多维项目反应理论补偿性模型参数估计:基于广义回归神经网络集合

运用广义回归神经网络(GRNN)方法对小样本多维项目反应理论(MIRT)补偿性模型的项目参数进行估计,尝试解决传统参数估计方法样本数量要求较大的问题。MIRT双参数Logistic补偿模型被设置为二级计分的二维模型。首先,模拟二维能力参数、项目参数值与考生作答矩阵。其次,把通过主成分分析得到的前两个因子在每个题目上的载荷作为区分度的初始值以及题目通过率作为难度的初始值,这两个指标的初始值作为神经网络的输入。集成100个神经网络,其输出值的均值作为MIRT的项目参数估计值。最后,设置2×2种(能力相关水平:0.3和0.7; 两种估计方法:GRNN和MCMC方法)实验处理,对GRNN和MCMC估计方法的返真性进行比较。结果表明,小样本的情况下,基于GRNN集成方法的参数估计结果优于MCMC方法。     

基于速度与准确率权衡的心理测量学模型及应用

在大规模被试评价中,任务完成的准确率一直是评价的主要指标。然而,在各种任务情景中,反映执行者素质的指标除了结果的正确性之外,决策过程的时效性同样是极其重要的。因此,开发一个适合大规模评价情景,同时考虑速度与准确率权衡的模型,探索被试作答准确率与速度间的关系将具有重要的价值。基于认知心理学实验中的速度与准确率研究,构建一个基于速度与准确率权衡的心理测量学模型。新模型参数能非常稳定而又精确地被估计,同时模型中的变量及其关系也能够很好地得到实测数据的支持。     

基于多层验证性因素分析的各种信度系数方法

近年来,心理学研究的复现性受到广泛关注,许多实证研究难以重复验证,信度较低。大量研究使用多层技术,但只报告整体信度,导致研究可重复危机。基于各种信度系数和验证性因素分析的理论,以二层模型为例,总结多层信度计算方法,通过文献综述检索应用情况,并用MPLUS进行实例演示,最后讨论单层信度估计存在的偏差及分层估计的好处。总之,对多层数据进行分层信度估计是很有必要的,可消除因测量工具缺乏信度而导致的研究不可重复。     

The use of the political categories of Brexiter and Remainer in online comments about the EU referendum

In June 2016, the United Kingdom held a referendum on EU membership; 52% of those who voted, voted to leave, and 48% voted to remain. During the referendum campaign, two identities emerged: “Brexiter” and “Remainer,” which remained salient post‐referendum. This study explores how the categories of Brexiter and Remainer were deployed by posters online. Data comprise comment threads collected from four online newspapers both during the campaign and after the vote, which focus on the Brexit campaign promise: “We send £350m a week to the EU. Let's fund our NHS instead.” We draw on membership categorization analysis and discursive psychology to analyse when categories were made salient and what responses to the invocation of categories were. Analysis revealed that posters explicitly categorize the out‐group and in doing so implicitly define their group. Posters resisted other political identities when attributed to them in relation to the referendum. The analysis shows how Brexiter and Remainer are new, albeit contested, political categories and identities in their own right, with other political identities resisted when used. The paper highlights implications for the political system in the United Kingdom and for social divisions within U.K. society.     

百善孝为先?大学生双元孝道信念与网络成瘾的关系研究

本研究基于双元孝道模型和病理性互联网使用的认知行为模型,考察双元孝道信念对网络成瘾的影响,并检验非适应性认知的中介作用。以1152名大学生为被试,采用双元孝道信念量表、非适应性认知量表和网络成瘾量表进行测试,结果发现：（1）互惠性孝道和非适应性认知显著负相关,和网络成瘾显著负相关;权威性孝道和非适应性认知显著正相关,和网络成瘾显著正相关;非适应性认知和网络成瘾显著正相关;（2）非适应性认知在孝道信念与网络成瘾之间起中介作用。具体而言,互惠性孝道负向影响非适应性认知,权威性孝道正向影响非适应性认知,非适应性认知正向影响网络成瘾。本研究的结果表明,根植于中国传统文化的孝道信念通过非适应性认知的中介作用影响网络成瘾;而且互惠性孝道和权威性孝道具有相反的作用。     

I Can't Get You Out of My Head: The Influence of Secrecy on Consumers’ Self‐Brand Connections

Although research provides anecdotal evidence of consumers keeping their brand consumption a secret, there is little empirical corroboration to justify its study or illuminate the resulting consequences. The goals of this research are to provide evidence for the prevalence of keeping brand consumption a secret and to understand the resulting cognitive processing and consequences. Specifically, we first explore the occurrence of secrecy in the context of brand consumption, its types, and its motivations, underscoring the need for further investigation. Then, through a series of three subsequent studies, we demonstrate that keeping brand consumption a secret can lead to enhanced self‐brand connections through the underlying processes of thought suppression and thought intrusion. We rule out alternative explanations and enhance generalizability through the examination of passive (e.g., avoiding) versus active (e.g., lying) secrecy and variations in agency (i.e., instructed vs. voluntary secrecy).     

Meta‐Analysis of Integrated Primary and Behavioral Health Care Interventions for Treating Substance Use Among Adults

This meta‐analysis detected low effects in reducing substance use (Hedges's g = –0.11) in favor of using integrated primary and behavioral health care (IPBH) compared with treatment‐as‐usual approaches for adult patients with substance use. The impact of IPBH was limited.     

Education for all: The Good Inclusion Game

The term “education for all” in the context of inclusive schooling describes the aim that children with identified special/additional needs are fully included in education together with their typically developing peers. However, this is easier said than done as there are few easy‐to‐use methods that teachers can use while at the same time teaching the approbate curriculum to children with a full range of different abilities. The Good Inclusion Game (GIG) is a group contingency‐based tool to create inclusive classrooms that utilizes principles of the applied branch of the science of behavior analysis (applied behavior analysis) and can be used across settings and academic subjects. The GIG was evaluated across nine classrooms including 93 boys and girls aged between 9 and 15 years of age, including 20 children with identified special educational needs. Findings show that the GIG reliably led to a significant increase of inclusive curriculum‐focused activities with the collateral effect of decreasing disruptive behaviors for all children. Findings are discussed in the context of inclusive schooling and evidence‐based education.     

Support gaps in parent‐emerging adult dyads: The role of support quality

This study used a support gap approach to examine parental support to college‐going emerging adults. Based on data from 156 parent–child dyads, this study investigated whether discrepancies between the amounts of support that children received from parents and what they desired would be associated with both persons’ perceived support quality and individual and relational well‐being. Using polynomial regressions with response surface modeling, analyses demonstrated that, generally, support deficits (i.e., receiving less support than desired) were associated with poorer outcomes. However, support surpluses (i.e., receiving more support than desired) were not necessarily problematic, and children and parents exhibited different viewpoints regarding support surpluses. In addition, child‐reported support quality mediated several associations between support gaps and relationship satisfaction reported by both persons.     

How does discourse among like-minded individuals affect their thinking about a complex issue?

Is there reason to be concerned about what has been seen as an increasing trend for discourse on complex issues to be confined to an “echo chamber” of like-minded individuals? To investigate how thinking regarding an issue changes in form following concentrated discourse with like-minded peers, we undertook a qualitative examination of changes in the reasoning that 34 older adults used to justify their positions prior to and following such discourse. They showed a divergence of patterns. A minority abandoned mention of previously expressed ideas at odds with their own positions or otherwise narrowed the reasoning used to justify their positions. Consistent with findings from previous studies, only a minority showed increased certainty and/or extremity (polarisation) on quantitative scales. Countering this minority were a larger proportion whose thinking appeared to benefit from such engagement in several respects that we describe. A follow-up study compares results for a sample of young adults.