A Comparison of Heuristic Procedures for Minimum Within-Cluster Sums of Squares Partitioning

Perhaps the most common criterion for partitioning a data set is the minimization of the within-cluster sums of squared deviation from cluster centroids. Although optimal solution procedures for within-cluster sums of squares (WCSS) partitioning are computationally feasible for small data sets, heuristic procedures are required for most practical applications in the behavioral sciences. We compared the performances of nine prominent heuristic procedures for WCSS partitioning across 324 simulated data sets representative of a broad spectrum of test conditions. Performance comparisons focused on both percentage deviation from the “best-found” WCSS values, as well as recovery of true cluster structure. A real-coded genetic algorithm and variable neighborhood search heuristic were the most effective methods; however, a straightforward two-stage heuristic algorithm, HK-means, also yielded exceptional performance. A follow-up experiment using 13 empirical data sets from the clustering literature generally supported the results of the experiment using simulated data. Our findings have important implications for behavioral science researchers, whose theoretical conclusions could be adversely affected by poor algorithmic performances.     

科学知识社会学相对主义知识观批判

“The Strong Programme” is put forward as a metaphysical theory of sociology by the Edinburgh School (SSK) to study the social causes of knowledge. Barry Barnes and David Bloor are the proponents of the School. They call their programme “the Relativist View of Knowledge” and argue against rationalism in the philosophy of science. Does their relativist account of knowledge present a serious challenge to rationalism, which has dominated 20th century philosophy of science? I attempt to answer this question by criticizing the main ideas of SSK and defending rationalism theories in modern philosophy of science.     

The Epistemic Criterion: A Response to Michael Hand

This article is a response to Michael Hand’s critique in this issue of my response to his use of the epistemic criterion as the sole means for identifying whether or not an issue should be identified as controversial. I argue that he has misunderstood my intention in suggesting that I was seeking to replace the epistemic criterion. Rather my purpose was to challenge his over-confidence in the decisiveness of reason and to temper his use of the epistemic criterion by suggesting it needed to embrace ethical concerns.     

Demand characteristics and research into drug use

Abstract

Health research makes frequent use of verbal reports; and such reports are usually assumed to be the surface indicants of measurable entities such as attitudes, beliefs, intentions, etc. which reside inside the person. It is our assertion that verbal reports themselves are motivated and variable; they are functional and context-dependent and their meaning and significance are localised. Consequently discourse cannot be treated as if it were fixed and categorical data to be retrieved from memory by asking questions that appear objective or disinterested. A method is postulated, using a social perception analogy with signal detection (SD) theory. The method allows for the assessment of subject criterion, and permits an examination of signal-strength in terms of the researcher's motivation. Data are evaluated in terms of their “robustness”, a measure which derives from the variability of response in different contexts, and under different elicitation procedures. The method requires few assumptions to be made about the “truth” or “falsity” of verbal reports. It focuses on the types of social activities that are performed by utterances and thus on the contextual variability revealed by asking questions in different situations and in different ways. This procedure does not presume a direct correspondence between verbal reports and mental representations.     

Development and Validation of the Xhosa Translations of the Beck Inventories: 3. Concurrent and Convergent Validity

This article provides data on the validity of the Xhosa versions of the Beck Depression Inventory-II (XBDI-II), the Beck Hopelessness Scale (XBHS) and the Beck Anxiety Inventory (XBAI) based on a sample of 122 Xhosa respondents which included students and patients. For patients, clinicians completed rating scales of Depression and Anxiety symptoms. In tests of concurrent validity, depressed patients had significantly higher scores on the XBDI-II and XBHS than students or patients who were not depressed. Similarly anxious patients had higher scores on the XBAI than students and patients who were not anxious. Correlations with clinicians' ratings were:.91 for XBDI-II scores and depression ratings, and.88 for XBAI scores and anxiety ratings. Correlations between the three translated scales were similar to those for the original scales. These analyses provide evidence that the translated scales have levels of concurrent and convergent validity comparable to the originals.     

Interactional Styles of Bullies and Victims Observed in a Competitive and a Cooperative Setting

Observational data were used to examine the interactional styles of bullies and victims while they participated in 2 different games: 1 cooperative and 1 competitive. Participants included 17 bullies, 18 victims, and 35 controls (all aged 8–11 years) selected through a peer nomination questionnaire. Bullies and victims were observed 4 times in total: twice during the cooperative game and twice during the competitive game, each time paired with a partner of opposite status or control status. Results indicated that when bullies and victims interacted together, bullies showed a dominant style in the dyad, often regulating and opposing victims' initiatives. Victims, on the other hand, complied with bullies' requests and presented a submissive style of interaction. However, interactions involving control partners indicated that (a) bullies opposed those partners' initiatives less frequently and (b) victims tended to affirm themselves by asking for help and explanations. These data provide evidence for an interactional model in explaining the dynamics between bullies and victims.     

Quality Improvement as a Tool for Translating Evidence Based Interventions Into Practice: What the Youth Violence Prevention Community can Learn from Healthcare

Health care has been working for the past 2 decades to improve the translation of evidence based practice (EBPs) into care. The strategies used to facilitate this, and lessons learned, can provide useful models for similar work taking place in youth violence prevention. This article discusses the history of evidence translation in health care, reviews key strategies used to support translation of evidence based practice into care, and suggests lessons learned that may be useful to similar efforts in youth violence prevention and intervention services.     

Effects of mastery criterion on the emergence of derived equivalence relations

In this study, we manipulated mastery criterion form (rolling or block) and stringency (across 6 or 12 trials) and measured the emergence of derived relations. College students learned neuroanatomy equivalence classes and experienced one of two rolling mastery criteria (6 or 12 consecutive correct responses) or a block mastery criterion (12 trials in a block) during training. The study found that block and rolling mastery criteria produced similar outcomes. Effectiveness was hampered when the criterion was less stringent.     

中外两种推理理论的实验比较研究

通过两个实验对中国学者提出的“推理题与推理者的推理知识双重结构模型”和Evans提出的“双重加工理论”进行了实验比较研究。实验一通过两种评定方法对相应性质命题进行评定后所得实验结果表明Evans等(1983)有关“信念效应”研究中的“结论可信性”变量可以视为与胡竹菁等(1996)实验中的“内容正确性”变量是同一性质的变量; 实验二根据“形式正确性”和“内容正确性”两个自变量设计的推理实验结果与Evans等(1983)的研究结果基本一致; 但增加“内容熟悉性”这一自变量设计的推理实验结果表明“推理题与推理者的推理知识双重结构模型”比Evans提出的“双重加工理论”能更好地解释推理者对性质三段论的推理结果。     

Decision framing in judgment aggregation

Judgment aggregation problems are language dependent in that they may be framed in different yet equivalent ways. We formalize this dependence via the notion of translation invariance, adopted from the philosophy of science, and we argue for the normative desirability of translation invariance. We characterize the class of translation invariant aggregation functions in the canonical judgment aggregation model, which requires collective judgments to be complete. Since there are reasonable translation invariant aggregation functions, our result can be viewed as a possibility theorem. At the same time, we show that translation invariance does have certain normatively undesirable consequences (e.g. failure of anonymity). We present a way of circumventing them by moving to a more general model of judgment aggregation, one that allows for incomplete collective judgments.