HPS301 Exam Notes- Contents Week 1 Research Design: What characterises different approaches 1 Experimental Design 1 Key Features 1 Criteria for establishing causality 2 Validity Internal Validity 2 Threats to Internal Validity 2 Major Types of Threats to Internal Validity 3 Threats to Internal Validity over Time 3 External Validity 3 Threats to External Validity 4 Threats to Internal Validity and External Validity 4 Extraneous Variables 5 How are extraneous variables controlled in experiments? 5 Group assignment and manipulation of IV 5 Techniques used to attempt equivalence of groups: 6 Keep Variables Constant 6 Balance or Match across Independent Variable 6 Control by Randomisation 6 Consequences of non-random assignment and not manipulating the IV 6 Between-subjects Design 6 Within-subjects Design 7 Non-experimental Design 7
Quasi-experimental research 8 Experiments vs Quasi Experiments 8 Correlational research 8 Strengths and Weaknesses 8 Third Variable Problem 9 Directionality 9 Descriptive research 9 Time Related Factors 9 Controlling for time related factors 10 Counterbalancing 10 Complete 10 Partial 10 Possible Orders 10 Non-experimental research 11 Non-equivalent group design 11 Problems 11 Differential Research 11 Posttest-Only NonEquivalent Control Group Design 11 Pretest-Posttest Nonequivalent Control Group Design 12 Pre-Post Designs 12 One-Group Pretest-Posttest 12 Time-Series 13 Interrupted Time-Series 13 Equivalent Time-Samples 13 Multiple Time-Series 13
Developmental 13 Cross-Sectional Research 13 Longitudinal Research 14 Cross-sectional Longitudinal 15 Longitudinal-Sequential Design 15 Summary of Designs 16 Week 2 Correlations and Bivariate Regression Correlations 17 Relationships/associations 17 Pearson Correlation 17 Strength of Relationships 18 Perfect Correlation 18 Strength and Direction 18 Coefficient of determination 19 Significance test of r 19 T-test 19 Spearman 20 Bivariate Regression 20 Regression lines 20 Regression equation 21 Example 22 Residual 22 Factors that affect correlations 22 Range restriction 22 Heterogeneous sample 23
Unequal split of responses on a dichotomous measure 23 Outliers 23 Assumptions of the General Linear Model 23 Normality 23 Linearity 24 Homogeneity of variance in arrays 24 Week 3 - Multiple Regression Multiple Regression 25 Bivariate regression vs Multiple regression 25 Problem with Bivariate regression 25 Multiple Regression 25 Questions that can be answered with multiple regression: 26 Combined predictive utility 26 Importance of the IVs 26 Uniqueness 27 Improvement in prediction 27 Introducing B, beta (β), r and sr 27 B = unstandardized coefficient 27 Beta = standardized coefficient 28 r = Pearson s correlation 28 sr = unique relation between IV and DV 28 Standard multiple regression 28 Squared semi-partial correlation (sr2) 29 Significance tests 29 Sequential multiple regression 30
Significance tests for Sequential multiple regression 30 Questions that can be answered with hierarchical multiple regression 31 Adjusted R2 31 Considerations when using Multiple Regression 32 Correlational model 32 Selection of predictor variables 32 Multicollinearity and singularity 32 Other considerations 33 Week 4 - ANOVA Overview 34 Questions answered by ANOVA 34 Logic of ANOVA 34 Assumptions of ANOVA 35 Homogeneity of variance 35 Normality 35 Independence of observations 35 Calculations 35 Definitions 35 Calculating Sum of Squared Differences 35 SStotal 35 SSwithin 36 SSbetween 36 Step by Step Example 36 Magnitude of experimental effect 38 Eta-Squared (η2)(1) 39
Power 39 Factors which increase power 39 Violation of the homogeneity of variance 40 What to do about violations of homogeneity of variance? 40 Welch and Brown-Forsythe 41 Answering ANOVA questions with tables using graphs 41 Week 5 Multiple Comparisons Things to consider when conducting multiple comparisons 42 Type 1 Error 42 Multiple tests of significance 42 Error Rate Per Comparison (PC) 43 Familywise error rate (FW) 43 Familywise error and planned comparisons 44 Post-hoc comparison 44 Problems with Post hoc 44 Priori comparisons 45 Multiple planned comparisons 45 Multiple t-tests 45 Linear contrasts 45 Steps to follow in using a linear comparison or contrast for hypothesis testing 45 Calculations 46 Rules for selecting coefficients 47 Sum of squares for contrasts 47 Understanding Logic of Coefficients 47 Orthogonal contrasts 48
Week 6 Factorial ANOVA Definitions 49 Interactions 49 Main Effects 50 Calculations 51 Several key steps involved 51 Calculations 52 Sums of squares 52 Main effects sums of squares 52 Interaction sums of squares 53 Degrees of freedom 54 Mean squares 54 F-values 55 Example 55 ANOVA Summary Table 56 Simple Effects 56 Three-Way ANOVA 57 Main Effects 57 Two-Way Interactions 58 Three-Way Interactions and Simple Interaction Effects 58 Interpreting Effects 58 Summary 59 Week 7 Repeated Measures and Mixed Model ANOVA Within subjects ANOVA 60
Advantages 60 Formulae for a one-way repeated-measures ANOVA 60 Treatment vs Subject Variability 62 Between subjects ANOVA 62 Mixed model 63 Sphericity 63 Failure of sphericity 63 Correcting for violation of sphericity 64 Which correction to apply 64 Week 8 - Sampling Population and Samples 65 Target Population 65 The Accessible Population 65 The Sample 65 Representative samples 65 Sampling error 65 Biased sample 65 Non-sampling error or measurement error: 66 Four key types that can occur 66 Probability Sampling 66 Simple Random Sampling 66 Sampling with replacement 67 Sampling without replacement 67 Systematic Sampling 67 Stratified Random Sampling 67
Proportionate Stratified Random Sampling 67 Cluster Sampling 67 Non-probability Sampling 68 Convenience Sampling 68 Quota Sampling 68 Purposeful Sampling 68 Qualitative sampling 69 Week 9 Surveys and Scales Construct 70 Operational definition 70 Construct validity 70 Generating an operational definition 71 Nomological net 71 Constructing self-report measures 71 Self-report 71 Likert scale 71 Response set 72 Administering surveys and scales 72 Nonresponse bias 72 Interviewer bias 72 Mail surveys 72 Telephone 72 In person 73 Item bias 73 Validity 73
Face validity 74 Concurrent validity 74 Predictive validity 74 Construct validity 74 Convergent validity 74 Divergent validity 74 Reliability 74 Observer error 74 Environmental error 75 Participant changes 75 Test-retest reliability 75 Inter-rater reliability 75 Split-half reliability 75 Assessing psychological tests of constructs 75 Validity 75 Reliability 75 Cronbach s Alpha 75 Split-Half Reliability 76 Week 10 Qualitative methods & interviewing Qualitative Methods 77 When to use it 77 Approaches 77 Phenomenology 77 Ethnography 77 Grounded theory 78
The design 78 Methods 78 Interviews 78 Observation 79 Document analysis 79 Preliminary data analysis 79 Thematic Analysis 79
HPS301 Exam Notes Week 1 Research Design Major Approaches to Research Experimental research Quasi-experimental research Non-experimental research Correlational research Descriptive research Qualitative research What characterises different approaches to research? Random assignment of participants Level of control Manipulation of independent variable Types of Quantitative Design Experimental Designs Used to assess causality, answer cause and effect questions about the relationships between variables Key features: Manipulation of IV - Measurement of DV Comparison across groups (e.g., treatment v control group, pre vs post-manipulation)
Other variables (extraneous variables) are controlled to rule out alternative explanations or prevent these from becoming confounding variables Random assignment of subjects to groups Conducted with rigorous control to ensure unambiguous conclusions Two main types: Between vs Within (can also have mixed design) Experimental designs are desirable, but not always possible Criteria for establishing causality Independent variable must precede dependent variable There needs to be a relationship between independent and dependent variable Need to rule out other possible causes Internal Validity Are conclusions valid for the study? The validity of a research study is the degree to which the study accurately answers questions it was intended to answer Threat to validity: questions or doubts about the study Threats to Internal Validity Any variable that a researcher is not directly interested in is an extraneous variable Any variable that influences two variables being studied is a confounding variable - Confounding variables threaten the internal validity of a study because they provide an alternative explanation for the relationship being investigated. - Changes in performance from condition 1 to condition 2 could be explained in terms of the experimental condition (music v no music), but it may instead be the result of fatigue (if participants perform worse in Condition 2) or practice (if they perform better in Condition 2). It provides an alternative explanation for the relationship - (e.g. depression, early pubertal development, poor attachment styles)
Major Types of Threats to Internal Validity Environmental variables - (e.g. time of testing, different experimenters) Assignment bias - (e.g. use of intact groups, males vs females) - groups may vary in participant characteristics Examining groups over time - the changes in the participants may be due to some other factor - (ie. history, maturation) Threats to internal validity limit your ability to draw clear, unambiguous conclusions from the data you have Threats to Internal Validity Over Time History - other events that happen during the study (e.g. sporting competition, other programs in the school) Maturation - changes in the participants (e.g. changes in height/ weight) Instrumentation - technical issues or researcher skill (e.g., better skilled or more complacent) Testing effects - fatigue or practice Regression toward the mean - extreme scores on first testing tend to be less extreme on second testing External Validity Extent findings can be generalized to people, settings, times, and other conditions beyond the scope of a particular study
Sample to population One study to another From research study to real world Threats to external validity limit your ability to generalise findings beyond the present sample Threats to External Validity Generalizing across participants - How representative is sample of the target population? - Selection bias - Volunteer bias - Participant characteristics - Cross-species generalisation Generalizing across features of a study - Novelty effect - Reactivity - Multiple treatment interference - Experimenter characteristics Generalizing across features of the measures - Repeated measurement (sensitization) - Type of measure used - Timing of the measurement Threats to Internal Validity and External Validity Threats to both: experimenter expectancy, demand characteristics We often make a trade-off between external and internal validity A high degree of control (e.g., lab conditions) may ensure internal validity, but are the results generalizable to real-world contexts? Conversely, behavioural observation in naturalistic settings may have generalizability, but we have less control of extraneous influences Research that is designed to have high external validity will often involve creation of an experimental environment that mimics the real world. Unfortunately, real world scenarios often involve a host of extraneous influences that can impact internal validity of the results.