Psych 300

  1. Randomized Experiments
    • 1. high degree of control
    • 2. usually low contextual realism
    • 3. allow to test cause/effect relations - causal relations
  2. Between (independent groups) design
    different individuals serve as Ps in different conditions
  3. Within subjects (repeated measures) design
    • the same individuals serve as Ps in multiple conditions
    • (note: covered in chp 8)
  4. Logic of randomized experiments
    while other variables are held constant, the researcher manipulates 1+ variables and measures its effect on 1+ other variables

    • Conditions:
    • treatment/ experimental
    • control/ comparison
    • ex: drug grp vs. placebo grp
    • Variables
    • - IVs/DVs
    • - control- variables held constant across all conditions (purpose- find causality)
  5. ....
    If the only difference between the conditions is the treatment any difference between the conditions in terms of the DV must be caused by the treatment
  6. 3 criteria of Causation*
    • 1. the IV and DV have to covary
    • 2. the IV precedes the DV in time
    • 3. There are no other possible explanations for difference among conditions in terms of the DV
  7. Rival Hypotheses
    alternative explanations for the results
  8. confounds
    the source of the rival hyptothesis
  9. Internal validity
    validity of statements about whether the IV causes changes in the DV
  10. Some threats to Internal validity
    • - selection
    • - experimenter bias
    • - history
    • - others (specific to given experiments)
    • also...order effects
  11. Selection confounds
    when group differences in the DV are due to differences btwn people who made up the groups
  12. Random Assignment
    1. assign Ps to treatment and control groups such that all Ps have the same probability of being placed in a given group

    2. Distributes impact of uncontrolled variables evenly across all groups

    3. what about matching "matched groups", matching on; known correlates, purposeful matching to both conditions

    4. note- not the same as random sampling (taking a sample from a population in a random way to get a representative of a population)
  13. Experimenter bias (experimenter effects)
    when the experimenter (unknowingly) does something to bias results

    ex: only look for certain/specific results may act differently to each of the groups

    to prevent- keep them blind to conditions or hypotheses "double blind"
  14. History
    when there is an event (unrelated to IV) that occurs during the study and affects the DV

    • time 1time 2
    • intervention
    • stress __________________________ stress
    • DV IV
  15. maturation
    gradual changes in Ps over the course of the study

    ex: normal development (language performance) learning process in course of a semester
  16. Other confounds
    • Many possibilities
    • specific to certain study/studies
    • ex:different experimenter are running different conditions- might have inconsistencies
    • run all of 1 condition first then run 2nd condition- different levels of stress at different times
    • 1 condition watches a video on topic for 1 hr; other condition does nothing at all - content of video could change opinion

    can be prevented by controlling variables anything that could change the IV should be controlled
  17. Order effects
    when doing tasks/conditions in a certain order influences the results

    - more common w/repeated measure designs
  18. Repeated Measures Example
    (Hosking and Young 2009)
    Experiment on the effects of text messaging while driving
    • method
    • 1. 20 Ps 18-20 yrs old
    • 2. drive in simulator w/8 test events
    • 3. conditions
    • - received and sent texts on cue during test event
    • - not text message during test events
    • Results
    • in text conditions
    • - amount of time drivers spent not looking @ road was 4005 higher
    • - 140% more missed signals
  19. Types of Order Effects*
    • 1. Practice effect: performance is better over time due to practice
    • 2. Fatigue effect: performance is worse over time due to boredom, fatigue, ect.
    • (ex/vigilance task- sustained attention)
    • 3. contrast effect: perform in the 2nd condition is changed b/c the conditions are contrasted with each other
    • ex: IV crime severity
    • DV perceived guilt

    to eliminate - COUNTERBALANCE- randomly assign Ps to all possible orders of the conditions
Card Set
Psych 300
Section 3 notes