HOW TO PROCESS THE EFFECT OF MEDIATION AND

HOW TO PROCESS THE EFFECT OF MEDIATION AND MODERATION IN “PROCESS” 張偉豪 三星統計服務有限公司 執行長 Amos 亞洲一哥

Book � � Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regressionbased approach: Guilford Press. Hayes, A. F. (2012). Process: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling.

Outline � Questions of Whether, If, How and When � 自變數、應變數、中介變數、調節變數及控制變數 � � PROCESS安裝與介紹 Mediation Analysis � Simple mediation model � Parallel Multiple Mediator model � Serial Multiple Mediator model � Reporting a Mediation Analysis

Outline � Moderation Analysis � Fundamentals of Moderation Analysis � Moderation Involving a Dichotomous moderator � Interaction between tow Quantitative Variable � Plot an interaction � Truth or Myth about Mean Centering � Reporting a Moderation Analysis � Advanced model analysis � Moderated Mediation � Mediated Moderation

中介效果分析 (How or Why) UF EOU ATT BI

干擾效果分析 (When or What) ATT EOU UF BI

Introduction of PROCESS 2. 13 � 分析 迴歸 PROCESS

Introduction of PROCESS 2. 13 � PROCESS OPTIONS

Mediation Analysis � � Simple mediation model Parallel Multiple Mediator model Serial Multiple Mediator model Reporting a Mediation Analysis

Simple mediation model � 連續變數 �Y (REACTION) � M (PMI) � 類別變數 �X (COND)

Simple mediation model in PROCESS � � 分析 迴歸 PROCESS Options 勾選 Sobel test and Total effect model

Simple mediation model � 連續變數 �Y (WITHDRAW) � M (AFFECT) � X (ESTRESS)

Simple mediation model in PROCESS � � 分析 迴歸 PROCESS Options 勾選 Sobel test and Total effect model

Parallel Multiple Mediator model � 連續變數 �Y (REACTION) � M 1 (Import) � M 2 (PMI) � 類別變數 �X (COND)

Parallel Multiple Mediator model � 分析 迴歸 PROCESS Model Number 4 Options � Sobel test � Total effect model � Compare Indirect Effect

Serial Multiple Mediator model � 連續變數 �Y (REACTION) � M 1 (Import) � M 2 (PMI) � 類別變數 �X (COND)

Serial Multiple Mediator model � 分析 迴歸 PROCESS Model Number 6 Options � Sobel test � Total effect model � Compare Indirect Effect

Simple Mediator model for more than one Predictors

Simple Mediator model for more than one Predictors � 分析 迴歸 PROCESS Model Number 4 Options � Sobel test � Total effect model

Moderation Analysis � � � Fundamentals of Moderation Analysis Moderation Involving a Dichotomous moderator Interaction between tow Quantitative Variable Truth or Myth about Mean Centering Reporting a Moderation Analysis

Fundamentals of Moderation Analysis � 連續變數 �Y (Liking) � M (Sexism) � 類別變數 �X (Protest)

Conditional and Unconditional Effect X M Y = 4 + 1 X + 2 M + 1. 5 XM -1 -1 -1 0 0 0 1 1 1 2 2 2 0 1 2 3 5 7 4 6 8 5 7 9 6 8 10 3 3 4 4 6 8 5 8 12 6 11 16

Conditional and Unconditional Effect X M Y = 4 + 1 X + 2 M -1 -1 -1 0 0 0 1 1 1 2 2 2 0 1 2 3 5 7 4 6 8 5 7 9 6 8 10

Conditional and Unconditional Effect Y=b 0 + b 1 X + b 2 M + b 3 XM X M Y= 4+1 X+2 M+1. 5 XM -1 -1 -1 0 0 0 1 1 1 2 2 2 0 1 2 3 3 4 4 6 8 5 8 12 6 11 16

Fundamentals of Moderation Analysis � � 分析 迴歸 PROCESS Model 1 Options � Generate data for plotting � Conditioning � Percentiles � Johnson-Newman

Fundamentals of Moderation Analysis � Visualizing moderation (percentiles)

Fundamentals of Moderation Analysis � Visualizing moderation (Johnson-Newman)

Fundamentals of Moderation Analysis � 連續變數 �Y (Liking) � M (Protest) � 類別變數 �X (Sexism)

Fundamentals of Moderation Analysis � � 分析 迴歸 PROCESS Model 1 Options � Generate data for plotting � Conditioning � Percentiles

Fundamentals of Moderation Analysis � Visualizing moderation (percentiles)

Extending moderation analysis � 連續變數 �Y (GOVACT) � M (AGE) � X (NEGMOT) � 控制變數 � POSEMOT � IDEOLOGY � SEX

Extending moderation analysis � � 分析 迴歸 PROCESS Model 1 Options � Mean center for product � Generate data for plotting � Conditioning � Percentiles

Extending moderation analysis

Extending moderation analysis � Visualizing moderation (percentiles)

More than one moderators � � � Y (GOVACT) X (NEGMOT) M (SEX) W (AGE) Control variables � POSMOT � IDIOLOGY

More than one moderators � 分析 迴歸 PROCESS Model 2

More than one moderators

More than one moderators � � � Y (GOVACT) X (NEGMOT) M (SEX) W (AGE) Control variables � POSMOT � IDIOLOGY

More than one moderators � � 分析 迴歸 PROCESS Model 3 Options � Generate data for plotting � Conditioning � Johnson-Newman

More than one moderators � Plot mean ± 1 SD interaction (Aiken and West)

Moderated Mediation

Moderated Mediation � � Y (LIKING) X (PROTEST) M (RESPAPER) W (SEXISM)

Moderated Mediation � � 分析 迴歸 PROCESS Model 8 Conditioning � Percentiles

Can a variable simultaneously Mediate and Moderate another variable’s effect

Questions? 48

三星統計服務有限公司 http: //www. semsoeasy. com. tw/ 49