Sharon Seidman Ph D CAS 301 Inquiry Methodology Sharon Seidman Ph D CAS 301 More on Data Collection Group Comparison Part 2 Review Research Process Examine association between variables Develop hypothesis or question State specific testable prediction Collect data 1 Sharon Seidman Ph D CAS 301 Inquiry Methodology Variables Basis for research Individual elements of hypotheses predictions Must vary Has levels A variable Is a category Is a member of category Can change Cannot change Variable Definitions Concept Valid Face Predictive Concurrent Convergent Discriminant Operation Reliable Internal Test retest Inter rater Scale Ratio Interval Ordinal Nominal Variables Causality Types of variables Subject Situation Response Created vs pre existing variables Subject always pre existing Situation response may be Pre existing Created 2 Sharon Seidman Ph D CAS 301 Inquiry Methodology Goals of Research Descriptive Statistics Description Prediction Inferential Statistics Causation Describing a Group Norms Central Tendencies Mean Median Mode Individual Differences Variability Range Standard Deviation Kinds of Inferences Linear Relations Correlations Group Comparisons Experiments 3 Sharon Seidman Ph D CAS 301 Inquiry Methodology Comparison Between 2 Groups Based on Difference between group norms means Differences within groups standard deviation Odds that sample represents population degrees of freedom Produces probability of Type 1 Error Example Is the number of fights in classrooms with 20 children significantly greater than the number of fights in classrooms with 10 children Steps Sample classrooms in each group Count the fights in each room Compare groups to see if norm for 20 children classes is typical of 10 children classes Number of Classes Frequency of Fighting 20 Student Classes 4 5 6 7 8 9 10 Number of Fights 4 Sharon Seidman Ph D CAS 301 Inquiry Methodology Number of Classes Frequency of Fighting 10 Student Classes 2 3 4 5 6 7 8 Number of Fights Properties of a Normal Distribution 2 5 Area 2 2 2 5 13 6 34 1 34 1 13 6 2 2 Standard 3S 2S Deviation 1S Mean 1S 2S 3S 95 Comparing Groups 4 5 10 Children 6 7 8 9 10 20 Children 5 Sharon Seidman Ph D CAS 301 Inquiry Methodology T statistic Numerical representation of difference between groups t 18 3 75 p 05 Degrees of Freedom Probability of Type 1 Error Difference between means divided by standard deviation Degrees of freedom Related to sample size Possible variability in sample How much freedom is there to get a different score Probability Errors Research Decision Reality Null is True Accept Null Accept Hypothesis Type I Error Hypothesis is True Type II Error 6 Sharon Seidman Ph D CAS 301 Inquiry Methodology Variables in Between Group Comparisons Grouping Variable Called independent or predictor Used as nominal measure Outcome Variable Called dependent or criterion Must Be interval ratio or Likert Accommodate math Developmental Designs Longitudinal CrossSectional 1 1 1 2 2 2 3 3 4 4 Sequential 2 3 3 4 7 Sharon Seidman Ph D CAS 301 Inquiry Methodology Intervention Designs Frequency Score Principle of Intervention Studies Before After Intervention Frequency Score Baseline Analysis Single Case Before After Intervention 8 Sharon Seidman Ph D CAS 301 Inquiry Methodology Frequency Score Pretest Posttest Analysis After Group Before Intervention Averages Frequency Score Posttest vs Control Analysis Test Group Control Group After Group Before Intervention Averages Sampling Assignment 9 Sharon Seidman Ph D CAS 301 Inquiry Methodology Sampling Identification of participants Strategies Random Entirely random Stratified random Cluster Haphazard Truly haphazard Random Quota Random Stratified Easy Haphazard Quota Assignment Creation of experimental groups Must be determined by experimenter Only applies to Independent Variables Strategies Random Haphazard Methods of Data Collection 10 Sharon Seidman Ph D CAS 301 Inquiry Methodology Data Collection Techniques Testing Observation Survey questionnaire Interview Content analysis Pre existing source archival data Observation Strategies Naturalistic vs Systematic Participant vs Concealed Issues Reactivity Reliability coding recording Sampling vs comprehensive Questionnaire Issues Measure ideas rather than behavior Ease data collection Evaluate sample characteristics Distribution Find target population Ensure adequate response rate Format Close vs open ended questions Influence of wording Reactivity Social desirability Order effects 11 Sharon Seidman Ph D CAS 301 Inquiry Methodology Content Analysis Analysis of thing rather than person Requires coding counting Interviews Same risks as Observations Surveys Advantages Detailed qualitative information Clinical interview flexibility Factorial Designs 12 Sharon Seidman Ph D CAS 301 Inquiry Methodology More Complex Designs X Multiple levels of a variable More than 2 groups X Multiple independent variables 2 IVs and a DV Fear Dark Room Light Room People Stories Monster Stories Fear Main Effects Interaction Effects Fear Dark Room Light Room 13 Sharon Seidman Ph D CAS 301 Inquiry Methodology Numerical Analysis Dark Light Monster 6 1 People 3 1 Cell Means Show Interaction Effects Dark Light 3 5 Monster 6 1 2 People 3 1 4 5 1 Marginal Means Show Main Effects of Single IV Statistical Analysis for Factorials Analysis of Variance ANOVA Described by of IV Predictors One way Two way Etc Described by levels of IV Predictors 2x2 2x3 14 Sharon Seidman Ph D CAS 301 Inquiry Methodology One Way ANOVA Little Medium Lots Two Way ANOVA Fear Dark Room Light Room 3 x 2 ANOVA Age Grade th 7 9th 11th Male Gender Female 15 Sharon Seidman Ph D CAS 301 Inquiry Methodology Statistical Analysis Procedure ANOVA Statistic F Differences Between Groups Differences Within Groups Same procedure but more groups than t test Degrees of Freedom 1st Variability from groups Levels of IV 1 nd 2 Variability from subjects error Factorial Example Effect of Story Light on Fear 3 Comparisons 2 main effects 1 interaction Statistics 3 statistical statements Each analysis uses different means 16 Sharon Seidman Ph D CAS 301 Inquiry Methodology Example Dark Light Monster 6 1 3 5 People 3 1 2 4 5 1 Source of Variance Type of Story Room lighting Interaction Error df 1 1 1 36 F 15 20 22 Results Main effect Marginal Means F statement There was a significant main effect for type of story children were more fearful with monster stories M 3 5 than people stories M 2 0 F 1 36 15 p 05 Results Interaction effect Cell Means F