Overview Lecture 4 Research Methods Research Designs Review

  • Slides: 24
Download presentation
Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics Developmental Research Designs Measurement Effects: Confounds Better Ways? Psychometrics

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics Developmental Research Designs Measurement Effects: Confounds Better Ways? Psychometrics

Simple Correlational Designs Basic Concepts • Two variables • Statistic can range from -1

Simple Correlational Designs Basic Concepts • Two variables • Statistic can range from -1 to +1 • Cannot infer cause-effect relationships

Basic Concepts Simple correlation- Positive. . . . . Reaction time Age

Basic Concepts Simple correlation- Positive. . . . . Reaction time Age

Basic Concepts Simple correlation- Negative Depression . . Age

Basic Concepts Simple correlation- Negative Depression . . Age

Experimental Designs Design Basics Independent • Variable that is manipulated Variable • Characteristic along

Experimental Designs Design Basics Independent • Variable that is manipulated Variable • Characteristic along which people differ Dependent • Variable that is observed

Experimental Designs Experimental design • Manipulation of independent variable • Respondents randomly assigned to

Experimental Designs Experimental design • Manipulation of independent variable • Respondents randomly assigned to groups • Determine cause and effect Quasi-experimental design • Compare groups on predetermined characteristic • Cannot draw cause and effect conclusions • Used in most studies of aging

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics Developmental Research Designs Measurement Effects: Confounds Better Ways? Psychometrics

Developmental Research Designs Basic Designs Cross-Sectional Design Longitudinal Design

Developmental Research Designs Basic Designs Cross-Sectional Design Longitudinal Design

Cross-sectional Design Year of test 1970 Year of birth 1980 1990 2000 1920 50

Cross-sectional Design Year of test 1970 Year of birth 1980 1990 2000 1920 50 yrs. 60 yrs. 70 yrs. 80 yrs. 1930 40 yrs. 50 yrs. 60 yrs. 70 yrs. 1940 30 yrs. 40 yrs. 50 yrs. 60 yrs.

Longitudinal Design Year of test 1970 Year of birth 1980 1990 2000 1920 50

Longitudinal Design Year of test 1970 Year of birth 1980 1990 2000 1920 50 yrs. 60 yrs. 70 yrs. 80 yrs. 1930 40 yrs. 50 yrs. 60 yrs. 70 yrs. 1940 30 yrs. 40 yrs. 50 yrs. 60 yrs.

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics Developmental Research Designs Measurement Effects: Confounds Better Ways? Psychometrics

Measurement Effects: Confounds Effects Term Definition Index of: Change within individual Age Chronological age,

Measurement Effects: Confounds Effects Term Definition Index of: Change within individual Age Chronological age, measured in years Cohort Year or period of birth Influences relevant to history Time of testing Current influences Time of measurement

Measurement Effects: Confounds Basic problem: Cannot separate “age” from “time” Design Confound Age effects

Measurement Effects: Confounds Basic problem: Cannot separate “age” from “time” Design Confound Age effects cannot be separated from historical change Longitudinal Seem to measure age “changes” but. . Cross-Sectional Age effects may be Measures age group differences due to cohort differences because…

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics Developmental Research Designs Measurement Effects: Confounds Better Ways? Psychometrics

Better Ways? Sequential Research Designs (Shaie’s Designs) • Involve a “sequence” of designs •

Better Ways? Sequential Research Designs (Shaie’s Designs) • Involve a “sequence” of designs • Allow inferences to be made about age effects and historical influences (cohort effects and time of measurement/test year effects)

Sequential Research Designs: Schaie’s “Most Efficient Design” Comparison of Designs Design Factor 1 Factor

Sequential Research Designs: Schaie’s “Most Efficient Design” Comparison of Designs Design Factor 1 Factor 2 Time sequential Age Test Year Cohort sequential Age Cohort Cross sequential Cohort Test year

Alternative Sequential Designs (Baltes) Cross-Sectional Sequences Year of test 1970 Year of birth 1980

Alternative Sequential Designs (Baltes) Cross-Sectional Sequences Year of test 1970 Year of birth 1980 1990 2000 1920 50 yrs. 60 yrs. 70 yrs. 80 yrs. 1930 40 yrs. 50 yrs. 60 yrs. 70 yrs. 1940 30 yrs. 40 yrs. 50 yrs. 60 yrs.

Alternative Sequential Designs (Baltes) Longitudinal Sequences Year of test 1970 Year of birth 1980

Alternative Sequential Designs (Baltes) Longitudinal Sequences Year of test 1970 Year of birth 1980 1990 2000 1920 50 yrs. 60 yrs. 70 yrs. 80 yrs. 1930 40 yrs. 50 yrs. 60 yrs. 70 yrs. 1940 30 yrs. 40 yrs. 50 yrs. 60 yrs.

Correlational Approaches: Multivariate Correlational Designs • Multiple variables • Causal relationships estimated

Correlational Approaches: Multivariate Correlational Designs • Multiple variables • Causal relationships estimated

Correlational Approaches in Research on Aging: Multivariate Correlational Designs Types Multiple regression analysis •

Correlational Approaches in Research on Aging: Multivariate Correlational Designs Types Multiple regression analysis • Predictor variables used to estimate dependent variable • Can suggest and test causal relationships Structural equations • Diagram a set of relationships • Test data against hypothesized model

Correlational Approaches in Research on Aging: Multivariate Correlational Designs See text page 68

Correlational Approaches in Research on Aging: Multivariate Correlational Designs See text page 68

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics

Overview- Lecture 4 Research Methods • • • Research Designs: Review of the Basics Developmental Research Designs Measurement Effects: Confounds Better Ways? Psychometrics

Measurement Issues: Psychometrics Forms of reliability Test-retest Give test on two occasions Test gives

Measurement Issues: Psychometrics Forms of reliability Test-retest Give test on two occasions Test gives consistent Internal consistency responses Test items yield similar answers Content Forms of validity Factual content Test measures what it measured correctly Construct is supposed to Scores fit with theory measure Criterion Scores predict actual performance Need to be established on all age groups tested