认知诊断中基于条件期望的距离判别方法

分类是认知诊断评估的一个核心问题。基于观察反应模式与理想反应模式之间的距离的判别方法, 以确定性的理想反应模式为类中心, 而这没有考虑误差, 故未充分利用总体分布信息。为了更充分地利用总体分布信息、提高诊断分类效果和拓展诊断评估的适用性, 本研究提出给定知识状态条件下项目反应模式的条件期望向量为类中心的欧氏距离判别方法, 同时提出认知诊断模型下项目反应函数估计方法以获得这个条件期望向量。模拟研究表明：认知诊断模型下的项目反应函数估计方法得到的条件期望向量返真性较高, 获得的分布信息较准确; 在观察反应模式与理想反应模式差异大的情形下, 基于条件期望向量为类中心的欧氏距离判别方法优于基于理想反应模式为类中心的分类方法(广义距离方法和非参数方法）。研究可为认知诊断分类和等值方法提供一个参考。     

The Mid-Career Award

Working memory is an extremely influential concept within experimental psychology, with, at the time of writing, over 90 papers with this term in their title published in this journal alone since 2000. One reason for this interest is that measures of working memory tend to be strong correlates of important indices of real-world function. In addition, at first sight working memory appears to be a relatively simple concept to understand. However, despite this apparent simplicity, explaining working memory performance is not straightforward. In this paper I address this challenge, with a particular focus on the development of working memory performance in children; both children developing typically and those experiencing atypical development. I specifically highlight the multiple constraints on working memory performance, and how these constraints inter-relate. I then consider the broader theoretical implications of each of these constraints for current accounts of working memory and its development.     

Relations between children’s metamemory and strategic performance: Time-varying covariates in early elementary school

Although much is known about the development of memory strategies and metamemory during childhood, evidence for linkages between these memory skills, either concurrently or over time, has been limited. Drawing from a longitudinal investigation of the development of memory, repeated assessments of children’s (N = 107) strategy use and declarative metamemory were made to examine the development of these skills and the relations between them over time. Latent curve models were used first to estimate the trajectories of children’s strategy use and metamemory and then to examine predictors of children’s performance in each of these domains. Children’s metamemory at the beginning of Grade 1 was linked to child- and home-level factors, whereas the development of both skills was related to maternal education level. Additional modeling of the longitudinal relations between strategic sorting and metacognitive knowledge indicated that metamemory at earlier time points was predictive of subsequent strategy use.     

The form of the forgetting curve and the fate of memories

Psychologists have debated the form of the forgetting curve for over a century. We focus on resolving three problems that have blocked a clear answer on this issue. First, we analyzed data from a longitudinal experiment measuring cued recall and stem completion from 1 min to 28 days after study, with more observations per interval per participant than in previous studies. Second, we analyzed the data using hierarchical models, avoiding distortions due to averaging over participants. Third, we implemented the models in a Bayesian framework, enabling our analysis to account for the ability of candidate forgetting functions to imitate each other. An exponential function provided the best fit to individual participant data collected under both explicit and implicit retrieval instructions, but Bayesian model selection favored a power function. All analysis supported above chance asymptotic retention, suggesting that, despite quite brief study, storage of some memories was effectively permanent.     

A multivariate reduced-rank growth curve model with unbalanced data

A multivariate reduced-rank growth curve model is proposed that extends the univariate reducedrank growth curve model to the multivariate case, in which several response variables are measured over multiple time points. The proposed model allows us to investigate the relationships among a number of response variables in a more parsimonious way than the traditional growth curve model. In addition, the method is more flexible than the traditional growth curve model. For example, response variables do not have to be measured at the same time points, nor the same number of time points. It is also possible to apply various kinds of basis function matrices with different ranks across response variables. It is not necessary to specify an entire set of basis functions in advance. Examples are given for illustration.The work reported in this paper was supported by Grant A6394 from the Natural Sciences and Engineering Research Council of Canada to the second author. We thank Jennifer Stephan for her helpful comments on an earlier version of this paper. We also thank Patrick Curran and Terry Duncan for kindly letting us use the NLSY and substance use data, respectively. The substance use data were provided by Grant DA09548 from the National Institute on Drug Abuse.     

Area under the curve as a measure of discounting

We describe a novel approach to the measurement of discounting based on calculating the area under the empirical discounting function. This approach avoids some of the problems associated with measures based on estimates of the parameters of theoretical discounting functions. The area measure may be easily calculated for both individual and group data collected using any of a variety of current delay and probability discounting procedures. The present approach is not intended as a substitute for theoretical discounting models. It is useful, however, to have a simple, univariate measure of discounting that is not tied to any specific theoretical framework.     

Screens across the pond: Findings from longitudinal screen time research in the US and UK

Digital media are increasingly pervasive in the lives of young children. This increase in the availability of digital media might have long-run implications for child development; however, it is too soon to definitively conclude the direction of effects. In part due to this lack of certainty, leading health organizations have chosen to make different recommendations to parents of young children: Many international health organizations (e.g., the American Academy of Pediatrics, World Health Organization) recommend very young children be limited to under one hour of screen time daily, whereas others (e.g., Royal College of Paediatrics and Child Health) have intentionally opted not to make recommendations about specific limits. These guidelines might contribute to parents in different countries making meaningfully different choices about children’s use of digital media. Using a sample of N = 303 families recruited in Cambridgeshire, England and New York City prior to the birth of couples’ first child, we explore predictors of digital media use across the first two years of life. Data were collected when children were 4, 14, and 24 months of age. Results of latent growth curve analyses show that generally, children spend more time engaging with digital media as they grow older; however, growth mixture models reveal most children fit into one of two classes: One group of children (High Media Users; 52.2 %) engages with a substantial amount of digital media, whereas the other (Low Media Users; 48.8 %) engages with relatively little. Children in the US were approximately 30 % more likely to be in the Low Media Users group and there were no differences in group membership on the basis of parents’ psychosocial wellbeing. While these differences could be due to a number of factors, these findings may reflect the power of pediatric recommendations.     

Robust Bayesian growth curve modelling using conditional medians

Growth curve models have been widely used to analyse longitudinal data in social and behavioural sciences. Although growth curve models with normality assumptions are relatively easy to estimate, practical data are rarely normal. Failing to account for non-normal data may lead to unreliable model estimation and misleading statistical inference. In this work, we propose a robust approach for growth curve modelling using conditional medians that are less sensitive to outlying observations. Bayesian methods are applied for model estimation and inference. Based on the existing work on Bayesian quantile regression using asymmetric Laplace distributions, we use asymmetric Laplace distributions to convert the problem of estimating a median growth curve model into a problem of obtaining the maximum likelihood estimator for a transformed model. Monte Carlo simulation studies have been conducted to evaluate the numerical performance of the proposed approach with data containing outliers or leverage observations. The results show that the proposed approach yields more accurate and efficient parameter estimates than traditional growth curve modelling. We illustrate the application of our robust approach using conditional medians based on a real data set from the Virginia Cognitive Aging Project.     

多变量追踪研究的模型整合与拓展：考察往复式影响与增长趋势

追踪研究当中, 交叉滞后模型可以探究多变量之间往复式影响, 潜增长模型可以探究个体增长趋势。对两类模型进行整合, 例如同时关注往复式影响与个体增长趋势, 同时可以定义测量误差、随机截距等变异成分, 衍生出随机截距交叉滞后模型、特质-状态-误差模型、自回归潜增长模型、结构化残差潜增长模型等。以交叉滞后模型和潜增长模型分别作为基础模型, 从个体间/个体内变异分解的角度对上述各类模型梳理, 整合出此类模型的分析框架, 并拓展建立“因子结构化潜增长模型(factor latent curve model with structured reciprocals)”作为统合框架。通过实证研究(早期儿童的追踪研究-幼儿园版, ECLS-K), 建立21049名儿童的阅读和数学能力的往复式影响与增长趋势。研究发现, 分离了稳定特质的模型拟合最优。研究也对模型建模思路和模型选择提供了建议。     

Mentally walking through doorways causes forgetting: The location updating effect and imagination

Researchers have documented an intriguing phenomenon whereby simply walking through a doorway causes forgetting (the location updating effect). The Event Horizon Model is the most commonly cited theory to explain these data. Importantly, this model explains the effect without invoking the importance or reliance upon perceptual information (i.e., seeing oneself pass through the doorway). This generates the intriguing hypothesis that the effect may be demonstrated in participants who simply imagine walking through a doorway. Across two experiments, we explicitly test this hypothesis. Participants familiarised themselves with both real (Experiment 1) and virtual (Experiment 2) environments which served as the setting for their mental walk. They were then provided with an image to remember and were instructed to imagine themselves walking through the previously presented space. In both experiments, when the mental walk required participants to pass through a doorway, more forgetting occurred, consistent with the predictions laid out in the Event Horizon Model.