信号检测论的应用

信号检测论(Signal Detection Theory,简称SDT)是信息论的一个分支,其研究对象是信号传输系统中信号的接收部分。它最早用于通讯工程中,即借助于数学的形式描述“接收者”在噪声背景中信号的接收问题。1954年密西根大学的W.P.Tanner 和J.A.Swets 等人首先将SDT 应用于感知觉研究。目前,SDT 作为一种新的心理物理学方法正广泛地应用于心理学研究的各个领域。     

一种测量记忆监测相对准确性的新指标：da

记忆监测的相对准确性是元记忆研究中一个重要内容。研究者采用了各种方法来测量人们记忆监测的判断值对正确项目和错误项目的区分度,例如相关法、信号检测论(SDT)测量法。其中,gamma相关从1984年起一直被广泛用于记忆监测相对准确性的测量。基于SDT的新指标da弥补了gamma相关在实际应用中的不足,并引发了新的研究兴趣。文章对da的提出背景、计算方法和应用前景进行了详细的介绍,对da与其他常用方法在使用中的优势进行了比较。对da的具体应用进行了说明,同时指出了da的使用条件。     

运用蒙特卡罗方法对SDT和IRT的比较研究

本研究运用蒙特卡罗模拟方法。在80种不同的实验条件下产生被试的反应矩阵,再分别用信号检测理论(SDT)和项目反应理论(IRT)中的Rasch模型、三参数逻辑斯谛模型对被试的能力进行估计,其结果表明,用这三种不同方法所得到的能力估计值均与能力的真实值有较高的相关。     

认知诊断模型Q矩阵修正：完整信息矩阵的作用

Q矩阵是CDM的核心元素之一,反映了测验的内部结构和内容设计,通常由领域专家根据经验进行主观界定,因此需要对可能存在的错误进行修正。本研究提出了一种新的Q矩阵修正方法——基于完整经验交叉相乘信息矩阵的Wald-XPD方法。采用Monte Carlo模拟检验了新方法的表现,并与同类方法进行了比较。研究表明：新开发的Wald-XPD方法在Q矩阵恢复率、保留正确标定属性的比例以及修正错误标定属性的比例这3个主要指标上均有较好的表现,且整体上优于其他方法,尤其是在修正错误标定的属性方面。通过实证数据展示了Wald-XPD方法在Q矩阵修正中的良好表现。总之,本研究为Q矩阵修正提供了有效的方法。     

20世纪西方学者对英国解散修道院动因的解释

20世纪以来,随着西方史学从传统史学到新史学的转变,西方史学界对宗教改革以及解散修道院的研究,经历了从传统学派、修正学派到后修正学派的传承和变化。传统学派认为,中世纪晚期英格兰的修道院已经走向腐化,解散是自下而上、不可逆转的进步事件。20世纪80年代兴起的修正学派则为天主教辩护,把宗教改革看作是自上而下、迅速完成的政治行为,寻找修道院在英格兰原本可以继续存在的合理性。近十几年来,逐渐兴起的后修正学派开始挑战修正学派的观点,他们扩展了宗教改革史的研究范围,从民众对新教的适应过程这一角度质疑修正学派,试图平衡中世纪晚期修道院的颓势与活力,在解散修道院的研究上不断寻求突破。     

基于类别水平的多级计分认知诊断Q矩阵修正：相对拟合统计量视角

多级计分认知诊断模型的开发对认知诊断的发展具有重要作用, 但对于多级计分模型下的Q矩阵修正还有待研究。本研究尝试对多级计分认知诊断Q矩阵修正进行研究, 并聚焦更具诊断价值的基于项目类别水平的Q矩阵修正。将相对拟合统计量应用于多级计分认知诊断Q矩阵修正, 并与已有方法Stepwise方法( Ma & de la Torre, 2019)进行比较。研究表明：BIC方法对多级计分认知诊断模型的Q矩阵修正具有较高的模式判准率和属性判准率, 其对Q矩阵的恢复率也高于Stepwise方法, BIC方法修正后的Q矩阵与数据更加拟合; 在复杂模型中, 相对拟合指标BIC比AIC和-2LL表现更好, 在实践中, 使用者可以选择BIC法进行测验Q矩阵修正; Q矩阵修正效果受到被试人数的影响, 增加被试人数可以提高Q矩阵修正的正确率。总之, 本研究为多级计分认知诊断Q矩阵修正提供了重要的方法支持。     

基于DINA模型的Q矩阵修正方法

本研究开发了一种基于DINA模型的认知诊断测验Q矩阵修正方法—— g 法, 为侦查并修正Q矩阵中的错误提供方法学支持, 从而为保证Q矩阵的合理性提供基础, 并为进一步提高认知诊断的准确率服务。本文采用Monte Carlo模拟及与国外同类研究相比较的方法进行, 研究发现：(1)不论在何种作答失误概率(5%, 10%, 15%)情况下, 当s,g临界值为0.2, 0.25或0.3时, 本研究提出的g 法均能有效地修正错误Q矩阵; 同时, 当Q矩阵无错误时, g 法对该Q矩阵未做任何修改。表明g 法对Q矩阵是否存在错误具有较强的识别能力及修正能力。(2)与国外同类研究相比, 本研究提出的g 法具有较理想的修正率, 且与de la Torre (2008)提出d法的修正效果相当。但相比较而言, g 法较d 法更为简单。(3) g 法不仅能有效地修正错误的Q矩阵, 而且还可以进一步提高认知诊断的正确率, 尤其是对模式判准率(PMR), 诊断正确率的最高增幅高达40%, 大大改善了认知诊断的准确率。     

基本心理需要及其满足

Deci和Ryan在前人基础上提出基本心理需要的概念,并将人的基本心理需要划分为胜任需要、关系需要和自主需要。基本心理需要来源于自我决定理论(SDT)。目前,对于基本心理需要的测量主要是采用Gagné编写的一般需要满足量表(GNSS)。基本心理需要的满足主要受到个体所处环境的影响,且与个体的心理健康有着密切的联系,基本心理需要受挫则会导致个体的焦虑、压抑。对于基本心理需要的质疑主要集中在需要的内容、各需要之间是否矛盾、需要是否会改变以及需要的普适性四个方面。未来有必要加强对基本心理需要的测量及需要受挫的研究,同时加强纵向研究,以及对于基本心理需要满足策略的研究。     

认知诊断Q矩阵估计(修正)方法

Q矩阵代表着项目考察的属性, 反映了项目的重要特征, 其正确性是影响认知诊断分类准确性的关键因素。研究Q矩阵估计(修正)方法具有重要价值。首先, 研究从是否采用认知诊断模型将Q矩阵估计(修正)分为基于认知诊断模型视角下的参数化方法和基于统计视角下的非参数方法。然后, 分别从最优项目质量、最优模型数据拟合和参数估计视角对它们进行分类介绍, 评析不同方法的特征和表现、区别与联系、优势与不足。最后, 提出几个未来研究问题：在复杂测验条件下系统比较各种方法; 校准知识状态和参数估计误差、结合多种思路和方法等多角度提出Q矩阵估计(修正)方法; 研究多级评分项目、混合测验模型、属性多级、属性个数未知甚至Q矩阵元素为连续变量等条件下的Q矩阵估计(修正)方法。     

邻域语义与修正真理论

古普塔和赫兹伯格在1982年各自独立地提出了修正真理论,建立了可用于分析真与相关悖论的修正序列。修正真理论根据语句在所有修正序列中的表现,对语句进行分类。然而,修正真理论在某些语句的分类上不能令人满意,如修正真理论把柯瑞悖论的逆命题断定为绝对地真,这与直觉不一致。本文将从两种路径引入邻域语义研究修正真理论。路径一是在基模型上引入邻域基模型,建立邻域基模型修正序列。这类修正序列比经典修正序列更多,增加的修正序列可使包括柯瑞悖论的逆命题在内的一些语句的病态呈现出来。路径二是通过引入邻域语义模型,使得对任意不含模态词的公式φ,模态公式□φ在后继阶段的真值可以反映φ在上一阶段的真值,并且□φ在极限阶段的真值可以反映φ在至这个极限阶前是否稳定真。从而可以通过□φ的真值来限定Tφ的真值,使得满足相应限制的模型类表示了相应的修正序列。本文最后将对两个路径进行整合,构造出能表示邻域基模型修正序列的整体修正序列模型。     

Atkinson's “risk preference” model

Several revisions of Atkinson's risk preference model have been proposed in the last 10 years. The advantage of such revisions, shifting the maximum of the approach tendency to the difficult range in predicting choice behavior, is discussed. After summarizing the results of a series of studies on choice behavior in achievement-related situations, the conclusion is drawn that such a revision is unnecessary. Choice behavior is in agreement with the prediction of the original model when expectancies are assessed in the proper way. There may be other good reasons to revise Atkinson's original model. However, the tendency to choose difficulty levels with objective success probabilities less than .50 is not a reason.     

Aspects of psychoanalytic theory: drives, defense, and the pleasure-unpleasure principle

Freud explained certain fundamentally important aspects of mental motivation by assuming the existence of two drives, one libidinal and the other aggressive/destructive. Elements of this theory that seem invalid are identified and discussed, and revisions are proposed that appear to have more validity and greater clinical usefulness.     

FUNCTIONAL ASSESSMENT,CURRICULAR REVISION,AND SEVERE BEHAVIOR PROBLEMS

An adolescent female with multiple handicaps and a long history of severely disruptive behavior participated in a functional assessment linked directly to specific revisions in her school curriculum. During Phase 1, reversal designs were used to test hypotheses pertaining to antecedent and auricular influences on problem behavior. During Phase 2, a multiple baseline across afternoon and morning time periods demonstrated that the auricular revisions were effective in eliminating severely disruptive behavior and increasing on-task responding. Data also showed that inappropriate “psychotic” speech was reduced and appropriate social interactions were increased. Follow-up results showed that the changes were maintained throughout the school year. Questionnaire data provided social validation of the procedures and outcomes. The findings are discussed in relation to their implications for functional assessment, individualized curricula, and positive programming for students with disabilities and serious behavior problems.     

Are Shunt Revisions Associated with IQ in Congenital Hydrocephalus? A Meta -Analysis

Although it is generally acknowledged that shunt revisions are associated with reductions in cognitive functions in individuals with congenital hydrocephalus, the literature yields mixed results and is inconclusive. The current study used meta-analytic methods to empirically synthesize studies addressing the association of shunt revisions and IQ in individuals with congenital hydrocephalus. Six studies and three in-house datasets yielded 11 independent samples for meta-analysis. Groups representing lower and higher numbers of shunt revisions were coded to generate effect sizes for differences in IQ scores. Mean effect size across studies was statistically significant, but small (Hedges’ g = 0.25, p < 0.001, 95 % CI [0.08, 0.43]) with more shunt revisions associated with lower IQ scores. Results show an association of lower IQ and more shunt revisions of about 3 IQ points, a small effect, but within the error of measurement associated with IQ tests. Although clinical significance of this effect is not clear, results suggest that repeated shunt revisions because of shunt failure is associated with a reduction in cognitive functions.     

Slot Mereology Revised

The paper suggests two revisions of K. Bennett's system of slot mereology. The revisions do not touch on the philosophical rationale for this system, but are focused on certain logical deficiencies in her formalisation.     

An adolescent version of the Michigan Alcoholism Screening Test

Item content of the Michigan Alcoholism Screening Test (MAST) was modified to make it more appropriate for young persons. The resulting test was found to have lower internal consistency than the adult MAST, but the elimination of five items with comparatively poor psychometric properties yielded an acceptable alpha coefficient of .73. A unitary factor model for the adolescent MAST was not confirmed; indeed, the revised test appeared to be factorially complex. Recommendations for further revisions are discussed.     

Using computers to disseminate scientific information: The potential of optical disk based journals

Several potential advantages of computerized scientific journals delivered on erasable optical disks are explored, including the availability of hypertext functions; databases of evaluative peer ratings and comments, as well as of author responses; keyword, rating, and text searching; author revisions or retractions of articles; availability of original research data to other researchers; and the possibility of incremental construction of journal-based computer expert systems. Problems associated with current editorial review processes that are likely-to-be obviated or alleviated by computerized journals are also discussed.     

Teaching Children About Revision in Writing

We investigated effects of instruction in the process of revision on children's knowledge of the revision process, on their ability to make revisions on paper, and on the quality of their writing. Fifteen sixth graders received instruction in the process of revision; 15 controls read good literature. After instruction, each child wrote a story, was interviewed about potential revisions, and was given an opportunity to make changes on the first copy and then write another draft. The main variables for knowledge of the revision process were number of spots suggested for revision and average specificity of suggested changes. The main variable for ability to make revisions on paper was total number of revisions made. A quality score was obtained for the first and final drafts. Instruction did affect knowledge of the revision process, and it enhanced revision efforts. Quality judgments tended to increase across drafts for the revision group, whereas they remained stable for the controls.     

The dynamics of belief: Contractions and revisions of probability functions

Conclusion Using probability functions defined over a simple language as models of states of belief, my goal in this article has been to analyse contractions and revisions of beliefs. My first strategy was to formulate postulates for these processes. Close parallels between the postulates for contractions and the postulates for revisions have been established - the results in Section 5 show that contractions and revisions are interchangeable.As a second strategy, some suggestions for more or less explicit constructive definitions of the revision process (and indirectly also of the contraction process) were then presented. However, the results in Section 6 are less conclusive than in the earlier ones. This problem area still awaits further development.