260 likes | 645 Views
Research Methods for the Social Sciences. Lorne Campbell Christopher J. Wilbur University of Western Ontario. Philosophy of Human Behavior. (1) Behavior is influenced by outside circumstances Experimental approach (2) Behavior is influenced by the qualities possessed by the individual
E N D
Research Methods for the Social Sciences Lorne Campbell Christopher J. Wilbur University of Western Ontario
Philosophy of Human Behavior • (1) Behavior is influenced by outside circumstances • Experimental approach • (2) Behavior is influenced by the qualities possessed by the individual • Correlational approach
Interactionist Perspective • Behavior is a function of both context and individual differences • E.g., extraversion and social dominance
Types of Research Methods • Runkel and McGrath • Developed a circumplex model to describe the goals of the research process, and the basic types of research methods available • Helps structure our thinking of the types of methods available, and the pros/cons of each type of method
B Obtrusive Research Operations Laboratory Experiments Experimental Simulations Judgment Tasks II II Field Experiments III I III I Sample Surveys Field Studies IV IV Unobtrusive Research Operations C Formal Theory Computer Simulations A Universal Behavioral Systems Particular Behavioral Systems
Benefits of Multi-Method Research • Mono-operation bias • When using the same method time after time, your research suffers from the same set of limitations • Using different methods to address the same question(s) helps overcome the limitation of each method • E.g., research on self-concept
Inferential Statistics • Inferential statistics are usually preferred to simply looking at differences because we can conclude with more certainty that the difference accurately characterizes the population
Inferential Statistics • Inferential statistics are usually preferred to simply looking at differences because we can conclude with more certainty that the difference accurately characterizes the population • Is this difference a true difference in the general population or just a random effect based on the particular sample?
Inferential Statistics • Based on analysis of samples, we can make generalizations about the population of interest • H0: Sleep deprivation does not impair performance • H1: Sleep deprivation does impair performance • Compare two groups on performance measure • If a mean difference emerges that is unlikely by chance alone, we assume this difference is accurate of the population
Three Basic Statistical Methods • T-Test • Analysis of Variance (ANOVA) • Multiple Regression
Three Basic Statistical Methods • Analysis of Variance (ANOVA) • Categorical data (i.e., experimental conditions, demographic data) • Between-subjects or within-subjects • Can compare 3 or more conditions or groups • Can examine interactive effects of multiple variables
Three Basic Statistical Methods • Analysis of Variance (ANOVA)
Three Basic Statistical Methods • Analysis of Variance (ANOVA) • Examples • Psychology experiments • Voter intentions • Geographical differences
Three Basic Statistical Methods • Multiple Regression • Continuous data • But can also handle categorical data (subsumes ANOVA) • Y = b0 + b1X1 + b2X2 + ... + bkXk
Three Basic Statistical Methods • Multiple Regression • Y = b0 + b1X1 + b2X2 + b3X1X2
Three Basic Statistical Methods • Multiple Regression
Advanced Statistical Techniques • Structural Equation Modeling
Advanced Statistical Techniques • Structural Equation Modeling Extraversion Risky Sexual Behavior Talkative Condom Use Daring STI Testing Friendly Casual Sex
Advanced Statistical Techniques • Structural Equation Modeling • Hierarchical Linear Modeling
Advanced Statistical Techniques • Hierarchical Linear Modeling
Teaching Statistics • Bachelor’s Level • Year 2 • Introductory statistics course (mathematical; probabilities, logic of inferential statistics, t-tests, ANOVA, correlation/regression, some other assorted tests) • Year 3 • Advanced statistics course (logical; logic of the tests; application of the tests with SPSS)
Teaching Statistics • Masters Level • Year 1 • Advanced statistics course (refreshing and extending; large focus on t-tests, ANOVA, and correlation/multiple regression) • Beyond • Specialized courses in advanced topics (e.g., factor analysis, SEM, HLM, etc.)
How I have taught undergraduate courses on research methods • Brief version of syllabus • Week 1 – introduction to course • Week 2 – Validity and Reliability • Validity • Construct • Internal • External • Reliability • Psychometric properties of scales • Week 3 – Experimental design and the significance testing debate
Week 4 – Quasi-experimental designs • E.g., regression discontinuity design, field experiment • Week 5 – Field Studies, simulation methods • E.g., research by Doug Kenrick • Week 6 – Diary research • Week 7 – Multilevel modelling • Week 8 – Dyadic data (collection and analysis)
Week 9 – Social Relations Model (SRM) • E.g, loneliness study • Week 10 – Mediation and Moderation • Week 11 – Methods in Social Cognition • E.g., AMP model • Week 12 – Meta-analysis • Week 13 – Research Ethics