Research Methods Final Exam

Ch 7: The experimental research strategy

  1. Experimental and Control Conditions
  2. There must be at least two conditions in every experiment.
  3. Another name for control conditions = comparison conditions.
  4. Control Conditions: purpose is to rule out competing explanations to test hypotheses.
  5. The equivalence in experience between the experimental and control conditions allows you to rule out nonspecific treatment effects.That is, if the experiences of the participants in the experimental and comparison conditions differ in any except for presence/absence/levels of the independent variables, alternate explanations for the variance in the DV is possible.
  6. Characteristics of a Good Manipulation (4 ½)
  7. Construct validity: the manipulation of the IV should actually manipulate the IV. For example, if you’re trying to manipulate level of cognitive information processing, you need to make sure that what you do to manipulate CIP (e.g., by introducting a test and reward into the experimental task) actually manipulates the IV and only the. One can check this by:
  8. Manipulation check: used to assess the validity of the manipulation. Also can test divergent validity --> did the manipulation affect the IV and ONLY the IV?
  9. Salience: a good manipulation must be strong for several reasons. One of which is so the IV is noticed. If participants miss the presentation of the IV, then your IV will have no affect.
  10. Strength: a manipulation must be strong enough to reliably affect the DV in a measureable way. However, mundane realism comes into play here; if the levels of the IV are too extreme, the experiment may lack realism and not provide reliable results. Also, extremem manipulations may be unethical in certain circumstances.
  11. Reliability: everytime a manipulation is applied, it is applied in the same way. Some ways to obtain reliability is to use scripts or automation.
  12. Multiple stimuli: one way to improve validity of your manipulation is to use multiple stimuli. For example, if I am using pictures of human faces to manipulate level of unattractiveness, I should use several different unattractive faces rather than just one face. Using several faces rules out the possibilty that the variance in the DV is associated with special characteristics present in a single stimuli.
  13. Between vs. Within Designs
  14. Between Subjects Design: when each condition consists of a unique set of Ss. No subject is in more than one condition.
  15. Must use RANDOM ASSIGNMENT OF Ss to GROUPS – randomly distributes participant characteristics among conditions; therefore, conditions contain (theoretically) equal proportions of subject variables that could differentially affect the DV.
  16. Can use matched random assignment to control for sources of extraneous DV variance.
  17. Match participants on certain characteristics (either categorical or continuous) and then assign one of each pair to random condition.
  18. Within Subjects Design: when subjects participate in more than one condition.