Estimation and visualization of confusability matrices from adaptive measurement data

We present a simple but effective method based on Luce’s choice axiom [Luce, R.D. (1959). Individual choice behavior: A theoretical analysis. New York: John Wiley & Sons] for consistent estimation of the pairwise confusabilities of items in a multiple-choice recognition task with arbitrarily chosen choice-sets. The method combines the exact (non-asymptotic) Bayesian way of assessing uncertainty with the unbiasedness emphasized in the classical frequentist approach.We apply the method to data collected using an adaptive computer game designed for prevention of reading disability. A player’s estimated confusability of phonemes (or more accurately, phoneme-grapheme connections) and larger units of language is visualized in an easily understood way with color cues and explicit indication of the accuracy of the estimates. Visualization of learning-related changes in the player’s performance is considered.The empirical validity of the choice axiom is evaluated using the game data itself. The axiom appears to hold reasonably well although a small systematic violation is observable for the smallest choice-set sizes.     

Continuum modeling of the reverse Hall–Petch effect in nanocrystalline metals under uniaxial tension: how many grains in a finite element model?

Owing to their very high strength, nanocrystalline metals have been extensively studied over the recent years. The direct Hall–Petch law, empirically proportioning the material strength to the inverse square root of its grain size has been shown to break down below a grain size of the order of tenths of nanometers. This phenomenon has been widely rationalized as a gradual switch from intragrain mediated deformation mechanisms to grain boundary mediated deformation mechanisms. This transition has been observed in many finite element simulations, despite the intrinsic restriction of necessarily limiting the nanocrystalline representative assembly to only a few grains. Such a limitation is generally overlooked, and its influence on an uniaxial tension test – when compared to a complete sample of millions of grains – ignored. We propose here to quantify the approximation done by considering a finite number of grains by means of a simple analytical model based on the early work of Stevens [R.N. Stevens, Philos. Mag. 23 (1971) p. 265]. The finite element approximation is demonstrated to be relatively good, even down to only three grains in width, and a method to “correct” the stress-strain curves of small representative volumes is proposed.     

Becoming a philosopher: What Heidegger learned from Dilthey, 1919–25

Wilhelm Dilthey is, famously, an epistemological pioneer for a second, ‘human’ kind of science that ‘understands’ life as we live it, instead of ‘explaining’ things as we observe them. Even today, he is usually cited for his role in the Erklären–Verstehen debate. My article, however, follows Heidegger's suggestion that we make the existence of the debate itself the problem. Whether there are different sorts of entity, different reasons for studying them and different means for doing so – such issues raise questions about science itself, not just about how to do it better. Moreover, what sort of philosopher is competent to address such questions? Heidegger argues that Dilthey's later writings intimate that it must be one who thinks from the ‘standpoint of (historical) life itself.’ This issue, says Heidegger, is ‘alive’ in Dilthey but is continually short-circuited by his very traditional plan for a ‘Critique of Historical Reason.’ Dilthey's unsuccessful struggles to produce this Critique are his gift to us, however. They encourage us to explicitly reconsider, as Heidegger does not only in Being and Time but throughout his life, what Dilthey cannot: If philosophy, like all human practices, is historical to the core, what is it to ‘be’ philosophical, about science or anything else?     

Lost in Translation? The Need for ‘Upstream Engagement’ with Nanotechnology on Trial

Nowadays, the criticism of the so-called ‘deficit model’ and the need for ‘upstream engagement’ in science and technology are becoming part of the master narratives of public policies in many countries, especially concerning nanotechnology. This may be considered as a major success for STS scholars, whose research results have largely contributed to this change, especially those concerning the GMO controversies. Some STS scholars thus move from a position of distant and critical observers to the role of experts in social engineering or advisers of policy-makers. However, in their enthusiasm concerning the expected benefits of upstream engagement, institutions, TA practitioners and social scientists seem to ignore some important limitations as well as the implicit framing assumptions of the concept. Based on an experience made by a group of social scientists in the Grenoble area—one of the major ‘nanodistricts’ in Europe—our paper shows that the ‘upstream engagement’ concept is still embedded in a linear model of innovation and is not very useful to anyone pursuing the co-production of innovations. It is especially true when socio-technical networks are already aligned by powerful actors and a worldwide agenda as in the case of nanotechnology. In order to give an opportunity for public engagement to have a larger impact on decision-making, we propose an alternative approach, which combines Actor–Network Theory (ANT), as an analytical tool, with the reflexive and ongoing implementation of public participation. Public engagement is probably one of the critical loci where STS scholars must reflect on the articulation between the knowledge they produce and public policies in action.     

Influence of properties of layered silicate minerals on adsorbed DNA surface affinity,self-assembly and nanopatterning

We report on the specific interaction between DNA and some mica-family minerals and other layered silicate structures. The interaction depends on mineral surface's crystallography, chemistry and potential. These properties are responsible for a remarkable variety of adsorption mechanisms and can be used to modulate surface self-assembly and nanopatterning of DNA. The controlled deposition of DNA onto non-conductive mineral regions with atomically flat boundaries has very attractive applications in nano-biotechnology (for example, microelectronics, microarrays and sensors). In addition, the extreme affinity discovered for some mineral surfaces, together with their ability to organize the DNA molecules, could be an indication of their catalytic potential. It may also have had relevance in the prebiotic environment, with important implications for the earth and life sciences.     

Retrieval of names in face and object naming in an interference study

Two experiments using the interference paradigm are reported. In the first experiment, the participants spoke aloud the names of celebrities and the names of objects when presented with pictures while hearing distractors. In the case of proper names, we replicated the data obtained by Izaute and Bonin (2001) using the interference paradigm with a proper name written naming task. In the case of common names, the results replicated those obtained by Shriefers, Meyer, and Levelt (1990). In the second experiment, the participants produced the names of celebrities when presented with their faces while hearing distractors that were either proper names associated with the celebrities (associate condition), that belonged to a different professional category (different condition), or that corresponded to the proper names of the celebrities (identical condition). For negative SOAs, “associate” distractors were found to increase latencies compared to the “different category” condition. The implications of the findings for proper name retrieval are briefly discussed.     

Study repetition and the rejection of conjunction lures

Four experiments were conducted to provide further support for a dual-process approach to conjunction errors in recognition memory over a single-process approach. In Experiments 1, 2a, and 2b study words were repeated once (2 presentations) or several times (8 or 16 presentations), and a response signal delay (RSD) was manipulated (short vs long). Short RSD groups produced lower hit rates but higher conjunction error rates compared to long RSD groups. Repetition effects on conjunction errors occurred for short RSD but not long RSD groups. In Experiment 3, additional judgements on new items were solicited to show that the “null” effects for long RSD groups in the earlier experiments were, in fact, illusory. Together, the experiments provide evidence for recollection-based rejections of conjunction lures and support a dual-process theory of recognition memory over a single-process model.     

一种多级评分的广义认知诊断模型

认知诊断是近些年教育测量研究中的热点,大多数的认知诊断模型仅适用于0～1评分的情况.本文提出一种有多个潜变量多个滑动参数的多级评分认知诊断模型——GP-D1NA,只要由评分标准和知识状态能确定理想反应模式,就可以利用此方法进行认知诊断分析.在该方法中,我们给出项目滑动矩阵的概念,将被试的观测得分均看成由某个理想得分的滑动,并采用EM算法估计滑动矩阵.在模拟研究中,采用每掌握一个属性得1分的评分标准,结果表明线性型、收敛型、发散型、无结构型和独立型五种属性层级结构均有较高的判准率.     

认知诊断分类中心的确定

期望反应模式是认知诊断分类的类中心,确定了诊断范围中属性及其层级关系后,类中心的数目由测验Q矩阵决定.类中心应该是完备的,即理论上有多少个知识状态就应该有多少个类中心,这涉及评分方式、Q矩阵设计、Q矩阵评价与修改、认知诊断测验如何实施等问题的讨论.重点给出多级评分认知诊断测验不同属性层级对应的完备Q矩阵设计的例子和罗列了Wang等人（2013）的Q矩阵修改的方法.     

Are Past Normative Behaviors Predictive of Future Behavioral Intentions?

We acknowledge the limitations in measures of moral reasoning and pursue an alternative technique by investigating past behaviors as they relate to present behavioral intentions. Our purpose is to evaluate the merits of patterned normative behavior for predicting present and future, morally relevant outcomes. Participants (N = 177) completed a policy capturing experimental design responding to questions that orthogonally varied the situational nature of the decision context. Results indicate that past normative behaviors are significantly and directly related to ethical behavioral intentions. Moreover, they moderate the relationships between situational factors and intended outcomes as well as moral reasoning and intended outcomes.