Philosophy of Statistics

Different inferences in probabilistic logics of conditionals 'preserve' the probabilities of their premisses to different degrees. Some preserve certainty, some high probability, some positive probability, and some minimum probability. In the first case conclusions must have probability I when premisses have probability 1, though they might have probability 0 when their premisses have any lower probability. In the second case, roughly speaking, if premisses are highly probable though not certain then conclusions must also be highly probable. In the third case (...) conclusions must have positive probability when premisses do, and in the last case conclusions must be at least as probable as their least probable premisses. Precise definitions and well known examples are given for each of these properties, characteristic principles are shown to be valid and complete for deriving conclusions of each of these kinds, and simple trivalent truthtable 'tests' are described for determining which properties are possessed by any given inference. Brief comments are made on the application of these results to certain modal inferences such as "Jones may own a car, and if he does he will have a driver's license. Therefore, he may have a driver's license.". (shrink)

In the epigraph, Fisher is blaming two generations of theoretical biologists, from Darwin on, for ignoring Quetelet's statistical techniques and hence harboring confusions about evolution and natural selection. He is right to imply that Darwin and his contemporaries were aware of the core of Quetelet's work. Quetelet's seminal monograph, Sur L'homme, was widely discussed in Darwin's academic circles. We know that Darwin owned a copy (Schweber 1977). More importantly, we have in Darwin's notebooks two entries referring to Quetelet's work on (...) the cause of a large-scale global phenomenon where each year more boys were born than girls. The first entry is written sometime between April and July 1838. Darwin writes "Find out from the Statistical Society—where M. Quetelet has published his laws about sexes relative to age of Marriages" (C 268, Barrett et al., 1987, p. 324). The second is written sometime after October 16, 1838: "In the Atheneum Numbers 406, 407, 409, Quetelet papers are given, & I think facts there mentioned about proportion of sexes, at birth & causes" (Ibid, p. 379). So, even if Darwin did not read Sur L'homme directly it is likely (though not certain) that he read its review in the Atheneum. There is no doubt that Darwin eventually became familiar with Quetelet's work in statistics. The smoking gun is an essay that Darwin writes in 1874, entitled, "On the Males and Complemental Males of Certain Cirripedes, and on Rudimentary Structures" where he discusses Quetelet's laws of variation. (shrink)

Richard Jeffrey’s radical probabilism (‘probability all the way down’) is augmented by the claim that probability cannot be turned into certainty, except by data that logically exclude all alternatives. Once we start being uncertain, no amount of updating will free us from the treadmill of uncertainty. This claim is cast first in objectivist and then in subjectivist terms.

We show that it is possible to generalize the Ursell–Mayer cluster formalism so that it may cover also the statistics of Internet websites. Our starting point is the introduction of an extra variable that is assumed to take account, as will be explained, of the nature of the Internet statistics. We then show, following the arguments in Mayer, that one may obtain a phase transition-like phenomena.

This paper takes a critical look at theory-free, statistical methodologies for processing and interpreting data taken from respondents answering a set of dichotomous (yes-no) questions. The basic issue concerns to what extent theoretical conclusions based on such analyses are invariant under a class of "informationally equivalent" question transformations. First the notion of Boolean equivalence of two question sets is discussed. Then Lazarsfeld's latent structure analysis is considered in detail. It is discovered that the best fitting latent model depends on which (...) one of the many informationally equivalent question sets is used. This fact raises a number of methodological problems and pitfalls with latent structure analysis. Related problems with other methodologies are briefly discussed. (shrink)

It is sometimes claimed that the Bayesian framework automatically implements Ockham's razor---that conditionalizing on data consistent with both a simple theory and a complex theory more or less inevitably favours the simpler theory. It is shown here that the automatic razor doesn't in fact cut it for certain mundane curve-fitting problems.

Cancer research is experiencing ‘paradigm instability’, since there are two rival theories of carcinogenesis which confront themselves, namely the somatic mutation theory and the tissue organization field theory. Despite this theoretical uncertainty, a huge quantity of data is available thanks to the improvement of genome sequencing techniques. Some authors think that the development of new statistical tools will be able to overcome the lack of a shared theoretical perspective on cancer by amalgamating as many data as possible. We think instead (...) that a deeper understanding of cancer can be achieved by means of more theoretical work, rather than by merely accumulating more data. To support our thesis, we introduce the analytic view of theory development, which rests on the concept of plausibility, and make clear in what sense plausibility and probability are distinct concepts. Then, the concept of plausibility is used to point out the ineliminable role played by the epistemic subject in the development of statistical tools and in the process of theory assessment. We then move to address a central issue in cancer research, namely the relevance of computational tools developed by bioinformaticists to detect driver mutations in the debate between the two main rival theories of carcinogenesis. Finally, we briefly extend our considerations on the role that plausibility plays in evidence amalgamation from cancer research to the more general issue of the divergences between frequentists and Bayesians in the philosophy of medicine and statistics. We argue that taking into account plausibility-based considerations can lead to clarify some epistemological shortcomings that afflict both these perspectives. (shrink)

When observing or measuring phenomena, errors are inevitable, one can only aspire to reduce these errors as much as possible. An obvious strategy to achieve this reduction is by using more precise instruments. Another strategy was to develop a theory of these errors that could indicate how to take them into account. One of the greatest achievements of statistics in the beginning of the 19th century was such a theory of error. This theory told the practitioners that the best thing (...) they could do is taking the arithmetical mean of their observations. This average would give them the most accurate estimate of the value they were searching for. Soon after its invention, this method made a triumphal march across various sciences. However, not in all sciences one stood waving aside. This method, namely, only worked well when the various observations were made under similar circumstances and when there were very many of them. And this was not the case for e.g. meteorology and actuarial science, the two sciences discussed in this paper. (shrink)

Entertaining diverse assumptions about empirical research, commentators give a wide range of verdicts on the NHSTP defence in Statistical significance. The null-hypothesis significance- test procedure is defended in a framework in which deductive and inductive rules are deployed in theory corroboration in the spirit of Popper's Conjectures and refutations. The defensible hypothetico-deductive structure of the framework is used to make explicit the distinctions between substantive and statistical hypotheses, statistical alternative and conceptual alternative hypotheses, and making statistical decisions and drawing theoretical (...) conclusions. These distinctions make it easier to show that H0 can be true, the effect size is irrelevant to theory corroboration, and “strong” hypotheses make no difference to NHSTP. Reservations about statistical power, meta-analysis, and the Bayesian approach are still warranted. (shrink)

Prouver et gouverner étudie le rôle des institutions, des conventions et des enjeux normatifs dans la construction d’indicateurs quantitatifs. Desrosières pense qu’on ne peut étudier le développement scientifique des statistiques sans prendre en compte le développement institutionnel – en particulier le rôle de l’État – dans la constitution de cette discipline.

In order to perform certain actions – such as incarcerating a person or revoking parental rights – the state must establish certain facts to a particular standard of proof. These standards – such as preponderance of evidence and beyond reasonable doubt – are often interpreted as likelihoods or epistemic confidences. Many theorists construe them numerically; beyond reasonable doubt, for example, is often construed as 90 to 95% confidence in the guilt of the defendant. -/- A family of influential cases suggests (...) standards of proof should not be interpreted numerically. These ‘proof paradoxes’ illustrate that purely statistical evidence can warrant high credence in a disputed fact without satisfying the relevant legal standard. In this essay I evaluate three influential attempts to explain why merely statistical evidence cannot satisfy legal standards. (shrink)

In the face of continuing assumptions by many scientists and journal editors that p-values provide a gold standard for inference, counter warnings are published periodically. But the core problem is not with p-values, per se. A finding that “p-value is less than α” could merely signal that a critical value has been exceeded. The question is why, when estimating a parameter, we provide a range (a confidence interval), but when testing a hypothesis about a parameter (e.g. µ = x) we (...) proceed as if “=” entails exact equality of the parameter with x. That standard is hard to meet, and is not a standard expected for power calculations, where we are satisfied to reject a null hypothesis H0 if the result is merely “detectably” different from (exact) H0. This paper explores, with resampling (simulation) methods, the impacts on p-values, and alternatives, if the null hypothesis is defined as a thick or thin range of values. It also examines, empirically, the extent to which the p-value may or may not be a good predictor of the probability that H0 is true, given the distribution of the data. (shrink)

A pure significance test would check the agreement of a statistical model with the observed data even when no alternative model was available. The paper proposes the use of a modified p -value to make such a test. The model will be rejected if something surprising is observed. It is shown that the relation between this measure of surprise and the surprise indices of Weaver and Good is similar to the relationship between a p -value, a corresponding odds-ratio, and a (...) logit or log-odds statistic. The s -value is always larger than the corresponding p -value, and is not uniformly distributed. Difficulties with the whole approach are discussed. (shrink)

Papers in experimental philosophy rarely offer an account of what it would take to reveal a philosophically significant effect. In part, this is because experimental philosophers tend to pay insufficient attention to the hierarchy of models that would be required to justify interpretations of their data; as a result, some of their most exciting claims fail as explanations. But this does not impugn experimental philosophy. My aim is to show that experimental philosophy could be made more successful by developing, articulating, (...) and advancing plausible models of the data that are collected and the analyses that are employed. (shrink)

John Earman and John T. Roberts advocate a challenging and radical claim regarding the semantics of laws in the special sciences: the statistical account. According to this account, a typical special science law “asserts a certain precisely defined statistical relation among well-defined variables” and this statistical relation does not require being hedged by ceteris paribus conditions. In this paper, we raise two objections against the attempt to cash out the content of special science generalizations in statistical terms.

Over the past two decades, several consistent procedures have been designed to infer causal conclusions from observational data. We prove that if the true causal network might be an arbitrary, linear Gaussian network or a discrete Bayes network, then every unambiguous causal conclusion produced by a consistent method from non-experimental data is subject to reversal as the sample size increases any finite number of times. That result, called the causal flipping theorem, extends prior results to the effect that causal discovery (...) cannot be reliable on a given sample size. We argue that since repeated flipping of causal conclusions is unavoidable in principle for consistent methods, the best possible discovery methods are consistent methods that retract their earlier conclusions no more than necessary. A series of simulations of various methods across a wide range of sample sizes illustrates concretely both the theorem and the principle of comparing methods in terms of retractions. (shrink)

In this paper I provide a frequentist philosophical-methodological solution for the stopping rule problem presented by Lindley & Phillips in 1976, which is settled in the ecological realm of testing koalas’ sex ratio. I deliver criteria for discerning a stopping rule, an evidence and a model that are epistemically more appropriate for testing the hypothesis of the case studied, by appealing to physical notion of probability and by analyzing the content of possible formulations of evidence, assumptions of models and meaning (...) of the ecological hypothesis. First, I show the difference in the evidence taken into account in different frequentist sampling procedures presented in the problem. Next, I discuss the inapplicability of the Carnapian principle of total evidence in deciding which formulation of evidence associated with a given sampling procedure and statistical model is epistemically more appropriate for testing the hypothesis in question. Then I propose a double-perspective (evidence and model) frequentist solution based on the choice of evidence which better corresponds to the investigated ecological hypothesis, as well as on the choice of a model that embraces less unrealistic ontological assumptions. Finally, I discuss two perspectives of the stopping rule dependence. (shrink)

Proponents of evidence-based medicine (EBM) provide the “hierarchy of evidence” as a criterion for judging the reliability of therapeutic decisions. EBM's hierarchy places randomized interventional studies (and systematic reviews of such studies) higher in the hierarchy than observational studies, unsystematic clinical experience, and basic science. Recent philosophical work has questioned whether EBM's special emphasis on evidence from randomized interventional studies can be justified. Following the critical literature, and in particular the work of John Worrall, I agree that many of the (...) arguments put forward by advocates of EBM do not justify the ambitious claims that are often made on behalf of randomization. However, in contrast to the recent philosophical work, I argue that a justification for EBM's hierarchy of evidence can be provided. The hierarchy should be viewed as a hierarchy of comparative internal validity. Although this justification is defensible, the claims that EBM's hierarchy substantiates when viewed in this way are considerably more circumscribed than some claims found in the EBM literature. (shrink)

The main thesis of the paper is that in the case of modern statistics, the differences between the various concepts of models were the key to its formative controversies. The mathematical theory of statistical inference was mainly developed by Ronald A. Fisher, Jerzy Neyman, and Egon S. Pearson. Fisher on the one side and Neyman–Pearson on the other were involved often in a polemic controversy. The common view is that Neyman and Pearson made Fisher's account more stringent mathematically. It is (...) argued, however, that there is a profound theoretical basis for the controversy: both sides held conflicting views about the role of mathematical modelling. At the end, the influential programme of Exploratory Data Analysis is considered to be advocating another, more instrumental conception of models. Introduction Models in statistics—‘of what population is this a random sample?’ The fundamental lemma Controversy about models Exploratory data analysis as a model-critical approach. (shrink)

ABSTRACT. Probability and statistics play an important role in contemporary -philosophy of causality. They are viewed as glasses through which we can see or detect causal relations. However, they may sometimes act as blinding glasses, as I will argue in this paper. In the 19th century, Francis Galton tried to statistically analyze hereditary phenomena. Although he was a far better statistician than Gregor Mendel, his biological theory turned out to be less fruitful. This was no sheer accident. His knowledge of (...) statistics generated two explananda (unknown to Mendel) which in turn generated constraints for any theory of heredity. These constraints misguided Galtons search for the causal mechanism of inheritance. This is not just. an interesting case for philosophers and historians of ·science. Notwithstanding the progress made by statitics, Galtons problem is still relevant today. In the final section, I briefly explore the implications for statistical techniques such as structural equation modelling. (shrink)

“Dutch Book” arguments and references to gambling theorems are typical in the debate between Bayesians and scientists committed to “classical” statistical methods. These arguments have rarely convinced non-Bayesian scientists to abandon certain conventional practices, partially because many scientists feel that gambling theorems have little relevance to their research activities. In other words, scientists “don’t bet.” This article examines one attempt, by Schervish, Seidenfeld, and Kadane, to progress beyond such apparent stalemates by connecting “Dutch Book”–type mathematical results with principles actually endorsed (...) by practicing experimentalists. (shrink)

This article concerns the psychology of the paradoxical Two Envelope Problem. The goal is to find instructive variants of the envelope switching problem that are capable of clear-cut resolution, while still retaining paradoxical features. By relocating the original problem into different contexts involving commutes and playing cards the reader is presented with a succession of resolved paradoxes that reduce the confusion arising from the parent paradox. The goal is to reduce confusion by understanding how we sometimes misread mathematical statements; or, (...) to completely avoid confusion, either by reforming language, or adopting an unambiguous notation for switching problems. This article also suggests that an illusion close in character to the figure/ground illusion hampers our understanding of switching problems in general and helps account for the intense confusion that switching problems sometimes generate. (shrink)

I document some of the main evidence showing that E. S. Pearson rejected the key features of the behavioral-decision philosophy that became associated with the Neyman-Pearson Theory of statistics (NPT). I argue that NPT principles arose not out of behavioral aims, where the concern is solely with behaving correctly sufficiently often in some long run, but out of the epistemological aim of learning about causes of experimental results (e.g., distinguishing genuine from spurious effects). The view Pearson did hold gives a (...) deeper understanding of NPT tests than their typical formulation as accept-reject routines, against which criticisms of NPT are really directed. The Pearsonian view that emerges suggests how NPT tests may avoid these criticisms while still retaining what is central to these methods: the control of error probabilities. (shrink)

A review of A Philosophical Guide to Chance by Toby Handfield. Cambridge CUP 2012.

Simple regression (regression analysis with a single explanatory variable), and multiple regression (regression models with multiple explanatory variables), typically correspond to very different biological questions. The former use regression lines to describe univariate associations. The latter describe the partial, or direct, effects of multiple variables, conditioned on one another. We suspect that the superficial similarity of simple and multiple regression leads to confusion in their interpretation. A clear understanding of these methods is essential, as they underlie a large range of (...) procedures in common use in biology. Beyond simple and multiple regression in their most basic forms, understanding the key principles of these procedures is critical to understanding, and properly applying, many methods, such as mixed models, generalised models, and causal inference using graphs (including path analysis and its extensions). A simple, but careful, look at the distinction between these two analyses is valuable in its own right, and can also be used to clarify widely-held misconceptions about collinearity (correlations among explanatory variables). There is no general sense in which collinearity is a problem. We suspect that the perception of collinearity as a hindrance to analysis stems from misconceptions about interpretation of multiple regression models, and so we pursue discussions about these misconceptions in this light. In particular, collinearity causes multiple regression coefficients to be less precisely estimated than corresponding simple regression coefficients. This should not be interpreted as a problem, as it is perfectly natural that direct effects should be harder to characterise than univariate associations. Purported solutions to the perceived problems of collinearity are detrimental to most biological analyses. (shrink)

Let an unbiased coin be used to form an ω-sequence S of independent tosses. Let N be the positive integers. The finite initial segment of length n ∈ N is denoted by Sn (thus, S1 holds exactly the first toss). For n ∈ N , let Hn be the proportion of heads that show up in Sn.

About a British Academy collection of papers on Bayes' famous theorem.

I give a pedagogical derivation of the Cramer-Rao Bound, which gives a lower bound on the variance of estimators used in statistical point estimation, commonly used to give numerical estimates of the systematic uncertainties in a measurement.

The p value is the probability under the null hypothesis of obtaining an experimental result that is at least as extreme as the one that we have actually obtained. That probability plays a crucial role in frequentist statistical inferences. But if we take the word ‘extreme’ to mean ‘improbable’, then we can show that this type of inference can be very problematic. In this paper, I argue that it is a mistake to make such an interpretation. Under minimal assumptions about (...) the alternative hypothesis, I explain why ‘extreme’ means ‘outside the most precise predicted range of experimental outcomes for a given upper bound probability of error’. Doing so, I rebut recent formulations of recurrent criticisms against the frequentist approach in statistics and underscore the importance of random variables. (shrink)

An inductive logic is a system of inference that describes the relation between propositions on data, and propositions that extend beyond the data, such as predictions over future data, and general conclusions on all possible data. Statistics, on the other hand, is a mathematical discipline that describes procedures for deriving results about a population from sample data. These results include predictions on future samples, decisions on rejecting or accepting a hypothesis about the population, the determination of probability assignments over such (...) hypotheses, the selection of a statistical model for studying the population, and so on. Both inductive logic and statistics are calculi for getting from the given data to propositions or results that transcend the data. (shrink)

Advocates of the self-corrective thesis argue that scientific method will refute false theories and find closer approximations to the truth in the long run. I discuss a contemporary interpretation of this thesis in terms of frequentist statistics in the context of the behavioral sciences. First, I identify experimental replications and systematic aggregation of evidence (meta-analysis) as the self-corrective mechanism. Then, I present a computer simulation study of scientific communities that implement this mechanism to argue that frequentist statistics may converge upon (...) a correct estimate or not depending on the social structure of the community that uses it. Based on this study, I argue that methodological explanations of the “replicability crisis” in psychology are limited and propose an alternative explanation in terms of biases. Finally, I conclude suggesting that scientific self-correction should be understood as an interaction effect between inference methods and social structures. (shrink)

: Bayesians and error statisticians have relied heavily upon examples from physics in developing their accounts of scientific inference. The present essay demonstrates it is possible to analyze H.B.D. Kettlewell's classic study of natural selection from Deborah Mayo's error statistical point of view (Mayo 1996). A comparison with a previous analysis of this episode from a Bayesian perspective (Rudge 1998) reveals that the error statistical account makes better sense of investigations such as Kettlewell's because it clarifies how core elements in (...) the design of experiments are used to minimize erroneous inferences rather than dwelling on whether the strategies used are reasonable. (shrink)

The hydrodynamical formalism for the quantum theory of a nonrelativistic particle is considered, together with a reformulation of it which makes use of the methods of kinetic theory and is based on the existence of the Wigner phase-space distribution. It is argued that this reformulation provides strong evidence in favor of the statistical interpretation of quantum mechanics, and it is suggested that this latter could be better understood as an almost classical statistical theory. Moreover, it is shown how, within this (...) context, the Wigner and the Margenau-Hill functions are not equivalent, and that the latter is essentially unsatisfactory, as well as the associated symmetrization rule. Arguments in favor of a stochastic picture of the phenomena at the microscopic level are also presented. (shrink)

Amidst long-running debates within the field, high energy physics has adopted a statistical methodology that primarily employs standard frequentist techniques such as significance testing and confidence interval estimation, but incorporates Bayesian methods for limited purposes. The discovery of the Higgs boson has drawn increased attention to the statistical methods employed within HEP. Here I argue that the warrant for the practice in HEP of relying primarily on frequentist methods can best be understood as pragmatic, in the sense that statistical methods (...) are chosen for their suitability for the practical demands of experimental HEP, rather than reflecting a commitment to a particular epistemic framework. In particular, I argue that understanding the statistical methodology of HEP through the perspective of pragmatism clarifies the role of and rationale for significance testing in the search for new phenomena such as the Higgs boson. (shrink)

Some accounts of evidence regard it as an objective relationship holding between data and hypotheses, perhaps mediated by a testing procedure. Mayo’s error-statistical theory of evidence is an example of such an approach. Such a view leaves open the question of when an epistemic agent is justified in drawing an inference from such data to a hypothesis. Using Mayo’s account as an illustration, I propose a framework for addressing the justification question via a relativized notion, which I designate security , (...) meant to conceptualize practices aimed at the justification of inferences from evidence. I then show how the notion of security can be put to use by showing how two quite different theoretical approaches to model criticism in statistics can both be viewed as strategies for securing claims about statistical evidence. (shrink)

Stopping rules — rules dictating when to stop accumulating data and start analyzing it for the purposes of inferring from the experiment — divide Bayesians, Likelihoodists and classical statistical approaches to inference. Although the relationship between Bayesian philosophy of science and stopping rules can be complex (cf. Steel 2003), in general, Bayesians regard stopping rules as irrelevant to what inference should be drawn from the data. This position clashes with classical statistical accounts. For orthodox statistics, stopping rules do matter to (...) what inference should be drawn from the data. "The dispute over stopping rule is far from being a marginal quibble, but is instead a striking illustration of the divergence of fundamental aims and standards separating Bayesians and advocates of orthodox statistical methods" (Steel 2004, 195) ... (shrink)

Philosophers subscribing to particular principles of statistical inference and evidence need to be aware of the limitations and practical consequences of the statistical approach they endorse. The framework proposed (for statistical inference in the field of medicine) allows disparate statistical approaches to emerge in their appropriate context. My dissertation proposes a decision theoretic model, together with methodological guidelines, that provide important considerations for deciding on clinical trial conduct. These considerations do not amount to more stopping rules. Instead, they are principles (...) that address the complexity of interpreting and responding to interim data, based on a broad range of epistemic and ethical factors. While they are not stopping rules, they would assist a Data Monitoring Committee in judging its position with regard to necessary precautionary interpretation of interim data. By vindicating a framework that accommodates a wide range of approaches to statistical inference in one important setting (clinical trials), my results pose a serious challenge for any approach that advocates a single, universal principle of statistical inference. (shrink)

This paper analyzes statistical decisions during the interim analyses of clinical trials. After some general remarks about the ethical and scientific demands of clinical trials, I introduce the notion of a hard-case clinical trial, explain the basic idea behind it, and provide a real example involving the interim analyses of zidovudine in asymptomatic HIV-infected patients. The example leads me to propose a decision analytic framework for handling ethical conflicts that might arise during the monitoring of hard-case clinical trials. I use (...) computer simulations to show how the framework can assist in reconciling certain ethical conflicts. The framework is partial, lacking the precision of a complete systematization of statistical monitoring procedures in practice. (shrink)

The question of how the probabilistic opinions of different individuals should be aggregated to form a group opinion is controversial. But one assumption seems to be pretty much common ground: for a group of Bayesians, the representation of group opinion should itself be a unique probability distribution, 410–414, [45]; Bordley Management Science, 28, 1137–1148, [5]; Genest et al. The Annals of Statistics, 487–501, [21]; Genest and Zidek Statistical Science, 114–135, [23]; Mongin Journal of Economic Theory, 66, 313–351, [46]; Clemen and (...) Winkler Risk Analysis, 19, 187–203, [7]; Dietrich and List [14]; Herzberg Theory and Decision, 1–19, [28]). We argue that this assumption is not always in order. We show how to extend the canonical mathematical framework for pooling to cover pooling with imprecise probabilities by employing set-valued pooling functions and generalizing common pooling axioms accordingly. As a proof of concept, we then show that one IP construction satisfies a number of central pooling axioms that are not jointly satisfied by any of the standard pooling recipes on pain of triviality. Following Levi, 3–11, [39]), we also argue that IP models admit of a much better philosophical motivation as a model of rational consensus. (shrink)

Maxwell's deduction of the probability distribution over the velocity of gas molecules—one of the most important passages in physics (Truesdell)—presents a riddle: a physical discovery of the first importance was made in a single inferential leap without any apparent recourse to empirical evidence. -/- Tychomancy proposes that Maxwell's derivation was not made a priori; rather, he inferred his distribution from non-probabilistic facts about the dynamics of intermolecular collisions. Further, the inference is of the same sort as everyday reasoning about the (...) physical probabilities attached to such canonical chance setups as tossed coins or rolled dice. The structure of this reasoning is investigated and some simple rules for inferring physical probabilities from symmetries and other causally relevant properties of physical systems are proposed. -/- Not only physics but evolutionary biology and population ecology, the science of measurement error, and climate modeling have benefited enormously from the same kind of reasoning, the book goes on to argue. Inferences from dynamics to probability are so obvious to us, however, that their methodological importance has been largely overlooked. (shrink)

Elliott Sober argues that the statistical slogan “Absence of evidence is not evidence of absence” cannot be taken literally: it must be interpreted charitably as claiming that the absence of evidence is (typically) not very much evidence of absence. I offer an alternative interpretation, on which the slogan claims that absence of evidence is (typically) not objective evidence of absence. I sketch a definition of objective evidence, founded in the notion of an epistemically objective likelihood, and I show that in (...) Sober’s paradigm case, the slogan can, on this understanding, be sustained. (shrink)

Ex ante predicted outcomes should be interpreted as counterfactuals (potential histories), with errors as the spread between outcomes. But error rates have error rates. We reapply measurements of uncertainty about the estimation errors of the estimation errors of an estimation treated as branching counterfactuals. Such recursions of epistemic uncertainty have markedly different distributial properties from conventional sampling error, and lead to fatter tails in the projections than in past realizations. Counterfactuals of error rates always lead to fat tails, regardless of (...) the probability distribution used. A mere .01% branching error rate about the STD (itself an error rate), and .01% branching error rate about that error rate, etc. (recursing all the way) results in explosive (and infinite) moments higher than 1. Missing any degree of regress leads to the underestimation of small probabilities and concave payoffs (a standard example of which is Fukushima). The paper states the conditions under which higher order rates of uncertainty (expressed in spreads of counterfactuals) alters the shapes the of final distribution and shows which a priori beliefs about conterfactuals are needed to accept the reliability of conventional probabilistic methods (thin tails or mildly fat tails). (shrink)