Digital Module 16: Longitudinal Data Analysis

In this digital ITEMS module, Dr. Jeffrey Harring and Ms. Tessa Johnson introduce the linear mixed effects (LME) model as a flexible general framework for simultaneously modeling continuous repeated measures data with a scientifically defensible function that adequately summarizes both individual change as well as the average response. The module begins with a nontechnical overview of longitudinal data analyses drawing distinctions with cross-sectional analyses in terms of research questions to be addressed. Nuances of longitudinal designs, timing of measurements, and the real possibility of missing data are then discussed. The three interconnected components of the LME model—(1) a model for individual and mean response profiles, (2) a model to characterize the covariation among the time-specific residuals, and (3) a set of models that summarize the extent that individual coefficients vary—are discussed in the context of the set of activities comprising an analysis. Finally, they demonstrate how to estimate the linear mixed effects model within an open-source environment (R). The digital module contains sample R code, diagnostic quiz questions, hands-on activities in R, curated resources, and a glossary.     

想象训练研究现状的分析

本文采用文献资料法对近十年刊载在中美四种主要期刊上的想象训练研究报告作了分析。结果表明想象训练正在从有效性研究向过程变量研究深化 ,研究面也由简单技术向复杂技术扩展 ,指出了目前复杂技术研究明显不足 ,还指出了基础理论在过程变量控制方面的研究严重滞后 ,在一定程度上影响了想象训练研究的进一步深化。本文为想象训练的研究和应用提供了有益的综合信息。     

College course scarcity and time to degree

College students are taking longer to earn baccalaureate degrees now than ever before, but little is known about institutional factors that may contribute to this trend. In this paper we investigate an important institutional constraint—course scarcity—that we hypothesize may be associated with increased time to degree. We employ a unique administrative dataset from a large, moderately selective, public institution and use an instrumental variables approach, identifying off the random registration times assigned to students. Results suggest that course scarcity does not delay students’ graduation. We explore alternative explanations for our findings and discuss a variety of other factors correlated with time to baccalaureate completion.     

中文电子病历的分词及实体识别研究

[目的/意义]健康医疗大数据是我国重要的基础性战略资源，本研究对中文电子病历分词与实体识别的探讨与实证较好地完成了医疗数据的信息抽取任务，对今后医疗大数据在语义层面的应用发展具有重要意义。[方法/过程]本研究首先融合权威词表、官方标准、健康网站数据及其他医学补充词库构建了词语数量级达到10万的医学词表；然后对电子病历的字段进行分词，对比了jieba工具、导入词典后的jieba、无监督学习及AC自动机4种模型的分词效果；最后，以自动分词和人工标注结果为语料，实现基于条件随机场的电子病历实体识别研究，并比较不同实体类别以及不同文本特征下的实体识别效果，选出最优模板。[结果/结论]分词结果显示，AC自动机的效果最好，F值可达82%；实体识别结果表明，"检查"和"疾病"实体的识别效果最好，而"症状"的识别效果不太理想。     

Does working during higher education affect students’ academic progression?

This paper examines the effect of working during higher education on academic progression, in terms of number of credits acquired by first-year university students in Italy. We discuss different contrasting hypotheses on the role of employment during university on academic outcomes: the zero-sum perspective, the reconciliation thesis, the positive and the negative selection to work hypotheses. In the empirical part we analyze data from the Eurostudent survey, which collected data on a representative sample of university students who were enrolled in the academic year 2002/03, after the implementation of the ‘Bologna Process’. We use a negative binomial regression model considering work experience as an endogenous multinomial treatment. Results indicate that, conditional on observed covariates (socio-demographic variables, school-related and university-related variables), there is a positive self-selection into employment, especially for low-intensity work. Traditional multivariate regressions show a penalty in academic progression only for high-intensity workers, but once accounted for unobserved heterogeneity also the low-intensity work experience appears to negatively affect academic progression.     

A multi-criteria approach for the group assessment of an academic course: A case study

The present paper aims at proposing a structured methodology to evaluate the results of an academic training course addressed to postgraduate students. The evaluation process intends to identify the efficacy of an education course by means of a comparison between the general objectives and the results expected by the students. Since the evaluated aspects by the students, detected by means of questionnaires, are both quantitative and qualitative, the proposed methodology is based on the analytic hierarchy process (AHP). In order to facilitate the evaluation process, answers have been modelled by means of linguistic variables. In particular, different evaluation scales have been used and each one is characterized by different levels associated to different fuzzy numbers.     

An NCME Instructional Module on Data Mining Methods for Classification and Regression

Data mining methods for classification and regression are becoming increasingly popular in various scientific fields. However, these methods have not been explored much in educational measurement. This module first provides a review, which should be accessible to a wide audience in education measurement, of some of these methods. The module then demonstrates using three real‐data examples that these methods may lead to an improvement over traditionally used methods such as linear and logistic regression in educational measurement.     

学术文献引文上下文自动识别研究

[目的/意义] 引文内容分析能够帮助揭示文献引用关系的深层语义内涵,而引文上下文识别作为引文内容分析的基础显得尤为重要。[方法/过程] 梳理已有引文上下文研究的现状,总结当前引文上下文识别的不足,在此基础上归纳引文上下文识别的5类特征,并采用文本分类和序列标注两种方法开展引文上下文自动识别实验。[结果/结论] 实验结果表明,本文提出的特征能够很好地提升引文上下文识别效果,且基于文本分类的SVM分类效果要优于基于序列标注的CRF。     

最优购书量的讨论

运用概率论和微分学的知识,讨论在一定的广告费用和书本贮存费用的情况下,根据图书的购进价和售出价确定书店的最优购书量．     

一类随机向量函数序列的依概率收敛性

将依概率收敛的一维随机变量序列的连续函数仍依概率收敛的结论推广到随机向量序列的情形。