混合效应回归模型及其实证应用

海绵有理想 2022-05-18 发布于吉林

展开全文

混合效应回归模型及其实证应用

编语

近年来，混合效应回归模型（mixed effects regression modelling）开始越来越频繁地出现在应用语言学的研究中。混合效应回归模型是一种兼测固定效应（fixed effect）和随机效应（random effect）的回归模型。简单来说，固定效应就是模型中的自变量，随机效应则是其他一些随机特征，比如个体差异、群体差异等。

那么，为什么混合效应回归模型能受到一众研究者的“青睐”？与方差分析（ANOVA），协方差分析（ANCOVA）和传统的回归分析相比，混合效应回归模型究竟有哪些“独到之处”？在具体研究中，我们又将如何使用“混合效应回归模型”进行数据分析及结果报告？假如您也有同样的疑惑，不妨来看看本期推送。本期推送中，小编为大家准备了五篇与混合效应回归模型相关的文章：前两篇侧重介绍混合效应回归模型的原理，并辅以实例说明；后三篇为实证研究，三篇文章均运用了混合效应回归模型对数据进行统计分析。

小编还为大家整理了极简版的混合效应线性模型操作指南，欢迎加入本公众号微信交流群获取：helenIxj0501（陆老师）、hanchen3713（韩老师）（扫描文末二维码也可添加）。

·内容概览 / TABLE OF CONTENTS

1. 原理介绍 1

2. 原理介绍 2

3. 实证研究 1

4. 实证研究 2

5. 实证研究 3

原理介绍 1

使用R软件进行线性模型和线性混合模型分析

Linear models and linear mixed effects models in R: Tutorials 1 and 2

教程一：这两篇教程将向您分别介绍线性模型和线性混合模型，它们功能强大，应用灵活。本篇教程是第一篇，侧重概念理解，故省略了较为复杂的数学知识，旨在帮助大家理解两种模型的基本原理。之后，我们会花一些篇幅讲解两种模型的实际应用。本教程期待帮助您尽快掌握线性模型相关知识，并将其运用到自己的研究分析中。在此基础上，您可以根据自己的需求，更深入地阅读相关文献（提示：您大约要花一小时阅读教程一）。

Tutorial 1: Linear models and linear mixed models are an impressively powerful and flexible tool for understanding the world. This tutorial is the first of two tutorials that introduce you to these models. The tutorials are decidedly conceptual and omit a lot of the more involved mathematical stuff. The focus is on understanding what these models are doing... and then we’ll spend most of the time applying this understanding. The idea is to bootstrap your knowledge as quickly as possible so that you can start with your own analyses and then turn to more technical texts if needed. This tutorial will take you about 1 hour (possibly a bit more).

教程二：本教程是一个混合效应模型集训营，旨在让您快速上手分析自己的数据。不同于其他概念性教程，本教程将重点介绍为何和如何使用混合效应模型。很多缺乏统计学背景知识的读者看到此类教程或许会望而却步，不过不必担心，本教程会相对通俗易懂一些。本教程需要用到R软件，如果您还没有安装，请安装一下。此外，如果您没有阅读过教程一，我强烈建议您在阅读教程二之先前看一下（提示：您大约要花一小时阅读教程二）。

Tutorial 2: This tutorial serves as a quick boot camp to jump-start your own analyses with linear mixed effects models. This text is different from other introductions by being decidedly conceptual; I will focus on why you want to use mixed models and how you should use them. While many introductions to this topic can be very daunting to readers who lake the appropriate statistical background, this text is going to be a softer kind of introduction… so, don’t panic! The tutorial requires R – so if you haven’t installed it yet, go and get it! I also recommend reading tutorial 1 in this series before you go further. This tutorial will take you about 1 hour (possibly a bit more).

引用信息及教程获取地址：

Winter, B. (2013). Linear models and linear mixed effects models in R with linguistic applications. arXiv:1308.5499.

[http:///pdf/1308.5499.pdf]

作者介绍

Bodo Winter博士

Bodo Winter，加利福尼亚大学美熹德分校认知与信息科学博士，现任伯明翰大学认知语言学助理教授。研究方向主要包括手势体态、像似性、隐喻、数量认知、语料库语言学、数据可视化。

原理介绍 2

（广义线性）混合效应模型：以学习者语料库分析为例

(Generalized Linear) Mixed-Effects Modeling: A Learner Corpus Example

摘要：本文详细描述了如何通过混合效应模型对语料库数据进行分析，比较了母语为德语和西班牙语的英语学习者和英语母语者使用宾语补语和主语补语的不同。

“需要强调的是，我们并未宣称这是分析数据的唯一方法，但这种方法确实带来了令人满意的分析结果。”（Bates等，2018，第5页）

关键词：混合效应模型；广义线性模型；回归；补语标记that

Abstract: This methods showcase article provides a detailed overview of a mixed-effects modeling analysis of corpus data on the use of that in object and subject complementation by native speakers of English compared to its use by German and Spanish learners of English.

We emphasize that we do not claim that our illustrations are the only way to carry out these analyses, but the strategy outlined above has yielded satisfactory results.” (Bates et al., 2018, p. 5)

Key words: mixed-effects modeling; generalized linear modeling; regression; that-complementation

论文引用信息：

Gries, S. T. (2021), (Generalized linear) mixed-effects modeling: A learner corpus example. Language Learning, 71(3), 757-798.

原文获取地址：

https:///10.1111/lang.12448

作者介绍

Stefan Th. Gries教授

Stefan Th. Gries，加利福尼亚大学圣塔芭芭拉分校教授。研究方向主要包括语料库语言学、统计方法在语言学中的运用、认知语言学、构式语法、语料库语言学与法律。

实证研究 1

西班牙语作为二语学习中的动机、可理解性和口音：研究动机对语音发展的预测作用

Motivation, Comprehensibility, and Accentedness in L2 Spanish: Investigating Motivation as a Time-Varying Predictor of Pronunciation Development

摘要：本文探讨了语言学习动机与二语语音发展之间的关系。26名以英语为母语的西班牙语学习者在一年中（即学习的第二、第三和第四学期）完成了5次图片描述任务，并录音。同时，学习者每学期（共计3学期）完成一份调查二语动机自我系统的定量问卷以及一份调查语言学习信念的开放式问卷。18名西班牙语母语者从可理解性和口音两方面对录音进行了评分。虽然混合效应模型显示一年中学习者的动机略有下降，但针对个体趋势的定性分析却表明，学习者逐渐开始形成和评估自己的语言学习目标，并能将这些目标纳入到个人发展和职业发展的目标中。此外，混合效应模型显示，随着时间的推移，学习者发音的可理解性不断提高，口音问题不断改善。将动机因素加入模型后，学习者的努力程度与口音显著相关，这说明随着时间的推移，努力程度对发音的作用可能会越来越大。

关键词：发音；可理解性；口音；西班牙语；动机；跟踪研究

Abstract: This study examined relationships between language learning motivation and the longitudinal development of second language (L2) pronunciation. Twenty-six English-speaking learners of Spanish recorded a simplified picture description task 5 times over a year-long period spanning their 2nd, 3rd, and 4th semesters of Spanish language instruction. Learners also completed a quantitative motivation survey based on the L2 Motivational Self System and an open-ended questionnaire on their language learning beliefs once per semester, yielding 3 measurements. Eighteen native Spanish listeners rated the learners’ clips for comprehensibility and accentedness. Although mixed modeling of the motivation data revealed a slightly negative trajectory for motivational subcomponents, qualitative analyses of individual patterns indicated that learners were beginning to formulate and evaluate language learning goals that were set into a larger framework of personal and professional objectives. Mixed effects models of the pronunciation data demonstrated that both comprehensibility and accentedness improved over time. When the quantitative motivation measures were integrated into the modeling process as time-varying fixed effects, effort was significantly related to accentedness, which suggests that effort may have played an increasingly important role in shaping learners’ pronunciation over time.

Key words: pronunciation; comprehensibility; accentedness; Spanish; motivation; longitudinal research

论文引用信息：

Nagle, C. (2018), Motivation, Comprehensibility, and Accentedness in L2 Spanish: Investigating Motivation as a Time-Varying Predictor of Pronunciation Development. The Modern Language Journal, 102, 199-217.

原文获取地址：

https:///10.1111/modl.12461

作者介绍

Charles Nagle博士

Charles Nagle，乔治城大学西班牙语及应用语言学博士，现任爱荷华州立大学副教授。研究方向主要包括研究方法论、二语习得、二语语音研究。

实证研究 2

工作记忆在青少年二语写作中的作用

The role of working memory in young second language learners’ written performances

摘要：本文研究了工作记忆水平在青少年英语二语写作中的作用，探讨了任务类型和年级高低与二语写作成绩之间的关系，以及工作记忆水平的高低是否对上述关系有调节作用。本文的被试为94名六、七年级的青少年英语学习者，他们的母语为匈牙利语。他们完成了四种类型的写作任务（这些任务均来自青少年版托福考试）和工作记忆水平测试。研究结果显示，被试在电子邮件写作和先听后写两项任务中得分较高。无论工作记忆水平的高低，七年级被试在先听后写和电子邮件写作两项任务的得分上均高于六年级被试。与其他类型的写作任务相比，被试在非学术编辑任务中得分较低。除了学术编辑任务外，工作记忆水平对其他写作任务得分没有影响。工作记忆水平对七年级被试先听后写任务得分没有显著影响，但却能影响预期分数的变化。与工作记忆力水平较低的被试相比，六年级工作记忆水平较高的被试在各项写作任务中表现更为一致。

关键词：工作记忆；青少年学习者；写作评估；写作测试；个体差异

Abstract: This study investigated the role of working memory (WM) in the second language (L2) writing performance of young English language learners. It also examined how L2 writing achievement relates to task type and grade level and whether the effect of cognitive abilities varies across different task types and grade level. The participants were 94 young learners (Grades 6 and 7) in Hungary, who performed four writing task types as part of the TOEFL® Junior™ Comprehensive test-battery and completed cognitive tests that assessed their WM functions. Participants scored high on the email writing and integrated Listen-Write tasks. Irrespective of WM functions, on average learners in Grade 7 outperformed those in Grade 6 on the Listen-Write task and the Email task. Students gained lower scores on the non-academic version of an editing task than on most other types of tasks. WM functions had no significant relationship with L2 writing scores, except for the academic editing task. In Grade 7, the effect of WM was not significant on the integrated Listen-Write task, but it resulted in the change of expected score. Learners with high working memory in Grade 6 showed somewhat more consistent performance across tasks than did learners with low working memory.

Key words: working memory; young learners; writing assessment; testing writing; individual differences

论文引用信息：

Michel, M., Kormos, J., Brunfaut, T., & Ratajczak, M. (2019), The role of working memory in young second language learners’ written performances. Journal of Second Language Writing, 45, 31-45.

原文获取地址：

https:///10.1016/j.jslw.2019.03.002

作者介绍

Marije Michel博士

Marije Michel，格罗宁根大学副教授。研究方向主要包括任务教学法、二语习得、二语教学法、数字化二语教学、眼动研究等。

Judit Kormos教授

Judit Kormos，兰卡斯特大学教授。研究方向主要包括心理学视角下的二语习得、有特殊教育需要的语言学习者研究。

Tineke Brunfaut教授

Tineke Brunfaut，兰卡斯特大学语言学及英语语言学教授。研究方向主要包括语言测试、二语阅读、二语听力。

Michael Ratajczak

Michael Ratajczak，兰卡斯特大学博士生。研究方向主要包括阅读、二语习得、语言测试、行为科学。

实证研究 3

读前词汇教学对词汇习得的影响：基于母语和二语学习者的眼动研究

The Effect of Pre-reading Instruction on Vocabulary Learning: An Investigation of L1 and L2 Readers’ Eye Movements

摘要：本文研究了读前词汇教学对注意力分配和词汇学习效果的影响。我们将被试（92名英语母语者，88名英语学习者）随机分为四组。组一为接受读前词汇教学组，该组被试在阅读前学习6个新词，每一个新词都会在之后的阅读文本中重复出现8次。组二为阅读组，该组被试在阅读前不学习新词，仅阅读含有新词的文本。组三为阅读控制组，该组被试在阅读前同样不学习新词，仅阅读文本，但与组二不同的是，文本内的新词被替换为被试认识的词。组四为仅接受词汇教学组，该组被试仅学习6个新词，之后阅读一篇无关的文章。我们用眼动仪记录了阅读过程中被试对新词的关注程度。我们用三项测试（形式识别、意义回忆和意义识别）评估了被试对目标词汇的掌握程度。结果表明，进行读前词汇教学（加上文本阅读）能够促进被试习得新词，提高被试加工新词的效率。阅读时被试对新词关注时间的多少是预测意义回忆测试分数的重要指标。

关键词：词汇学习；阅读；读前词汇教学；眼动

Abstract: This study examined the effect of pre-reading vocabulary instruction on learners’ attention and vocabulary learning. We randomly assigned participants (L1 = 92; L2 = 88) to one of four conditions: pre-reading instruction, where participants’ received explicit instruction on six novel items and read a text with the items repeated eight times; reading-only, where participants simply read the same text with the novel items repeated eight times; reading-baseline, where participants read the same text with the repeated items replaced by known (control) words; and instruction-only, where participants received explicit instruction on the novel items and read an unrelated text. Eye-tracking was used to measure amount of attention to the vocabulary during reading. We assessed knowledge of the target vocabulary in three immediate posttests (form recognition, meaning recall, and meaning recognition). Results showed that pre-reading instruction (plus reading the text) led to both more vocabulary learning and a processing advantage. Cumulative reading times were a significant predictor of meaning recognition scores.

Key words: vocabulary learning; reading; pre-reading instruction; eye-tracking

论文引用信息：

Pellicer-Sánchez, A., Conklin, K., & Vilkaitė-Lozdienė, L. (2021), The effect of pre-reading instruction on vocabulary learning: An investigation of L1 and L2 readers’ eye movements. Language Learning, 71, 162–203.

原文获取地址：

https:///10.1111/lang.12430

作者介绍

Ana Pellicer-Sánchez博士