1. INTRO TO EXPERIMENTAL RESEARCH, continued Lawrence R. Gordon Psychology Research Methods I

2. “TRUE” EXPERIMENTS 4 Investigate the “effect(s) of X(s) on Y(s)” 4 At least one IV is manipulated (X) with two or more “levels” 4 All extraneous variables are controlled 4 At least one DV is measured (Y) 4 SIMPLEST EXPERIMENT: one IV with two levels and one measured DV

3. Independent Variables 4 Two or more “levels” (conditions, treatments) 4 Types: –Situational variables –Task variables –Instructional variables –“Control groups” - not all have one 4 Expression: –Manipulated variables (above) - assigned to participants –Subject variables - participants selected for (‘ex post facto’ study if none manipulated).

4. Extraneous Variables 4 Variables NOT of interest to our research question 4 But if they covary with the IV, become a “confound” or an “alternative” or “rival” explanation for effect on DV 4 Control procedures are all about this -- next chapter 4 Goodwin p. 152 has excellent tables for this; try to “fix” the example shown!

5. Dependent Variables 4 “...some characteristic of behavior or reported experience” (Woodworth, 1938) 4 Measured 4 Considerations –Operational definition of construct –Reliability and validity of measurement

6. A Simple Experiment: “Time Flies” 4 EXAMPLE: “Time flies when you’re having fun” 4 Hypothesis: IF one is “having more fun”, THEN time will seem to pass more quickly 4 Design: IV: 100 persons randomly assigned to two groups: –1: “Having more fun” –2: “Having less fun” –Procedure: manipulation of cartoon captions DV: Estimate of a standard 10 minute interval

7. A Simple Experiment, cont. 4 Results (cont.) –Group 1 = “More fun” Mean = 8.60, SD = 2.72, N = 50 –Group 2 = “Less fun” Mean = 12.48, SD = 3.35, N = 50 –Quickie summary of results: the “More fun” group gave shorter estimates of the 10-minute interval, on average, than the “Less fun” group. 4 Can you think of any possible confounds, or alternative explanations for this effect?

8. Research Validity 4 How do we know we’re answering the question we asked? 4 Statistical conclusion validity 4 Construct validity 4 External validity 4 Internal validity

9. External Validity 4 Generalization to –other populations? who –Other environments? where –Other occasions? When 4 Affects the scope of our inference; usually addressed in “discussion” of the research 4 Often the target when adding additional IVs

10. Internal Validity 4 Does X indeed cause Y? 4 Analysis of possible confounds 4 There are special “threats” to internal validity that the “design” of research attempts to address –Affecting Pre-Post research –Concerning our participants

11. Threats to Internal Validity 4 Affecting Pre-Post research –Pre X Y Post X –History and maturation –Regression to the mean –Testing and instrumentation 4 Major solutions –Eliminate Pre-X (“Posttest only design”) –Add control condition without Y: Pre X Y Post X Pre X Post X

12. Threats to Internal Validity 4 Concerning our participants –Subject selection (are groups equivalent?) –Section by Other interactions (Sel  History) –Attrition (“mortality”)

13. Wrap up: The Bower Experiment 4 IV(s)? 4 DV(s)? 4 EVs? 4 Results 4 Problems?

14. Bower: Histograms (F’02)

15. BOWER: Descriptives (F’02)

16. BOWER: Inferential (a peek) (F’02) “ Significant difference” if “Sig (2-tailed)” is <.05

17. CORRELATION: The Problem 4 Are two variables related? –Does one increase as the other increases? e. g. skills and income –Does one decrease as the other increases? e. g. health problems and nutrition 4 How can we get a numerical measure of the degree of relationship? SPSS, for now...

18. Correlation: A Quick Introduction 4 Descriptive: “Pearson product-moment correlation coefficient,” r 4 Values: -1 __________ 0 _________ +1

19. Correlation Coefficient 4 A measure of degree of relationship. 4 Sign refers to direction. 4 Based on covariance –Measure of degree to which large scores go with large scores, and small scores with small scores

20. Correlation: A Quick Introduction 4 Descriptive: “Pearson product-moment correlation coefficient,” r 4 Values: -1 __________ 0 _________ +1 4 Visualization: SCATTERPLOTS

21. {X = 6, Y = 11} N=19 data pairs

22. What Does the Scatterplot Show? 4 As smoking increases, so does coronary heart disease mortality. 4 Relationship looks strong 4 Not all data points on line. –These are “residuals” or “errors of prediction”

23. Correlation: A Quick Introduction 4 Descriptive: “Pearson product-moment correlation coefficient,” r 4 Values: -1 __________ 0 _________ +1 4 Visualization: SCATTERPLOTS 4 Inferential -- is the computed r unlikely from a population with a correlation of 0? 4 EXAMPLES -- from Bower & Class Survey 

24. BOWER, revisited, F’02 4 Are Correct vs. Incorrect recall related? 4 That is, do they predict one another “really”? 4 YES - why?  4 Bower: Correct vs. Incorrect

25. CLASS SURVEY 2002 4 Are the scales measuring Need for Cognition and Concern for Future Consequences related? 4 That is, do they predict one another “really”? 4 YES - why?  CFC Scale by NC Scale