Share Course Ware
Medicine > Psychiatry, Neurology > Statistics in Psychosocial Research: Measurement
 Statistics in Psychosocial Research: Measurement   posted by  boym   on 3/24/2008  Add Courseware to favorites Add To Favorites  
Abstract/Syllabus
Courseware/Lectures
Test/Tutorials
Further Reading
Webliography
Downloads
More Options
 
Abstract/Syllabus:

330.657

Statistics in Psychosocial Research: Measurement


Photo courtesy of the OpenEye via flickr. Some rights reserved.

Staff

Instructors: William Eaton, Elizabeth Garrett-Mayer, and Jeannie-Marie Leoutsakos

Originally Offered

Fall 2006

Offered By

Department of Mental Health

 

Description

Presents quantitative approaches to measurement in the psychological and social sciences. Topics include the principles of psychometrics, including reliability and validity; the statistical basis for latent variable analysis, including exploratory and confirmatory factor analysis and latent class analysis; and item response theory. Draws examples from the social sciences, including stress and distress, social class and socioeconomic status, personality; consumer satisfaction, functional impairment and disability, quality of life, and the measurement of overall health status. Intended for doctoral students.

 

OCW offers a snapshot of the content used in courses offered by JHSPH. OCW materials are not for credit towards any degrees or certificates offered by the Johns Hopkins Bloomberg School of Public Health.

Schedule

 


SESSION # TOPIC ACTIVITIES
 
1

Introduction to Measurement

After this class students will be able to (1) briefly describe the concept of reliability in both intuitive and statistical terms, and (2) identify the key assumptions of classical test theory.

Introduction

Syllabus review

Classical test theory

Introduction to reliability

2

Measuring Association and Dimensionality

After this class students will be able to (1) measure associations between continuous observed variables using covariances and correlations, (2) measure magnitudes of association between discrete observed variables, and (3) define multidimensionality regarding latent variables.

Covariance

Pearson & Spearman correlation

Correlations with non-linear data

Polychoric correlation

Covariance, correlation, and odds ratio matrices

Dimensionality

3

Principles of Psychometrics: Reliability I

After this class students will be able to (1) describe two definitions of the concept of reliability, (2) predict how long a scale should be, and (3) estimate reliability for continuous and categorical measures

Types of reliability (Interrater, test-retest reliability, internal consistency)

Different types of reliability coefficients (correlation, split half measures, Alpha coefficient, Kuder Richardson Coefficient, Kappa

4

Principles of Psychometrics: Reliability II

After this class students will be able to (1) describe the relationship of the intraclass correlation coefficient to other measures of reliability, and (2) correctly identify which intraclass correlation to use for different research designs

ANOVA model for reliability

Intra-class Correlation Coefficient

Research Designs

5

Principles of Psychometrics: Validity I

After this class students will be able to (1) distinguish four different types of validity, (2) describe the conceptual and quantitative relationship of reliability to validity, (3) estimate a true correlation from an observed correlation

Types of Validity (face, content, criterion, construct)

Relationship of Reliability to Validity

Correction for attenuation

6

Principles of Psychometrics: Validity II

After this class students will be able to evaluate the relative utility of different cutoffs for a measure in relation to a gold standard.

Internal Construct Validity

External Construct Validity

Multi-trait Multimethod Matrix

Sensitivity and Specificity

ROC Curves

7

Scale Development

After this class students will be able to describe procedures for constructing a scale from scratch

Scale Construction
8

Factor Analysis I

After this class students will be able to (1) identify when a factor analysis is appropriate and when it is not, (2) run a one-factor and multi-factor analysis, (3) interpret the results from a factor analysis

Introduction to factor analysis

The orthogonal factor model

Loadings

Principal components

Eigenvalues

Introduction to rotation

Communalities/Uniqueness of items

9

Factor Analysis II

After this class students will be able to (1) use the statistical procedure of rotation to aid in the interpretation of results from a factor
analysis, (2) be able to apply both orthogonal and oblique rotations and identify the assumptions
underlying each, and (3) apply the appropriate method of estimation for factor analysis

Factor extraction

Methods of estimation and rotation: orthogonal and oblique

Choosing the number of factors

Factor scores

Confirmatory factor analysis

Conditional independence Dichotomous factor analysis

10

Factor Analysis III

In this class students will (1) apply factor analysis to real data, and (2) critique published use of factor analysis

Journal examples
11

Latent Class Analysis I

After this class students will be able to (1) differentiate when to use factor analysis and when to use latent class analysis, and (2) interpret output from a latent class analysis

The latent class model

The response pattern matrix

Choosing the number of classes

Conditional probabilities

Interpreting the model

Examples: depression and functioning

12

Latent Class Analysis II

After this class students will be able to (1) estimate a latent class analysis, and (2) to interpret different criterion to choose among alternative models

Statistical model and assumptions

Exploration: response patterns

Issues of model fitting

Identifiability

Checking the model: tests and displays

13

Latent Class

In this class students will (1) apply latent class analysis to real data, and (2) critique published use of latent class analysis

Journal Examples
14

Sample Size in Reliability and Factor Analysis

After this class students will be able to (1) estimate the sample size needed to for scales with targeted reliability levels, (2) estimate the sample size needed for pilot studies that will use factor analysis, and (3) understand issues involved with sample size in latent class analysis.

Targeted Reliability Levels

Pilot Studies

Latent Class Analysis

15 Review
16 Final Exam




www.sharecourseware.org   Tell A Friend