• The independent variable is • A. Cheering • B. Noise • C. Number of people in the group • D. Social loafing 25
Identify Key Experimental Features Well-being is related to by amount of small talk in faculty parties What kind of study? A.Correlational B.Experimental 26
Identify Key Experimental Features • In a representative sample of Florida mothers, maternal employment is related to child behavior problems What kind of study? A.Correlational B.Experimental In this study, the independent variable is ______ and the dependent variable is _______. a.Behavior problems, mother employment b.Mother employment, behavior problems 27
The Four Validities
• Reliability & validity of measured variables • Short version: how well the measure of the variable is capturing the idea • Consider depression • Diagnostic interview • Observer judgment • Self assessment (survey question) • Validity (via manipulation check) of manipulated variables • Consider induction of negative affect by sad movie 30
External Validity: Generalizing Your Experiment to the Outside • External Validity : Do your results apply to populations and situations that are different from those of your experiment? • Men to women? Labs to bars to sporting events? Young to old? • Generalization : (the process of) applying the results from an experiment to a different situation or population. – We want to take our results beyond the narrow confines of our specific experiment. 31
Statistical Validity • False alarms (Type I Error) • Too many analyses • Misses (Type II Error) • Too small of sample size Fire Alarm No Yes Silent Working Yikes! Miss: Type II Goes off False Alarm: Type I Working Fire 32
Statistical Validity • Strength of association – Magnitude of effect is not the same as statistical significance • Outliers, subgroups, and nonlinear relationships • (to be covered during correlational designs lecture) ( r = 0) N = 2500; r = .08) 33
Statistical Validity • Significance frequently depends on sample size • To get around this problem standardized measures of effect size (not sample size dependent) • Cohen’s guidelines for correlation coefficients • .10 = small/weak • .30 = medium • .50 = strong N = 2500 r =.08 p < .05 34
d=(M 1 -M 2 )/SD p A large value of d means that the treatment or association is strong. Anything over 1 is BIG 35
Internal Validity • Internal validity – The extent to which the observed effect is caused only by the experimental treatment condition; freedom from confounds • The ability to draw conclusions about a causal relationship from our data • In sum: Did your IV actually cause change that you observe in your DV, or was it something else? 36
Comprehension • Dr. Scott wants to assess the relations between negative affect and alcohol consumption. He administers a survey to a sample of college students asking about their emotional experience and the amount they drink. In this study, construct validity would
You've reached the end of your free preview.
Want to read all 54 pages?
- Fall '08
- external validity, internal validity, Confounding, • Association