While many philosophers of science have accorded special evidential significance to tests whose results are "novel facts", there continues to be disagreement over both the definition of novelty and why it should matter. The view of novelty favored by Giere, Lakatos, Worrall and many others is that of use-novelty: An accordance between evidence e and hypothesis h provides a genuine test of h only if e is not used in h's construction. I argue that what lies behind the intuition (...) that novelty matters is the deeper intuition that severe tests matter. I set out a criterion of severity akin to the notion of a test's power in Neyman-Pearson statistics. I argue that tests which are use-novel may fail to be severe, and tests that are severe may fail to be use-novel. I discuss the 1919 eclipse data as a severe test of Einstein's law of gravity. (shrink)

A family of familiar linguistic tests purport to help identify when a term is ambiguous. These tests are philosophically important: a familiar philosophical strategy is to claim that some phenomenon is disunified and its accompanying term is ambiguous. The tests have been used to evaluate disunification proposals about causation, pain, and knowledge, among others. -/- These ambiguity tests, however, have come under fire. It has been alleged that the tests fail for polysemy, a common type (...) of ambiguity, and one that is at issue in philosophically interesting cases. Furthermore, the objection that the tests fail for polysemy is often taken to be an undeniable bit of linguistic data. -/- We argue that this is mistaken. The objection implicitly relies on controversial assumptions about how to account for copredicational sentences, in which a single argument is ascribed prima facie incompatible properties. Furthermore, on several viable theories of copredication, the objection fails. However, our discussion also reveals that even if ambiguity tests are preserved, they may be significantly harder to execute than previously thought. (shrink)

Research risks have to meet minimal risk requirements in order for the research to qualify for expedited ethics review, to be exempted from ethics review, or to be granted consent waivers. The definition of “minimal risk” in the Common Rule (45 CFR 46) relies on the risks-of-daily-life and risks-of-routine-tests as comparators against which research activities are assessed to meet minimal risk requirements. While either or both comparators have been adopted by major ethics codes, they have also been criticized. In (...) response to criticisms, elaborations, and alternative comparators have been proposed. In this paper, I approach the search for workable comparators from the point of view that ethical reasoning about minimal risk involves analogical reasoning using comparators. In this regard, I develop two necessary conditions for an adequate minimal risk conception, which I use to assess three comparators. I conclude that the risks-of-routine-tests best fits the analogical reasoning operating in minimal risk assessments. (shrink)

One of the chief branches of Industrial Psychology is Vocational Guidance. This attempts to discover the capacities required for different types of work and to guide young people into occupations for which their endowments fit them. Investigations have been made to determine vocational fitness in a variety of occupations; for instance, in many engineering processes, in printing, in telegraphic and telephonic work, and in clerical occupations. The article which follows gives a practical illustration of some of the methods used in (...) ovcational investigations. (shrink)

The robustness of the classical tests on means (Z, t, and F) and variances (chi square and F) was investigated by obtaining 30,000 (or, sometimes, 10,000 or 150,000) values of the test statistic under assumption-violating conditions and comparing the actual proportion of Type I errors with the proportion expected when all assumptions are met. The sampling and testing conditions investigated were: population shape (L-shape or bell-shape), relative population variance (1 or 4), sample size (8, 16, or 24), nominal significance (...) level (.05,.01, or.001), and location of rejection region (left-tailed, right-tailed, or two-tailed). All tests on means were nonrobust under most of the investigated conditions, and one was nonrobust under all of them. Both tests on variances were extremely nonrobust under virtually all conditions. The worst nonrobustness of the one- and two-independent-sample Z and t tests rivaled or exceeded that of the notoriously nonrobust chi-square and F tests on variances. (shrink)

Preregistration has been proposed as a useful method for making a publicly verifiable distinction between confirmatory hypothesis tests, which involve planned tests of ante hoc hypotheses, and exploratory hypothesis tests, which involve unplanned tests of post hoc hypotheses. This distinction is thought to be important because it has been proposed that confirmatory hypothesis tests provide more compelling results (less uncertain, less tentative, less open to bias) than exploratory hypothesis tests. In this article, we challenge (...) this proposition and argue that there are several advantages of exploratory hypothesis tests that can make their results more compelling than those of confirmatory hypothesis tests. We also consider some potential disadvantages of exploratory hypothesis tests and conclude that their advantages can outweigh the disadvantages. We conclude that exploratory hypothesis tests avoid researcher commitment and researcher prophecy biases, reduce the probability of data fraud, are more appropriate in the context of unplanned deviations, facilitate inference to the best explanation, and allow peer reviewers to make additional contributions at the data analysis stage. In contrast, confirmatory hypothesis tests may lead to an inappropriate level of confidence in research conclusions, less appropriate analyses in the context of unplanned deviations, and greater bias and errors in theoretical inferences. (shrink)

This paper seeks to show that Achinstein's recent attempt to establish that both parties to the wave-particle debate in 19th-century optics were Bayesian conditionalizers forces us to ignore several of the key conceptual issues in that controversy-not least the role of the vera causa principle and, more important still, the role of positive evidence in securing acceptance for the wave theory of light.

In a paper published in Psychological Review, Tyler Burge has offered a unified non-mentalistic account of a wide range of social cognitive developmental findings. His proposal is that far from attributing mental states, young children attribute to humans the same kind of internal generic states of sensory registration that biologists attribute to e.g. snails and ticks. Burge’s proposal deserves close attention: it is especially challenging because it departs from both the mentalistic and all the non-mentalistic accounts of the data so (...) far. Moreover Burge has been one of the leading philosophers of mind of the past 40 years and some of his writings on the objectivity of perception display a deep understanding of the relevance of science for sharpening our understanding of the mind. After taking a close look at the developmental evidence, in particular at false-belief studies, I argue that Burge’s, 409–434, 2018) account faces severe obstacles. To give one telling example: if young children can only attribute to others sensory registrations, then it is hard to explain the evidence showing that they respond differently to an agent’s ignorance and to her false belief. (shrink)

In lieu of an abstract, here is a brief excerpt of the content:Diagnostic Criteria, Psychological Tests, and Ratings Scales: Extending the HistoryPeter Zachar, PhD (bio)Le moigne narrates a history of the development of psychiatric ratings scales as hybrids between psychological tests and diagnostic categories. In his telling, psychological tests seek to quantify population-based traits on which every person has a position and which tend to be conceptualized as being stable. Personality traits are often conceptualized as dispositions. Diagnostic (...) categories represent not trait-like properties of populations but episodic states consisting of clusters of symptoms experienced by individuals with disorders. Ratings, scales, he notes, are hybrids between the two. They are used to quantify psychiatric symptom clusters so that change over time can be measured. Le Moigne argues that these distinctions were initially suggested during the development of measures that could assess outcomes in psychopharmacological treatment studies.With respect to depression, Le Moigne writes that in the 1970s and 1980s something akin to depressive personality was reconceptualized as a temperament. A key move, somewhat reflective of Kraepelin’s notion of a depressive temperament, was to view a dysthymic temperament as both a predisposing factor and an attenuated form of depression. Thus, an episodic state commandeered the predisposing–dispositional role that was assigned to stable personality traits.This conceptual shift allowed test-like ratings scales to be seen as useful for quantifying not just stable traits, but salient features of episodic states that are potentially subject to change in response to interventions. To put it another way, in the development of a psychological test, one goal is to have high test-retest reliability (i.e., stability of measurement). With rating scales used to measure change, one wants the scores to not be as stable; ideally the scores on the later assessments will be lower. According to Le Moigne, when dysthymia was subsequently placed on axis I of the Diagnostic and Statistical Manual of Mental Disorders (DSM)-III (as a syndrome) and segregated from personality disorders which were placed on axis II, personality traits were evicted from the conceptualization of depressive disorders.One perennial feature of psychiatry and psychology is that even if an older idea is superseded [End Page 253] in the view of some thinkers, it can still attract adherents. Indeed, interest in understanding a relationship between personality and attenuated states did not fully disappear. For example, based on research in genetics, the schizophrenia spectrum concept included not only schizophrenia, but also milder conditions such that were conceptualized as personality disorders; that is, schizotypal personality disorder and paranoid personality disorder (Kety, Rosenthal, Wender, & Schulsinger, 1971).During the development of the DSM-IV, some people wanted to add depressive personality disorder to the manual (Phillips, Hirschfeld, Shea, & Gunderson, 1995). It seems to have become a territorial dispute between mood disorder specialists and personality disorder specialists The mood disorders specialists challenged the personality disorder specialists to differentiate depressive personality disorder from early onset, chronic dysthymia. They were able to do so by using criteria such as critical, blaming, and derogatory toward self. After the mood disorders work group proposed incorporating some of those criteria into a revised criteria set for dysthymia, the two sets overlapped. The result was the old criteria for dysthymia were printed in DSM-IV and criteria for depressive personality disorder and the revised criteria for dysthymia were both placed in the appendix.In the DSM-5, Axis II was eliminated and personality disorders were placed alongside all the syndromic symptom clusters at the same level. Even so, interest in attenuated presentations as personality-related lingered still. During the development of the DSM-5 there was some thought to using diagnostic spectra as a meta-structure for organizing the manual (Andrews et al., 2009). One idea was to move some personality disorders into a best fitting spectrum, akin to attenuated forms. This did occur with schizotypal personality disorder which is both in the DSM-5 chapter on schizophrenia and in the chapter on personality disorder. In the schizophrenia chapter it is positioned as the earliest emerging disorder on the spectrum. There was also some interest in grouping avoidant personality disorder with the anxiety disorders, but the evidence supporting such a change was not available.These issues may... (shrink)

The most interesting feature about Mr. Richardson's criticism of my paper is that it reveals the typical attitude of the traditional intelligence tester which I set out to criticize. He accepts the view that intelligence deals mainly with the grasping of relationships, that intelligence thus defined is an innate ability, that it may be relatively isolated by the use of suitably designed tests and treated independently of other abilities.

This paper looks at the question of what it means for a psychological test to have construct validity. I approach this topic by way of an analysis of recent debates about the measurement of implicit social cognition. After showing that there is little theoretical agreement about implicit social cognition, and that the predictive validity of implicit tests appears to be low, I turn to a debate about their construct validity. I show that there are two questions at stake: First, (...) what level of detail and precision does a construct have to possess such that a test can in principle be valid relative to it? And second, what kind of evidence needs to be in place such that a test can be regarded as validated relative to a given construct? I argue that construct validity is not an all-or-nothing affair. It can come in degrees, because both our constructs and our knowledge of the explanatory relation between constructs and data can vary in accuracy and level of detail, and a test can fail to measure all of the features associated with a construct. I conclude by arguing in favor of greater philosophical attention to processes of construct development. (shrink)

We distinguish traditional cognition theories from hierarchically complex stacked neural networks that meet many of Newell's criteria. The latter are flexible and can learn anything that a person can learn, by using their mistakes and successes the same way humans do. Shortcomings are due largely to limitations of current technology.

I highlight a category of experiment-what I am calling 'demonstrations'-that differs in justificatory mode and argumentative role from the more familiar 'crucial tests'. 'Tests' are constructed such that alternative results are equally and symmetrically informative; they help discriminate between alternative solutions within a problem-field, where questions are shared. 'Demonstrations' are notably asymmetrical (for example, "failures" are often not telling), yet they are effective, if not "crucial," in interparadigm dispute, to legitimate questions themselves. The Ox-Phos Controversy in bioenergetics serves (...) as an integral case study. (shrink)

Diagnostic tests allow better informed medical decisions when there is uncertainty about a patient’s health status and, therefore, about the desirability to undertake treatment. This paper studies the relation between the expected value of diagnostic information and a patient's risk aversion. We show that the ex ante value of diagnostic information increases with risk aversion for diseases with low prevalence, but decreases with risk aversion for diseases with high prevalence. On the other hand, the ex post value of diagnostic (...) information always increases with the patient's degree of risk aversion. (shrink)

While I concur that the conclusions reached in the target article by Burger and Kass (2009) in this issue are essentially on target, the authors glossed over an important consideration in ethics an...

This experiment investigated the use of positive and negative hypothesis and target tests by groups in an adaptation of the 2-4-6 Wason task. The experimental variables were range of rule (small vs large), amount of evidence (low vs high), and trial block (1 vs 2). The results were in accordance with Klayman and Ha's (1987) analysis of base rate probabilities of falsification and with additional theoretical considerations. Base rate probabilities were more descriptive of participants' behaviour in target than in (...) hypothesis tests, under low than under high amount of evidence, and at the beginning of the process than at its end. The percentage of positive tests was higher under small than large range of rule. More falsifications than verifications resulted from hypothesis tests than would be expected by a random process. When evidence is richly available, the relative importance of falsification seems to decrease. An analysis of the group compositions before and after group discussion by the PCD model (Crott, Werner, & Hoffmann, 1996) revealed that the normative weight was approximately twice as large as the informational. Groups produced fewer false answers than their members individually. (shrink)

On a literal reading of `Computing Machinery and Intelligence'', Alan Turing presented not one, but two, practical tests to replace the question `Can machines think?'' He presented them as equivalent. I show here that the first test described in that much-discussed paper is in fact not equivalent to the second one, which has since become known as `the Turing Test''. The two tests can yield different results; it is the first, neglected test that provides the more appropriate indication (...) of intelligence. This is because the features of intelligence upon which it relies are resourcefulness and a critical attitude to one''s habitual responses; thus the test''s applicablity is not restricted to any particular species, nor does it presume any particular capacities. This is more appropriate because the question under consideration is what would count as machine intelligence. The first test realizes a possibility that philosophers have overlooked: a test that uses a human''s linguistic performance in setting an empirical test of intelligence, but does not make behavioral similarity to that performance the criterion of intelligence. Consequently, the first test is immune to many of the philosophical criticisms on the basis of which the (so-called) `Turing Test'' has been dismissed. (shrink)

Susan Schneider (2019) has proposed two new tests for consciousness in AI (artificial intelligence) systems, the AI Consciousness Test and the Chip Test. On their face, the two tests seem to have the virtue of proving satisfactory to a wide range of consciousness theorists holding divergent theoretical positions, rather than narrowly relying on the truth of any particular theory of consciousness. Unfortunately, both tests are undermined in having an ‘audience problem’: Those theorists with the kind of architectural (...) worries that motivate the need for such tests should, on similar grounds, doubt that the tests establish the existence of genuine consciousness in the AI in question. Nonetheless, the proposed tests constitute progress, as they could find use by some theorists holding fitting views about consciousness and perhaps in conjunction with other tests for AI consciousness. (shrink)