples and can be unnecessarily wide in real applications. To address this issue, we propose a summary procedure that generates a single estimate, one confidence interval, and identifies a set of units for which the causal effect estimate remains valid, provided at least one adjustment set is valid. The width of our proposed confidence interval shrinks to zero with sample size at rate, unlike the original range which is of constant order. Thus, our assumption-robust approach enables reliable causal inference on the ATE even in scenarios where most of the adjustment sets are invalid. Admittedly, this robustness comes at a cost: our inferential guarantees apply to a target population close to, but different from, the one originally intended. We use synthetic and real-data examples to demonstrate that our proposed procedure provides substantially tighter confidence intervals for the ATE as compared to the whole range. In particular, for a real-world dataset on 401(k) retirement plans our method produces a confidence interval 50\% shorter than the whole range of confidence intervals based on multiple adjustment sets.
... Of course, if these conditions hold (along with non-parametric identification), then estimated propensity scores converge to the true propensity scores, and causal effects may be learned. RCTs, because of associated "shoe-leather," bypass this necessity by knowing the propensity score by design (Freedman, 1991). ARSSS demonstrate this with an adversarial example where function complexity increases with sample size. ...
We respond to Aronow et al. (2025)’s paper arguing that randomized controlled trials (RCTs) are “enough,” while nonparametric identification in observational studies is not. We agree with their position with respect to experimental versus observational research, but question what it would mean to extend this logic to the scientific enterprise more broadly. We first investigate what is meant by “enough,” arguing that this is a fundamentally a sociological claim about the relationship between statistical work and larger social and institutional processes, rather than something that can be decided from within the logic of statistics. For a more complete conception of “enough,” we outline all that would need to be known – not just knowledge of propensity scores, but knowledge of many other spatial and temporal characteristics of the social world. Even granting the logic of the critique in Aronow et al. (2025), its practical importance is a question of the contexts under study. We argue that we should not be satisfied by appeals to intuition about the complexity of “naturally occurring” propensity score functions. Instead, we call for more empirical metascience to begin to characterize this complexity. We apply this logic to the example of recommender systems developed by Aronow et al. (2025) as a demonstration of the weakness of allowing statisticians’ intuitions to serve in place of metascientific data. Rather than implicitly deciding what is “enough” based on statistical applications the social world has determined to be most profitable, we argue that practicing statisticians should explicitly engage with questions like “for what?” and “for whom?” in order to adequately answer the question of “enough?”
... Another important challenge for causal inference based on observational studies is whether it is possible to distinguish successful and unsuccessful uses of a model or procedure, by now an old question raised by Freedman (1991). This is a fundamentally important question for statisticians and data scientists as causal inference becomes popular in social science and other areas where observational studies are more common than randomized experiments. ...
This project was sponsored by the National Science Foundation and organized by a steering committee and a group of theme leaders. The six-member steering committee, consisting of James Berger, Xuming He, David Madigan, Susan Murphy, Bin Yu, and Jon Wellner, was responsible for the overall planning of the project. This report is designed to be accessible to the wider audience of key stakeholders in statistics and data science, including academic departments, university administration, and funding agencies. After the role and the value of Statistics and Data Science are discussed in Section 1, the report focuses on the two goals related to emerging research and data-driven challenges in applications. Section 2 identifies emerging research topics from the data challenges arising from scientific and social applications, and Section 3 discusses a number of emerging areas in foundational research. How to engage with those data-driven challenges and foster interdisciplinary collaborations is also summarized in the Executive Summary. The third goal of creating a vibrant research community and maintaining an appropriate balance is addressed in Sections 4 (Professional Culture and Community Responsibilities) and 5 (Doctoral Education).
... Of course, if these conditions hold (along with non-parametric identification), then estimated propensity scores converge to the true propensity scores, and causal effects may be learned. RCTs, because of associated "shoe-leather," bypass this necessity by knowing the propensity score by design (Freedman, 1991). ARSSS demonstrate this with an adversarial example where function complexity increases with sample size. ...
We respond to Aronow et al. (2025)'s paper arguing that randomized controlled trials (RCTs) are "enough," while nonparametric identification in observational studies is not. We agree with their position with respect to experimental versus observational research, but question what it would mean to extend this logic to the scientific enterprise more broadly. We first investigate what is meant by "enough," arguing that this is fundamentally a sociological claim about the relationship between statistical work and larger social and institutional processes, rather than something that can be decided from within the logic of statistics. For a more complete conception of "enough," we outline all that would need to be known -- not just knowledge of propensity scores, but knowledge of many other spatial and temporal characteristics of the social world. Even granting the logic of the critique in Aronow et al. (2025), its practical importance is a question of the contexts under study. We argue that we should not be satisfied by appeals to intuition about the complexity of "naturally occurring" propensity score functions. Instead, we call for more empirical metascience to begin to characterize this complexity. We apply this logic to the example of recommender systems developed by Aronow et al. (2025) as a demonstration of the weakness of allowing statisticians' intuitions to serve in place of metascientific data. Rather than implicitly deciding what is "enough" based on statistical applications the social world has determined to be most profitable, we argue that practicing statisticians should explicitly engage with questions like "for what?" and "for whom?" in order to adequately answer the question of "enough?"
... However, in complex and diverse societies, certain data are too strict. Therefore, the academic community has adopted an ostrich policy, gradually shifting its focus to statistical significance and maintaining an open or suspenseful attitude toward whether the data meet model assumptions (Freedman 1991). Second, the conclusion is about the mechanism of the model rather than the mechanism of the facts. ...
The inductive logic of grounded theory and the principle of avoiding theoretical preconceptions are significantly different from the deductive logic and hypothesis testing of traditional quantitative research. Based on the limitations of theory production in quantitative research, this paper proposes a Computing Grounded Theory (CGT) approach that directly quantitatively assists theories. With the help of machine learning and attribution algorithms, CGT identifies variables that have not been the focus of previous studies based on the predictive power of the independent variables to propose new theoretical hypotheses, following the principle that causality is a sufficient and unnecessary condition for predictability. This paper systematically discusses CGT’s basic idea, logical premise, and methodological foundation while providing an empirical example. This method bridges the gap in the theoretical production of quantitative research and is of great value in theory, discipline, knowledge systems and social governance.
... There is a long history of using linear regression models in education and social science research to attempt to answer causal research questions with observational study data (Freedman, 1991;Berk, 2004;Morgan & Winship, 2015). Throughout that history, arguments about controlling for the right set of covariates, satisfying the assumptions of the particular modeling approach (e.g., independent and normally distributed errors with constant variance and zero mean), and having the right temporal ordering (i.e., the treatment was implemented before the outcome was measured) have been used to justify causal attribution. ...
Causal inference involves determining whether a treatment (e.g., an education program) causes a change in outcomes (e.g., academic achievement). It is well-known that causal effects are more challenging to estimate than associations. Over the past 50 years, the potential outcomes framework has become one of the most widely used approaches for defining, identifying, and estimating causal effects. In this paper, we review the potential outcomes framework with a focus on potential outcomes notation to define individual and average causal effects. We then show how three canonical assumptions, Unconfoundedness, Positivity, and Consistency, may be used to identify average causal effects. The identification results motivate methods for estimating causal effects in practice, which include model-based estimators, such as regression, inverse probability weighting, and doubly robust estimation, and procedures that target covariate balance, such as matching and stratification. Examples and discussion are grounded in the context of a running example of a study aimed at assessing the causal effect of receipt of special education services on 5th grade mathematics achievement in school-aged children. Practical considerations for education research are discussed.
... 212-217] for a particularly concise review. See also [1,6,9,26,39,40,48]. 13 Marie-Françoise Jozeau has documented the competition of the two traditions in geodesy [36,47]. ...
When testing a statistical hypothesis, is it legitimate to deliberate on the basis of initial data about whether and how to collect further data? Game-theoretic probability’s fundamental principle for testing by betting says yes, provided that you are testing the hypothesis’s predictions by betting and do not risk more capital than initially committed. Standard statistical theory uses Cournot’s principle, which does not allow such optional continuation. Cournot’s principle can be extended to allow optional continuation when testing is carried out by multiplying likelihood ratios, but the extension lacks the simplicity and generality of testing by betting. Testing by betting can also help us with descriptive data analysis. To obtain a purely and honestly descriptive analysis using competing probability distributions, we have them bet against each other using the principle. The place of confidence intervals is then taken by sets of distributions that do relatively well in the competition. In the simplest implementation, these sets coincide with R. A. Fisher’s likelihood ranges.
... Here again, both kinds of misclassification error decrease the chance of showing that the intervention works. Furthermore, lapses in treatment integrity at the design level cannot be rectified at the statistical analysis level (Freedman, 1991). A policy of adhering to the "intention-totreat" principle (Peto et al., 1977) so often used in clinical trials (analyzing results according to the original assignment of experimental and control group participants) also biases the conclusions in the direction of finding no difference. ...
The article focuses on the validity of measures of service delivery; any attempt to establish a linkage between service and outcomes follows the assumption that services have been given and received. Claims relating to the benefits of attending self-help groups on the mental health functioning of individuals with depression or manic depression were examined in the context of a well-controlled study and a validated measure of individual attendance. A test was conducted to probe the conjecture that the greater the degree of involvement in self-help groups, the greater the improvement in functioning. Additional questions related to individual patterns of attendance, generalization of the intervention's effects, and migration from “home” to other meeting sites.
... 212-217), is particularly concise and revealing. See alsoAlexander (2015);Berk and Freedman (2003);Bru et al. (1997);Freedman (1991);Mason (1991);Matthews (1995);Shafer (2019a).16 Marie-Françoise Jozeau has documented the competition of the two traditions in geodesy(Jozeau, 1997;Shafer, 2019b). ...
When testing a statistical hypothesis, is it legitimate to deliberate on the basis of initial data about whether and how to collect further data? Game-theoretic probability's fundamental principle for testing by betting says yes, provided that you are testing by betting and do not risk more capital than initially committed. Standard statistical theory uses Cournot's principle, which does not allow such optional continuation. Cournot's principle can be extended to allow optional continuation when testing is carried out by multiplying likelihood ratios, but the extension lacks the simplicity and generality of testing by betting. Game-theoretic probability can also help us with descriptive data analysis. To obtain a purely and honestly descriptive analysis using competing probability distributions, we have them bet against each other using the Kelly principle. The place of confidence intervals is then taken by a sets of distributions that do relatively well in the competition. In the simplest implementation, these sets coincide with R. A. Fisher's likelihood intervals.
... Regressions for selection into fertility are estimated using the specification We control for parental schooling in all our regressions to deal with this potential selection bias. We also check regression results without controlling for parental schooling to identify how much heavy lifting is required by this control (Freedman, 1991). We then consider parental schooling as the outcome variable in a falsification test to check whether there is a bias arising from selective parental schooling (Angrist and Pischke, 2008). ...
... methods influence how people self-present themselves (Freedman 1991;Kiesler and Sproull 1992;Spears and Lea 1994). While in-person field research has historically been viewed as "the gold standard against which the performance of computer-mediated interaction is judged," digital methods have now proven to be equally valid and legitimate approaches to research (Hine 2005, 4). ...
The COVID-19 pandemic has underscored the unpredictability and instability of fieldwork as a method for data collection. As the pandemic prompted unprecedented political dynamism and social and economic disruptions at both domestic and global levels, in-person fieldwork became challenging, if even possible, in the two years following March 2020. While scholars are again using traditional fieldwork methods, we have seen an increased use of digital tools to conduct research remotely since the pandemic due to international travel bans and social distancing measures. Although not yet widely discussed, these new approaches pose new ethical questions as understandings of both our “fields” and “homes” evolve. In this paper, we stress the need for scholars to reconsider how we conceive of our ethical obligations in situations wherein we have conducted research without ever physically accessing our field sites or interacting in person with our participants. We particularly urge researchers to re-evaluate their ethical responsibilities around transparency and replicability in the dissemination and publication of findings when engaging in fieldwork “from home.” These considerations were necessary prior to 2020 but are especially relevant within the context of the pandemic as scholars enter new field sites remotely or return to those previously visited in person. As a result, this paper starts a critical conversation about ethical practices in remote and digital fieldwork, which will continue to prove significant as digital and remote methods are used for data collection in a post-pandemic world.
... Donald Schön describes how the high hard ground where practitioners can make effective use of their techniques is quite distinct from "swampy lowlands", which represent messes incapable of technical solution and require critical reflection [46]. David A. Freedman highlights how statistical methods (upon which modern data science and machine learning is based) is seldom an adequate substitute for good research design, relevant data, and expending of "shoe leather" testing predictions against reality in a wide variety of settings [50]. This is the reason why apprenticeship, internships, and case studies are methods of choice for exposing students to authentic experiences from which the students can learn insights after reflection to guide future practice [51]. ...
In a volatile, uncertain, complex, and ambiguous (VUCA) 21st-century world, there is a great need for drawing upon capacity for holistic thinking and wisdom. We need to leverage multiple transdisciplinary thinking that allows a multiplexity of ways of thinking and knowing to deal with multi-layered complex reality. Our society is rife with elusive “wicked problems” that defy narrow technological solutions emerging from disciplinary siloed thinking. A narrow education in such times is not only sub-optimal but also dangerous. With the advancement in artificial intelligence, we now have too advanced a technology to be able to survive without wisdom. We require a broad range of experiences and models to study the same problem and the wisdom to choose and combine different models in appropriate ways. We need not only mathematics, engineering, and sciences for engineers but also an exposure to the important ideas in philosophy, psychology, history, economics, sustainability, and significant insights about subjects of enduring interest to human beings.
... The first desideratum to consider is predictive accuracy. If the constructed model does not accurately represent the underlying problem, any subsequent analysis will be suspect [34,84]. Second, the main purpose of model-based interpretation methods is to increase descriptive accuracy. ...
The recent surge in highly successful, but opaque, machine-learning models has given rise to a dire need for interpretability. This work addresses the problem of interpretability with novel definitions, methodology, and scientific investigations, ensuring that interpretations are useful by grounding them in the context of real-world problems and audiences. We begin by defining what we mean by interpretability and some desiderata surrounding it, emphasizing the underappreciated role of context. We then dive into novel methods for interpreting/improving neural network models, focusing on how to best score, use, and distill interactions. Next, we turn from neural networks to relatively simple rule-based models, where we investigate how to improve predictive performance while maintaining an extremely concise model. Finally, we conclude with work on open-source software and data for facilitating interpretable data science. In each case, we dive into a specific context which motivates the proposed methodology, ranging from cosmology to cell biology to medicine.
... Donald Schön describes how the high hard ground where practitioners can make effective use of their techniques is quite distinct from "swampy lowlands", which represent messes incapable of technical solution and require critical reflection [46]. David A. Freedman highlights how statistical methods (upon which modern data science and machine learning is based) is seldom an adequate substitute for good research design, relevant data, and expending of "shoe leather" testing predictions against reality in a wide variety of settings [50]. This is the reason why apprenticeship, internships, and case studies are methods of choice for exposing students to authentic experiences from which the students can learn insights after reflection to guide future practice [51]. ...
In a volatile, uncertain, complex, and ambiguous (VUCA) 21st-century world, there is a great need for drawing upon capacity for holistic thinking and wisdom. We need to leverage multiple transdisciplinary thinking that allows a multiplexity of ways of thinking and knowing to deal with multi-layered complex reality. Our society is rife with elusive "wicked problems" that defy narrow technological solutions emerging from disciplinary siloed thinking. A narrow education in such times is not only sub-optimal but also dangerous. With the advancement in artificial intelligence, we now have too advanced a technology to be able to survive without wisdom. We require a broad range of experiences and models to study the same problem and the wisdom to choose and combine different models in appropriate ways. We need not only mathematics, engineering, and sciences for engineers but also an exposure to the important ideas in philosophy, psychology, history, economics, sustainability, and significant insights about subjects of enduring interest to human beings.
... Considerations of model stability have emerged in Bayesian statistics (Box, 1980;Skene et al., 1986), causal inference (Leamer, 1983;LaLonde, 1986;Rosenbaum, 1987;Imbens and Rubin, 2015) and in discussions about the data science lifecycle (Yu, 2013;Steegen et al., 2016;Yu and Kumbier, 2020). Using different estimation strategies is commonly recommended to corroborate a causal hypothesis (Freedman, 1991;Rosenbaum, 2010;Karmakar et al., 2019). In particular, to evaluate omitted variable bias, it is a common recommendation to consider the between-estimator variation of several adjusted regressions (Oster, 2019). ...
During data analysis, analysts often have to make seemingly arbitrary decisions. For example during data pre-processing there are a variety of options for dealing with outliers or inferring missing data. Similarly, many specifications and methods can be reasonable to address a domain question. This may be seen as a hindrance for reliable inference, since conclusions can change depending on the analyst's choices. In this paper, we argue that this situation is an opportunity to construct confidence intervals that account not only for sampling uncertainty but also some type of distributional uncertainty. The distributional uncertainty model is related to a variety of potential issues with the data, ranging from dependence between observations to selection bias and confounding. The final procedure can be seen as an extension of current statistical practice. However, the underlying assumptions are quite different. Standard statistical practice often relies on the i.i.d. assumption. We rely on a strictly weaker symmetry assumption stating that the empirical distribution and the target distribution differ by an isotropic distributional perturbation.
... Les travaux développés dans ce cadre accordent une place centraleà la régression : historiquement la quantification de la causalité y prend essentiellement la forme du calcul de régression. L'approche s'est donc révélée vulnérable aux critiques visant la régression et son utilisation de plus en plus systématique et automatisée en sciences sociales ( [Freedman, 1991] par exemple). Il est cependant possible de considérer,à l'instar de Blalock dans sa réponsè a Freedman (dans [Blalock, 1991]), que les critiques formulées n'atteignent pas la méthode ellemême mais seulement ses applications mal informées ou peu réflexives. ...
... Measurement remains one of the most poorly developed parts of the tool kit of the social sciences, especially the measurement of dynamical properties of processes (see, e.g., Collins and Sayer, 2001). Little wonder that statisticians have concluded that so much statistical inference testing in regression-type analyses, lacking firm basis in substantive knowledge of the phenomenon being studied, is mostly empty ritual (Lieberson 1985;Freedman 1991;Wasserstein, Schirm, and Lazar 2019). ...
A short primer on inference (observational, mensural, causal, and statistical), requiring very little mathematics.
... Regressions for selection into fertility are estimated using the specification We control for parental schooling in all our regressions to deal with this potential selection bias. We also check regression results without controlling for parental schooling to identify how much heavy lifting is required by this control (Freedman, 1991). We then consider parental schooling as the outcome variable in a falsification test to check whether there is a bias arising from selective parental schooling (Angrist and Pischke, 2008). ...
We use childhood exposures to disasters as natural experiments inducing variations in adulthood outcomes. Following the fetal origin hypothesis, we hypothesize that children from households with greater exposure will have poorer health, schooling, and consumption outcomes. Employing a unique dataset from Bangladesh, we test this hypothesis for the 1970 cyclone that killed over 300,000 people in southern Bangladesh. We find that children surviving the cyclone experience significant health, schooling and consumption adversities, and during their adulthood, have lower probabilities of good health and primary schooling; and lower durations of good health, schooling and consumption. Such adversities are further heightened among the rural and less-educated households. Therefore, public programs benefiting the females and the poor, alongside the development of healthcare and schooling infrastructure, can be useful protective measures against the long-term harms of a disaster.
... Joka tapauksessa luonnollisista koeasetelmista kiinnostuneiden tutkijoiden kannattaa varautua siihen, että rekisteriaineistojen sielunelämää ja syntyprosessia joutuu selvittämään myös laadullisin tutkimusmenetelmin. Datan yksityiskohtien penkominen voi olla puuduttavaa, mutta "kengänpohjien kuluttaminen" (Freedman, 1991) ja tutkittavan muutoksen tosiasialliseen toteutukseen perehtyminen nostaa eittämättä tutkimuksen laatua. ...
Rikollisuutta koskevien rekisteriaineistojen tutkimuskäyttö on lisääntynyt 2000-luvulla voimakkaasti Pohjoismaissa ja myös Suomessa. Kriminologisen tutkimuksen mahdollisuudet ovat parantuneet rekisterilinkkausaineistojen aikaisempaa paremman saatavuuden myötä, ja erityisesti Tilastokeskuksen etäkäyttöjärjestelmä FIONA:n perustaminen on ollut tärkeä uudistus. Rekisteriaineistoja on hyödynnetty viime vuosina sekä rikollisuuden syiden että rikosseuraamusten vaikutusten pohjoismaisessa tutkimuksessa varsin monipuolisesti. Yksilötasoisilla pitkittäisaineistoilla päästään kiinni huomattavasti aggregaattitilastoja tarkempaan rikollisuuden muutosten analyysin ja niitä voidaan hyödyntää aikaisempaa uskottavammassa kausaalivaikutusten tutkimuksessa. Luonnollisten koeasetelmien ja kvasikokeellisten tutkimusmenetelmien kehitys on hyödyttänyt rekisteritutkimusta selvästi ja nostanut osaltaan rekisteriaineistojen käyttöarvoa. Kirjoitus perustuu 14.12.2020 pidettyyn Veli Verkko -luentoon.
... Snow carried out a multifaceted and mixed-method investigation using both qualitative and quantitative data. Nevertheless, many scholars believe that his strongest piece of evidence is a natural experiment during the cholera epidemic of 1853-54 (Freedman 1991(Freedman , 1999Dunning 2012). We focus on this component of his study. ...
In this article, we import set-theoretic methods from the qualitative research tradition into the quantitative research tradition. We focus specifically on set-theoretic methods designed to analyze the extent to which a condition is necessary and is sufficient for an outcome of interest. We use these methods to reanalyze four major studies from the quantitative tradition. We find that set-theoretic methods call attention to asymmetrical patterns in the data that otherwise go unnoticed and unanalyzed. We develop a general set-theoretic framework for the study of necessity and sufficiency. We conclude that the use of this framework can enrich existing and future quantitative research in the social sciences.
... Finally, in some cases data are physically veriable and researchers can use a little of what Freedman (1991) called shoe leather and simply verify behavior. For instance, in Mexico, the government sent administrators to audit self-reported asset data used to decide who was in or out of a cash transfer program and found underreporting of assets to increase eligibility (Martinelli and Parker, 2009). ...
... The first example of a problem-focused science comes from the discovery of cholera in England. The efforts of John Snow have been retold by numerous authors (e.g., Freedman, 1992) including Snow himself (1936), and his story is quite simple. People were getting sick and dying from a new medical condition in London. ...
... Här kan det handla om att till exempel jämföra studieprestationer bland placerade med icke-placerade barn när vi statistiskt kontrollerar för psykiatriska problem och andra faktorer som vi bedömer som relevanta. Regressionsmodeller har använts inom samhällsvetenskapen i över hundra år (Freedman 1991) och är enkla att utföra med sedvanliga statistikprogram. Det finns en omfattande metodlitteratur att utgå ifrån där grundantaganden hos regression förtydligas. ...
Sedan länge har dygnsvården för barn ifrågasatts utifrån de bristfälliga resultat empirisk forskning har kunnat påvisa. Registerstudier visar genomgående att placerade barn har sämre utfall än barn från normalpopulationen, och dessa studier har spelat en viktig roll för att uppmärksamma de placerade barnens utsatta situation. Den här artikeln diskuterar utförligt frågan om de negativa utfallen hos de placerade går att hänföra till placeringen som sådan. Forskningen om dygnsvården bygger på observationsstudier och det är svårt att komma åt kausala effekter via en sådan design. Det gäller särskilt för samhällsvård där de placerade barnen utgör en starkt selekterad grupp. I frånvaron av randomiserade kontrollerade experiment kan vi i princip aldrig veta om de sämre utfallen bland de placerade barnen beror på placeringen som sådan eller på andra faktorer som vi saknar data om. Ett centralt argument är att inte ens de mest avancerade statistiska teknikerna kan lösa svårigheten med att komma åt effekterna av samhällsvården i observationsstudier. En del av svårigheterna går dock att hantera via design, där ett starkare metodologiskt upplägg är att inkludera jämförelsegrupper som är mer lika de placerade barnen. Ett fåtal metodologiskt starkare studier med ett sådant upplägg ger betydligt färre indikationer på att vården har en negativ inverkan.
... Shoeleather epidemiology, on the other hand, pertains to Collection of epidemiological and other pertinent data relevant to an epidemiological investigation by painstaking direct inquiry among all or a representative sample of the affected people, for example by 48 walking door to door (wearing out shoe leather in the process, hence the term) to ask direct questions. (Last 2007) We are using the term "shoe-leather public health" for a public health approach of epidemic control investigation, planning and implementation that is based on familiarity with the specifi c context and grounded in real-life evidence from local conditions (Freedman 1991). ...
The evolving COVID-19 pandemic requires that data and operational responses be examined from a public health perspective. While there exist deep contestations about the epidemic control strategies to be adopted, past experience seems to be corroborated in the present epidemic that a contextually rooted “shoe-leather public health” approach provides the most effective interventions and operational strategies, more so in a society as diverse as ours. Drawing from this, an analysis of the COVID-19 situation in India is put forth, and debates on mitigation strategies, optimisation of testing, and the essential steps for a comprehensive set of interventions in order to minimise human suffering are addressed.
... Regression. Regression analysis is the most basic and commonly used predictive tool, especially in practical applications [72], and is based on a statistical process for estimating the relationships among variables. ...
The increasing power of computer technology does not dispense with the need to extract meaningful in-formation out of data sets of ever growing size, and indeed typically exacerbates the complexity of this task. To tackle this general problem, two methods have emerged, at chronologically different times, that are now commonly used in the scientific community: data mining and complex network theory. Not only do complex network analysis and data mining share the same general goal, that of extracting information from complex systems to ultimately create a new compact quantifiable representation, but they also often address similar problems too. In the face of that, a surprisingly low number of researchers turn out to resort to both methodologies. One may then be tempted to conclude that these two fields are either largely redundant or totally antithetic. The starting point of this review is that this state of affairs should be put down to contingent rather than conceptual differences, and that these two fields can in fact advantageously be used in a synergistic manner. An overview of both fields is first provided, some fundamental concepts of which are illustrated. A variety of contexts in which complex network theory and data mining have been used in a synergistic manner are then presented. Contexts in which the appropriate integration of complex network metrics can lead to improved classification rates with respect to classical data mining algorithms and, conversely, contexts in which data mining can be used to tackle important issues in complex network theory applications are illustrated. Finally, ways to achieve a tighter integration between complex networks and data mining, and open lines of research are discussed.
... hindsight, it is not at all surprising that Berlin failed to produce its own John Snow as the latter's analysis proted from unique circumstances approximating a natural experiment (Freedman, 1991). ...
Did cholera function as a potent catalyst for the reform of urban water infrastructure in 19th century Europe's disease-ridden cities, serving as "our old ally" in the struggle for urban sanitation (Robert Koch)? Based on a detailed case study of Berlin's hydrological reconfiguration, this paper challenges popular narratives that paint the emergence of safe tap water supplies and sanitary sewers as an efficient, scientifically motivated reaction to Europe's recurrent cholera epidemics since 1831. While historians have long stressed the dominance of aesthetical and industrial over sanitary concerns, the study of Berlin's contemporary discourse suggest that the causal link between cholera and water infrastructure reform was not only weak, but ambiguous. Far from motivating the right actions for the wrong reasons, cholera's conception through the dominant miasmatist frameworks and limited proto-epidemiological tools of the prebacteriological era inspired inefficient, at times even counterproductive approaches that potentially deepened the urban mortality penalty. Berlin's role as a political and scientific center of 19th century Europe suggests that her experience was the norm rather than the exception. A nuanced understanding of Western Europe's sanitary past has important implications for the continuing struggle for urban sanitation in today's developing world. JEL Codes: N33, N53, N93
... For instance, Philip Dawid (23) drew connections between statistical inference and prediction under the name "prequential statistics," highlighting the importance of forecasts in statistical analyses. David Freedman (24) argued that when a model's predictions are not tested against reality, conclusions drawn from the model are unreliable. Seymour Geisser (25) advocated that statistical analyses should focus on prediction rather than parametric inference, particularly in cases where the statistical model is an inappropriate description of reality. ...
Building and expanding on principles of statistics, machine learning, and scientific inquiry, we propose the predictability, computability, and stability (PCS) framework for veridical data science. Our framework, composed of both a workflow and documentation, aims to provide responsible, reliable, reproducible, and transparent results across the data science life cycle. The PCS workflow uses predictability as a reality check and considers the importance of computation in data collection/storage and algorithm design. It augments predictability and computability with an overarching stability principle. Stability expands on statistical uncertainty considerations to assess how human judgment calls impact data results through data and model/algorithm perturbations. As part of the PCS workflow, we develop PCS inference procedures, namely PCS perturbation intervals and PCS hypothesis testing, to investigate the stability of data results relative to problem formulation, data cleaning, modeling decisions, and interpretations. We illustrate PCS inference through neuroscience and genomics projects of our own and others. Moreover, we demonstrate its favorable performance over existing methods in terms of receiver operating characteristic (ROC) curves in high-dimensional, sparse linear model simulations, including a wide range of misspecified models. Finally, we propose PCS documentation based on R Markdown or Jupyter Notebook, with publicly available, reproducible codes and narratives to back up human choices made throughout an analysis. The PCS workflow and documentation are demonstrated in a genomics case study available on Zenodo.
... These studies, which were carried out in the United States, include longitudinal information on (a) teachers' evaluations of children's academic skills (a proxy for teachers' potentially biased perceptions of children), (b) different aspects of children's cultural capital (participation in performing arts, reading interests and participation in athletics and clubs), and (c) children's educational performance (standardized test scores in reading and math). This means that, in addition to offering a more robust methodological framework than previous research, we are able to replicate all analyses using two separate datasets, thereby increasing the external validity of our results (Bloome, 2015;Freedman, 1991). ...
In this paper, we test two mechanisms through which cultural capital might affect educational performance: (a) teachers misinterpreting cultural capital as signals of academic brilliance and (b) cultural capital fostering skills in children that enhance educational performance. We analyse data from the ECLS‐K and ECLS‐K:2011 from the United States and focus on three aspects of children’s cultural capital: participation in performing arts, reading interest and participation in athletics and clubs. We find that (1) none of the three aspects of cultural capital that we consider affects teachers’ evaluations of children’s academic skills; (2) reading interest has a direct positive effect on educational performance; and (3) the direct effect of reading interest on educational performance does not depend on schooling context. Our results provide little support for the hypothesis that cultural capital operates via signals about academic brilliance. Instead, they suggest that cultural capital fosters skills in children that enhance educational performance. We discuss the theoretical implications of our findings.
... However, this literature has difficulty accounting for the role of economic structure, tending to focus either on the specificities of a given time or place (Mckenzie, 2015;Shildrick and MacDonald, 2013), or making broad generalisations using vague concepts such as neo-liberalism (Jensen and Tyler, 2015;Tyler, 2008 My empirical approach has both advantages and disadvantages: by using quantitative methods I trade off the empirical and conceptual richness possible in more qualitative work for broader empirical reach. As I use observational data that is largely cross-sectional in nature, many of the results need to be viewed as descriptive rather than as providing strong evidence for causality (Freedman, 1991;Gerring, 2012). Chapter 4 is an exception to this, as it uses longitudinal data with fixed effects regression models to deal with a variety of threats to causal inference (Allison, 1994). ...
In this thesis I investigate how an individual’s economic position and the context they live in affects their sympathy for the poor. Poverty and welfare receipt are stigmatised across high income countries; such attitudes reduce support for redistribution and exacerbate the negative impact of poverty on wellbeing. Across three empirical chapters, I use attitudinal data from the UK and Europe to investigate the relationship between individual advantage, broader economic context, and the prevalence of stigmatising stereotypes about welfare recipients and the poor. I apply an innovative perspective combining qualitative research on the experiences of people in poverty and comparative political economy work on inequality and redistribution to address neglected topics in the study of deservingness perceptions. In the first empirical chapter I argue that those in more disadvantaged economic positions have more sympathetic attitudes towards welfare recipients. However, this relationship is counteracted by the role of social status and authoritarian attitudes, which can make the disadvantaged hold less sympathetic views. The second chapter uses survey data from twenty-seven European countries to show that individuals in more unequal nations are more likely to believe that laziness rather than injustice is the cause of poverty. I argue that a plausible explanation of this relationship is status anxiety among disadvantaged individuals. In the third chapter I conduct the first longitudinal analysis of the association between area level unemployment and attitudes towards the unemployed, finding little evidence of a meaningful effect of exposure on stigmatising stereotypes. Overall, this thesis argues that status anxiety plays a major role in shaping stigmatising stereotypes, explaining why people are less sympathetic towards the poor in high inequality contexts, and why disadvantaged individuals often hold especially negative attitudes.
Findings from tests of the predictive validity of causal models estimated using ordinary-least-squares (OLS) multiple regression were presented. Out-of-sample forecasts from 99 published models were compared with those from models estimated using nine simple and conservative alternative methods and one naive model. Forecast errors from models estimated using one of the alternative methods (multiple least absolute deviation regression, or median, regression) were smaller than those from the published OLS models on average and for most individual models. The findings have implications for model building practice and the quantitative estimation of causal relationships from empirical non-experimental data, and for AI (machine learning), which is based on OLS regression analysis.
This handbook is currently in development, with individual articles publishing online in advance of print publication. At this time, we cannot add information about unpublished articles in this handbook, however the table of contents will continue to grow as additional articles pass through the review process and are added to the site. Please note that the online publication date for this handbook is the date that the first article in the title was published online. For more information, please read the site FAQs.
This handbook is currently in development, with individual articles publishing online in advance of print publication. At this time, we cannot add information about unpublished articles in this handbook, however the table of contents will continue to grow as additional articles pass through the review process and are added to the site. Please note that the online publication date for this handbook is the date that the first article in the title was published online. For more information, please read the site FAQs.
Situated at the intersection of natural science and philosophy, Our Genes explores historical practices, investigates current trends, and imagines future work in genetic research to answer persistent, political questions about human diversity. Readers are guided through fascinating thought experiments, complex measures and metrics, fundamental evolutionary patterns, and in-depth treatment of exciting case studies. The work culminates in a philosophical rationale, based on scientific evidence, for a moderate position about the explanatory power of genes that is often left unarticulated. Simply put, human evolutionary genomics - our genes - can tell us much about who we are as individuals and as collectives. However, while they convey scientific certainty in the popular imagination, genes cannot answer some of our most important questions. Alternating between an up-close and a zoomed-out focus on genes and genomes, individuals and collectives, species and populations, Our Genes argues that the answers we seek point to rich, necessary work ahead.
In October 2019, Abhijit Banerjee, Esther Duflo, and Michael Kremer jointly won the 51st Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel "for their experimental approach to alleviating global poverty." But what is the exact scope of their experimental method, known as randomized control trials (RCTs)? Which sorts of questions are RCTs able to address and which do they fail to answer? This book provides answers to these questions, explaining how RCTs work, what they can achieve, why they sometimes fail, how they can be improved and why other methods are both useful and necessary. Chapters contributed by leading specialists in the field present a full and coherent picture of the main strengths and weaknesses of RCTs in the field of development. Looking beyond the epistemological, political, and ethical differences underlying many of the disagreements surrounding RCTs, it explores the implementation of RCTs on the ground, outside of their ideal theoretical conditions and reveals some unsuspected uses and effects, their disruptive potential, but also their political uses. The contributions uncover the implicit worldview that many RCTs draw on and disseminate, and probe the gap between the method's narrow scope and its success, while also proposing improvements and alternatives. This book warns against the potential dangers of their excessive use, arguing that the best use for RCTs is not necessarily that which immediately springs to mind, and offering opportunity to come to an informed and reasoned judgement on RCTs and what they can bring to development.
This book of proceedings contains peer-reviewed papers that were presented at the 29th ISTE International Conference on Transdisciplinary Engineering (TE2022), organized by System Design and Management (SDM) at the Massachusetts Institute of Technology in Cambridge, MA, United States from July 5–8, 2022.
TE2022 brought together a diverse global community of scholars and practitioners in dialogue and reflection on engineering itself. Engineering is changing rapidly. The
connectedness of the world’s most critical systems along with rapid advancement of
methods push us to ask “How will we teach, research, and practice engineering?”
This compact course is written for the mathematically literate reader who wants to learn to analyze data in a principled fashion. The language of mathematics enables clear exposition that can go quite deep, quite quickly, and naturally supports an axiomatic and inductive approach to data analysis. Starting with a good grounding in probability, the reader moves to statistical inference via topics of great practical importance – simulation and sampling, as well as experimental design and data collection – that are typically displaced from introductory accounts. The core of the book then covers both standard methods and such advanced topics as multiple testing, meta-analysis, and causal inference.
Fairfield and Charman provide a modern, rigorous and intuitive methodology for case-study research to help social scientists and analysts make better inferences from qualitative evidence. The book develops concrete guidelines for conducting inference to best explanation given incomplete information; no previous exposure to Bayesian analysis or specialized mathematical skills are needed. Topics covered include constructing rival hypotheses that are neither too simple nor overly complex, assessing the inferential weight of evidence, counteracting cognitive biases, selecting cases, and iterating between theory development, data collection, and analysis. Extensive worked examples apply Bayesian guidelines, showcasing both exemplars of intuitive Bayesian reasoning and departures from Bayesian principles in published case studies drawn from process-tracing, comparative, and multimethod research. Beyond improving inference and analytic transparency, an overarching goal of this book is to revalue qualitative research and place it on more equal footing with respect to quantitative and experimental traditions by illustrating that Bayesianism provides a universally applicable inferential framework.
The most important step in social science research is the first step – finding a topic. Unfortunately, little guidance on this crucial and difficult challenge is available. Methodological studies and courses tend to focus on theory testing rather than theory generation. This book aims to redress that imbalance. The first part of the book offers an overview of the book's central concerns. How do social scientists arrive at ideas for their work? What are the different ways in which a study can contribute to knowledge in a field? The second part of the book offers suggestions about how to think creatively, including general strategies for finding a topic and heuristics for discovery. The third part of the book shows how data exploration may assist in generating theories and hypotheses. The fourth part of the book offers suggestions about how to fashion disparate ideas into a theory.
Empirical research in operations management (OM) has made rapid strides in the last 30 years, and increasingly, OM researchers are leveraging methods used in the econometrics and statistics literature for assessing the causal effects of interventions. We discuss the two key challenges in assessing causality with observational data (i.e., baseline bias, differential treatment effect bias) and how dominant identification approaches such as matching, instrumental variables, regression discontinuity, difference‐in‐differences, and fixed effects deal with such challenges. We surface the key underlying assumptions of different causal estimation methods and discuss how OM scholars have used these methods in the last few years. We hope that reflecting on the plausibility and substantive meaning of underlying assumptions regarding different identification strategies in a particular context will lead to a better conceptualization, execution, evaluation, dissemination, and consumption of OM research. We conclude with a few thoughts that authors and reviewers may find helpful in their research as they engage in discourse related to causality. This article is protected by copyright. All rights reserved
Herein, the socio-psychological narrative of sexual harassment (SH) is critically evaluated. The notion of systemic SH in university departments of science, technology, engineering, and mathematics (STEM) is contradicted by the overwhelming (>90%) career satisfaction among female STEM academics. The Sexual Experiences Questionnaire (SEQ), central to the study of SH, inheres the nominalistic fallacy. SEQ usage deploys subjectivist methodologies, categorical ambiguity, the post hoc ergo propter hoc fallacy, and treats respondents as cyphers. Intercorrelation of SEQ factors reduces response statistics by 42%, while phase-space vector geometry indicates the SEQ does not measure SH. Personality analysis implies that serial abusers dominate the incidence of SH. The widespread notion that 20–25% of female college students suffer violent sexual assault rests on a misreading of published work. The 2016 Campus Climate Survey permits an upper limit estimate that 3.2% of female college students suffer rape at the hands of 4.3% of male student perpetrators, largely accompanied by drugs or alcohol. The 2018 National Academy (NAS) Report on sexual harassment in STEM exhibits negligent scholarship and carelessly generalizing statistics and may itself promote violation of the EEOC legal definition of SH. Despite instances of grievous sex-based abuse, there is no evidence that female STEM academics face systemic sexual harassment. Finally, evolutionary psychology and the social significance of personality provide a scientific understanding of SH.
This graduate textbook on machine learning tells a story of how patterns in data support predictions and consequential actions. Starting with the foundations of decision making, we cover representation, optimization, and generalization as the constituents of supervised learning. A chapter on datasets as benchmarks examines their histories and scientific bases. Self-contained introductions to causality, the practice of causal inference, sequential decision making, and reinforcement learning equip the reader with concepts and tools to reason about actions and their consequences. Throughout, the text discusses historical context and societal impact. We invite readers from all backgrounds; some experience with probability, calculus, and linear algebra suffices.
Jeffrey Friedman’s Power Without Knowledge argues forcefully that there are inherent limitations to the predictability of human action, due to a circumstance he calls “ideational heterogeneity.” However, our resources for predicting human action somewhat reliably in the light of ideational heterogeneity have not been exhausted yet, and there are no in-principle barriers to progress in tackling the problem. There are, however, other strong reasons to think that disagreement among epistocrats is bound to persist, such that it will be difficult to decide who has “the right answer” to a given technocratic problem. These reasons have to do with competing visions of the good society, fact/value entanglement, and the fragility of the facts of the social sciences.
Happiness/well-being researchers who use quantitative analysis often do not give persuasive reasons why particular variables should be included as controls in their cross-sectional models. One commonly sees notions of a “standard set” of controls, or the “usual suspects”, etc. These notions are not coherent and can lead to results that are significantly biased with respect to a genuine causal relationship.
This article presents some core principles for making more effective decisions of that sort. The contribution is to introduce a framework (the “causal revolution”, e.g. Pearl and Mackenzie 2018) unfamiliar to many social scientists (though well established in epidemiology) and to show how it can be put into practice for empirical analysis of causal questions. In simplified form, the core principles are: control for confounding variables, and do not control for intervening variables or colliders. A more comprehensive approach uses directed acyclic graphs (DAGs) to discern models that meet a minimum/efficient criterion for identification of causal effects.
The article demonstrates this mode of analysis via a stylized investigation of the effect of unemployment on happiness. Most researchers would include other determinants of happiness as controls for this purpose. One such determinant is income—but income is an intervening variable in the path from unemployment to happiness, and including it leads to substantial bias. Other commonly-used variables are simply unnecessary, e.g. religiosity and sex. From this perspective, identifying the effect of unemployment on happiness requires controlling only for age and education; a small (parsimonious) model is evidently preferable to a more complex one in this instance.
This paper offers an overview of methods for causal inference in electoral studies. We first summarize typical definitions of causality used by social scientists in Japan and then discuss their limitations in application. Next, we review a definition of causality based on potential outcomes and extend its principle to demonstrate the problems of a conventional method combining regression analysis and cross-sectional data for causal inference. Further, we introduce three sets of approaches for building a causal relationship and review their applications in electoral studies. Those approaches include an experimental method, a quasiexperimental method, and a statistical method.
The recent literature on evaluating manpower training programs demonstrates that alternative nonexperimental estimators of the same program produce a array of estimates of program impact. These findings have led to the call for experiments to be used to perform credible program evaluations. Missing in all of the recent pessimistic analyses of nonexperimental methods is any systematic discussion of how to choose among competing estimators. This paper explores the value of simple specification tests in selecting an appropriate nonexperimental estimator. A reanalysis of the National Supported Work Demonstration Data previously analyzed by proponents of social experiments reveals that a simple testing procedure eliminates the range of nonexperimental estimators that are at variance with the experimental estimates of program impact.
Theories of regime-society relations in Communist states stress the central role of coercion in maintaining political control. Based on a survey of Soviet emigrants, we examine whether Soviet citizens are deterred from nonconformity by the punitive actions of the KGB (individual deterrence), a perception of the KGB's coercive potential (general deterrence), or mistrust of other people. We find that few respondents were directly coerced by the KGB (and those who were had engaged in the most serious kinds of nonconformity); that those who had punitive contacts with the KGB in the past were not deterred from subsequent nonconformity; that the KGB's competent image was a general deterrent; and that trust in other people facilitated both nonconformist and compliant political activism. Those who came of political are under Khrushchev and Brezhnev were more likely to be involved in both kinds of activism than those who came of age under Stalin. NOTE: The original data and documentation for this study are archived at the Inter-University Consortium for Political and Social Research (ICPSR) in Ann Arbor, MI: http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/8694
I test several hypotheses concerning the origins of political repression in the states of the United States. The hypotheses are drawn from the elitist theory of democracy, which asserts that repression of unpopular political minorities stems from the intolerance of the mass public, the generally more tolerant elites not supporting such repression. Focusing on the repressive legislation adopted by the states during the McCarthy era, I examine the relationships between elite and mass opinion and repressive public policy. Generally it seems that elites, not masses, were responsible for the repression of the era. These findings suggest that the elitist theory of democracy is in need of substantial theoretical reconsideration, as well as further empirical investigation.
This study investigates empirically the strengths and limitations of using experimental versus nonexperimental designs for evaluating employment and training programs. The assessment involves comparing results from an experimental-design study-the National Supported Work Demonstration-with the estimated impacts of Supported Work based on analyses using comparison groups constructed from the Current Population Surveys. The results indicate that nonexperimental designs cannot be relied on to estimate the effectiveness of employment programs. Impact estimates tend to be sensitive both to the comparison group construction methodology and to the analytic model used. There is currently no way a priori to ensure that the results of comparison group studies will be valid indicators of the program impacts.
Genes and operons that encode bacterial virulence factors are often subject to coordinate regulation. These regulatory systems are capable of responding to various environmental signals that may be encountered during the infectious cycle. For some pathogens, proteins that mediate sensory transduction and virulence control are similar to components of other bacterial information processing systems. Understanding the molecular mechanisms governing global regulation of pathogenicity is essential for understanding bacterial infectious diseases.
Salmonella bacteria are capable of entering (invading) and multiplying within eukaryotic cells. Stable adherence to and invasion
of epithelial cells by S. choleraesuis and S. typhimurium were found to require de novo synthesis of several new bacterial
proteins. This inducible event appears to be a coordinately regulated system dependent on trypsin- and neuraminidase-sensitive
structures present on the epithelial cell surface. Mutants of S. choleraesuis and S. typhimurium were unable to synthesize
these proteins and did not stably adhere to nor invade eukaryotic cells. Two such S. typhimurium mutants were avirulent in
mice, an indication that these proteins are required for Salmonella virulence.