CHAPTER 5 Introduction to Experimental Research Chapter 5. Introduction to Experimental Research Chapter Objectives • Define a manipulated independent variable and identify examples that are situational, task, and instructional variables • Distinguish between experimental and control groups • Recognize the presence of confounding variables in an experiment and understand why confounding creates serious problems for interpreting the results of an experiment Chapter Objectives • Identify independent and dependent variables, given a brief description of any experiment • Distinguish between manipulated independent variables and those that are subject variables • Describe the interpretation problems that accompany the use of subject variables Chapter Objectives • Recognize the factors that can reduce the statistical conclusion validity of an experiment • Distinguish between the internal and external validity of a study • Describe the various ways in which an experiment’s external validity can be reduced The meaning of Experiment • Experiment: a systematic research study in which the investigator directly manipulates some factor, holds all other factors constant, and observes the results of the variation.(contrast with correlation) • “investigating the effect of X on Y” Essential Features of Experimental Research • Establishing independent variables (IVs) • Manipulated IVs : must have minimum of 2 levels • Situational: environmental features • Task: different problems to solve • Instructional: performing tasks in different ways • Experimental groups • given treatment • Research Example given a golf ball and told it was a “lucky” ball • Control groups • treatment withheld • Research Example given a golf ball and not told it was a “lucky” ball Essential Features of Experimental Research • Controlling extraneous variables: variables which are not of interest to the researcher but which might influence the behavior being studied if not controlled for properly. • Confounds • Any uncontrolled extraneous variable • Covaries with the IV; results could be due to IV or to confound • Distributed practice example Essential Features of Experimental Research • Measuring dependent variables (DVs) • DVs are any behaviors measured in an experiment • Review scales of measurement (Ch. 4)! • Problems: • Ceiling effects • task is too easy, all scores very high, disguising any differences • Floor effects • Task too difficult, all scores very low, disguising any differences • Solution: • Task of moderate difficulty, determined through pilot testing Manipulated versus Subject Variables • Subject Variables • Already-existing attributes of subjects in a study • Examples gender, age, personality characteristic • Anxiety example • As a manipulated variable induce different degrees of anxiety in participants • As a subject variable choose participants who have different degrees of their typical anxiety Manipulated versus Subject Variables • Cannot draw certain conclusions when using subject variables • With a manipulated IV • Assuming no confounds IV causes DV • With a subject IV • Groups may differ in several ways IV cannot be said to cause DV • All that can be said the groups differ from each other Manipulated versus Subject Variables • Using both manipulated and subject IVs • Bandura’s Bobo study (Box 5.2) • Manipulated type of exposure to violence • Subject gender Independent: Type of aggression, gender Extraneous: Emotional arousal, Proportion of aggressor to Bobo size Dependent: Aggression score The Validity of Experimental Research • Statistical conclusion validity • Proper statistical analyses and conclusions • Construct validity • Well-chosen and well-defined IVs and DVs • External validity: the degree to which research findings generalize beyond the specific context being studied (subject pools, the college sophomore problem, the male problem) • Internal validity- methodologically sound and confound-free (confidence that IV is directly responsible for results) Threats to Internal Validity • Studies extending over time (may have pretests) • Example: Test-anxiety reduction History – courses become Pass/Fail Maturation Regression to the mean Testing and instrumentation- practice effects, change in measurement instruments/coders • Importance of using a control group • • • • Threats to Internal Validity • Participant problems • Subject selection: method/criterion for participation produces a confound • The Brady study ulcers in executive monkeys • Attrition • Loss of data • Subject selection problem Summary • Experimental research involves independent and dependent variable, in an effort to test the effects of the IV on the DV. • We attempt to control for confounding variables to increase the internal validity of our study. • We must consider other possible threats to internal validity as they pertain to our study. • Once we identify IVs, DVs, and threats to validity, we design a study to control those threats.