Fun With Structural Equation Modelling in Psychological Research Jeremy Miles IBS, Derby University.

Slides:



Advertisements
Similar presentations
Basic Statistics Correlation.
Advertisements

Cause (Part II) - Causal Systems I. The Logic of Multiple Relationships II. Multiple Correlation Topics: III. Multiple Regression IV. Path Analysis.
On / By / With The building blocks of the Mplus language.
Writing up results from Structural Equation Models
Multiple Regression and Model Building
Structural Equation Modeling. What is SEM Swiss Army Knife of Statistics Can replicate virtually any model from “canned” stats packages (some limitations.
SEM PURPOSE Model phenomena from observed or theoretical stances
Lesson 10: Linear Regression and Correlation
StatisticalDesign&ModelsValidation. Introduction.
Structural Equation Modeling Using Mplus Chongming Yang Research Support Center FHSS College.
General Structural Equation (LISREL) Models
Structural Equation Modeling: An Overview P. Paxton.
Chapter 12 Simple Linear Regression
Mean, Proportion, CLT Bootstrap
 A description of the ways a research will observe and measure a variable, so called because it specifies the operations that will be taken into account.
Hypothesis Testing Steps in Hypothesis Testing:
Irwin/McGraw-Hill © Andrew F. Siegel, 1997 and l Chapter 12 l Multiple Regression: Predicting One Factor from Several Others.
Linear Regression and Binary Variables The independent variable does not necessarily need to be continuous. If the independent variable is binary (e.g.,
Chapter 12 Simple Linear Regression
1 1 Slide © 2014 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole.
Structural Equation Modeling
The Simple Regression Model
Intro to Statistics for the Behavioral Sciences PSYC 1900
“Ghost Chasing”: Demystifying Latent Variables and SEM
Structural Equation Modeling
Topics: Regression Simple Linear Regression: one dependent variable and one independent variable Multiple Regression: one dependent variable and two or.
1 Simple Linear Regression Chapter Introduction In this chapter we examine the relationship among interval variables via a mathematical equation.
Analysis of Individual Variables Descriptive – –Measures of Central Tendency Mean – Average score of distribution (1 st moment) Median – Middle score (50.
Simple Linear Regression. Introduction In Chapters 17 to 19, we examine the relationship between interval variables via a mathematical equation. The motivation.
AM Recitation 2/10/11.
Regression Analysis Regression analysis is a statistical technique that is very useful for exploring the relationships between two or more variables (one.
Inference for regression - Simple linear regression
1 Least squares procedure Inference for least squares lines Simple Linear Regression.
BPS - 3rd Ed. Chapter 211 Inference for Regression.
1 1 Slide Simple Linear Regression Coefficient of Determination Chapter 14 BA 303 – Spring 2011.
Educational Research: Competencies for Analysis and Application, 9 th edition. Gay, Mills, & Airasian © 2009 Pearson Education, Inc. All rights reserved.
Correlation and Regression Used when we are interested in the relationship between two variables. NOT the differences between means or medians of different.
Multiple Regression The Basics. Multiple Regression (MR) Predicting one DV from a set of predictors, the DV should be interval/ratio or at least assumed.
1 Chapter 12 Simple Linear Regression. 2 Chapter Outline  Simple Linear Regression Model  Least Squares Method  Coefficient of Determination  Model.
Multiple Regression and Model Building Chapter 15 Copyright © 2014 by The McGraw-Hill Companies, Inc. All rights reserved.McGraw-Hill/Irwin.
Lecture 8 Simple Linear Regression (cont.). Section Objectives: Statistical model for linear regression Data for simple linear regression Estimation.
Basic Concepts of Correlation. Definition A correlation exists between two variables when the values of one are somehow associated with the values of.
MGS3100_04.ppt/Sep 29, 2015/Page 1 Georgia State University - Confidential MGS 3100 Business Analysis Regression Sep 29 and 30, 2015.
Measurement Models: Exploratory and Confirmatory Factor Analysis James G. Anderson, Ph.D. Purdue University.
Educational Research Chapter 13 Inferential Statistics Gay, Mills, and Airasian 10 th Edition.
Chapter 16 Data Analysis: Testing for Associations.
Controlling for Baseline
CFA: Basics Beaujean Chapter 3. Other readings Kline 9 – a good reference, but lumps this entire section into one chapter.
Multiple Regression. Simple Regression in detail Y i = β o + β 1 x i + ε i Where Y => Dependent variable X => Independent variable β o => Model parameter.
G Lecture 91 Measurement Error Models Bias due to measurement error Adjusting for bias with structural equation models Examples Alternative models.
SEM Basics 2 Byrne Chapter 2 Kline pg 7-15, 50-51, ,
CJT 765: Structural Equation Modeling Class 8: Confirmatory Factory Analysis.
Structural Equation Modeling Mgmt 291 Lecture 3 – CFA and Hybrid Models Oct. 12, 2009.
Chapter 12 Simple Linear Regression n Simple Linear Regression Model n Least Squares Method n Coefficient of Determination n Model Assumptions n Testing.
Chapter 13 Understanding research results: statistical inference.
Copyright © 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins Chapter 18 Multivariate Statistics.
©The McGraw-Hill Companies, Inc. 2008McGraw-Hill/Irwin Linear Regression and Correlation Chapter 13.
BPS - 5th Ed. Chapter 231 Inference for Regression.
The SweSAT Vocabulary (word): understanding of words and concepts. Data Sufficiency (ds): numerical reasoning ability. Reading Comprehension (read): Swedish.
ESTIMATION.
REGRESSION G&W p
12 Inferential Analysis.
CHAPTER 29: Multiple Regression*
Prepared by Lee Revere and John Large
Writing about Structural Equation Models
Structural Equation Modeling
12 Inferential Analysis.
Causal Relationships with measurement error in the data
Structural Equation Modeling
Presentation transcript:

Fun With Structural Equation Modelling in Psychological Research Jeremy Miles IBS, Derby University

Structural Equation Modelling Analysis of Moment Structures Covariance Structure Analysis Analysis of Linear Structural Relationships (LISREL) Covariance Structure Models Path Analysis

Normal Statistics Modelling process –What is the best model to describe a set of data –Mean, sd, median, correlation, factor structure, t-value DataModel

SEM Modelling process –Could this model have led to the data that I have? ModelData

Theory driven process –Theory is specified as a model Alternative theories can be tested –Specified as models Data Theory A Theory B

Ooohh, SEM Is Hard It was. Now its not Jöreskog and Sörbom developed LISREL –Matrices: x y –Variables: X Y –Intercepts:

The Joy of Path Diagrams Variable Causal Arrow Correlational Arrow

Doing Normal Statistics xy Correlation

Doing Normal Statistics xy T-Test

Doing Normal Statistics x1x1 y One way ANOVA (Dummy coding) x2x2 x3x3

Doing Normal Statistics x1x1 y Two- way ANOVA (Dummy coding) x2x2 x 1 * x 2

Doing Normal Statistics x y Regression x x

Doing Normal Statistics MANOVA x1x1 x2x2 y1y1 y2y2 y3y3

Doing Normal Statistics ANCOVA xy z

etc...

Identification Often thought of as being a very sticky issue Is a fairly sticky issue The extent to which we are able to estimate everything we want to estimate

X = 4 Unknown: x

x = 4 y = 7 Unknown: x, y

x + y= 4 x - y = 1 Unknown: x, y

x + y = 4 Unknown: x, y

Things We Know Things We Want to Know = x=4 x + y = 4, x - y = 2 Just identified Can never be wrong Normal statistics are just identified

Things We Know Things We Want to Know < x + y = 7 Not identified Can never be solved

Things We Know Things We Want to Know > x + y = 4, x - y = 2, 2x - y = 3 over-identified Can be wrong SEM models are over-identified

Identification We have information –(Correlations, means, variances) Normal statistics –Use all of the information to estimate the parameters of the model –Just identified All parameters estimated Model cannot be wrong

Over-identification SEM –Over-identified –The model can be wrong If a model is a theory –Enables the testing of theories

Parameter Identification x - 2 = y x + 2 = y Should be identified according to our previous rules –its not though There is model identification –there is not parameter identification

Sampling Variation and 2 Equations and numbers –Easy to determine if its correct Sample data may vary from the model –Even if the model is correct in the population Use the 2 test to measure difference between the data and the model –Some difference is OK –Too much difference is not OK

Simple Over-identification xy Estimate 1 parameter -just-identified xy Estimate 0 parameters -over-identified

Example 1 R ab = 0.3, N = 100 Estimate = 0.3, SE = 0.105, C.R. = The correlation is significantly different from 0 ab

Model Tests the hypothesis that the correlation in the population is equal to zero –It will never be zero, because of sampling variation –The 2 tells us if the variation is significantly different from zero ab

Example 2 Test the model Force the value to be zero –Input parameters = 1 –Parameters estimated = 0 The model is now over-identified and can therefore be wrong ab

The program gives a 2 statistic The significance of difference between the data and the model –Distributed with df = known parameters - input parameters 2 = 9.337, df = = 1, p = So what? A correlation of 0.3 is significant?

Hardly a Revelation No. We have tested a correlation for significance. Something which is much more easily done in other ways But –We have introduced a very flexible technique –Can be used in a range of other ways

Testing Other Than Zero Estimated parameters usually tested against zero –Reasonable? Model testing allows us to test against other values 2 = 2.3, n.s. Example 3 ab 0.15

Example 4: Comparing correlations 4 variables –mothers' sensitivity –mothers' parental bonding –fathers' sensitivity –fathers' parental bonding Does the correlation differ between mothers and fathers?

M S M PB F PB F S

Example 4a –analyse with all parameters free –0 df, model is correct Example 4b –fix FS-FPB and MS-MPB to be equal. –See if that model can account for the data

M S M PB F PB F S dave 2 = 1.82, df = 1 p = dave = 0.41 (s.e. 0.08)

Latent Variables The true power of SEM comes from latent variable modelling Variables in psychology are rarely (never?) measured directly –the effects of the variable are measured –Intelligence, self-esteem, depression –Reaction time, diagnostic skill

Measuring a Latent Variable Latent variables are drawn as ellipses –hypothesised causal relationship with measured variables Measured variable has two causes –latent variable –other stuff random error Latent Measured

x = t + e Reliability is: the square root of proportion of variance in x that is accounted the correlation between x and e Measured True Score Error

Identification and Latent Variables 1 measured variable –not (even close to) identified 4 measured variables –6 known, 4 estimated model is identified

Need four measured variables to identify the model Need to identify the variance of the latent variable –fix to 1

Why oh why oh why? Why bother with all these tricky latent variables? 2 reasons –unidimensional scale construction –attenuation correction

Unidimensionality Correlation matrix 2 = 3.65, df = 2, p =

Attenuation Correction Why bother? –Gets accurate measure of correlation between true scores Why bother –theories in psychology are ordinal –attenuation can only cause relationships to lower

The Multivariate Case Much more complex and unpredictable x1x1 y1y1 x2x2 y2y2 a c d e b

Some More Models Multiple Trait Multiple Method Models (MTMM) Temporal Stability Multiple Indicator Multiple Cause (MIMIC)

MTMM Multiple Trait –more than one measure Multiple Method –using more than one technique Variance in measured score comes from true score, random error variance, and systematic error variance, associated with the shared methods

What? Example 6 (From Wothke, 1996) –Three traits Getting along with others (G) Dedication (D) Apply learning (L) Three methods Peer nomination (PN) Peer Checklist (PC) Supervisor ratings (SC)

Matrix g.pn d.pn l.pn g.pc d.pc l.pc g.sc d.sc l.sc

Analysis g.pnl.pnd.pc pn g.pnl.pcd.pc pc g.scl.scd.sc sc gl d

Temporal Stability Usually –sum the items –correlate them BUT –items may not be unidimensional –relationship will be attenuated due to measurement error –relationship will be inflated, due to correlated error

L1L1 X 3.1 X 4.1 X 5.1 X 2.1 X 1.1 L2L2 X 3.2 X 4.2 X 5.2 X 2.2 X 1.2 Corrects for attenuation But - correlated errors may be a problem

Added correlated errors Example 7b L1L1 X 3.1 X 4.1 X 5.1 X 2.1 X 1.1 L2L2 X 3.2 X 4.2 X 5.2 X 2.2 X 1.2

MIMIC Model Conventional wisdom in psychological measurement is that a latent variable is the cause of the measured variables Assumption is made (implicitly) in many types of measurement –Bollen and Lennox (1989) –not necessarily the case

Value of a Car Causes –type, size, age, rustiness –no reason they should, or should not, be correlated Effects –assessment of value by people who know

Level of Depression Questionnaire items –causes or effects? been feeling unhappy and depressed? been having restless and disturbed nights? found everything getting 'on top' of you? MIMIC

Example 8: MIMIC L1L1 c1c1 c2c2 c3c3 y4y4 y1y1 LY 1 LY 2 y2y2 y3y3 y5y5 y6y6 y7y7 y8y8

Concluding remarks Given a taster –some may be too simple? Much more to say –no time to say it See further reading (Books and WWW)

Further Info SEMNET - list (messages) bama.ua.edu (leave) – the semnet FAQ

Books See web page

References See web page