272 U4

  1. What are the 3 reasons for assessing IOA?
    Consistency of measurement

    Minimizes observer bias

    reflects the quality of operational definition
  2. What is the difference between agreement between observers and accuracy?  Is it possible to have one but not the
    other?  Explain.
    Agreement - how well data from independent observers correspond

    Accuracy - How well the data reflects what actually happens

    Yes it is possible, two people can make mistakes in the same way therefore you can have agreement but not accuracy

    definition could be too broad, thus you have accuracy but not agreement
  3. What is the difference between and reliability and validity? Can you have one
    without the other? Explain.
    Reliability - consistent measurement of the construct

    Validity - measurement reflects the construct

    • yes. you can measure something reliably and it not reflect the construct
    • and you can measure a construct inaccurately
  4. List and describe the 3 methods for calculating
    interobserver agreement that were compared?
    • Frequency Ratio - Compares totals of observers
    • Can be used for frequency, intervals of Bx, or
    • duration
    • Typically used when measuring free-operant bx

    • Formula
    • Small freq/large freq x 100

    • o Point by point agreement - Assess agreement on each instance of BX
    • Use when there are discrete opportunities for BX (e.g., present-absent, correct-incorrect)

    Formula - PBP IOA = A/(A + D) X 100

    • o  Frequency within interval recording
    • Formula - Subtract # of disagreements form the total
    • number of intervals

    • For each interval that was scored as a
    • disagreement divide the small frequency by the larger frequency

    • (Total # of disagreements – total intervals) +
    • (sum of all disagreement when small freq is divided by large freq)/ total
    • intervals
  5. In general, describe results of the study.
    (3 ways to calc IOA) Repp et al.
    You must select the appropriate method for your study

    Exact agreement was the most conservative method.
  6. 1.    
    What can be done to reduce the likelihood that a high
    percentage of agreement between two observers is simply the result of chance
    given that the level of chance agreement changes as the percent of intervals in
    which the behavior is recorded changes?
    If there is a slight effect the acceptable agreement must be higher to compensate for errors in recording
  7. What are potential sources of bias discussed by Kazdin?

    Observer Drift

    Observer expectancies & feedback

    Complexity of observations
  8. What are the two purposes of baseline observations? What is the purpose of the
    intervention phase?
    The descriptive function - provides the existing level of performance or the extent to which the client engages in the bx

    The predictive function - allows for the prediction of future performance without intervention

    Intervention - describe current performance and predict future performance if future conditions were unchanged
  9. What is the purpose of the second A phase (i.e., the second baseline phase)?
    • describe current performance
    • Predict future performance
    • test the prediction of the prior phase and baseline or the first A condition
  10. Describe different procedures that can be implemented to reverse behavior.
    withdraw the intervention

    make the intervention ineffective or less effective (lower quality of praise, add a delay b/t praise and bx)

    continue the consequences but alter associated with the consequence
  11. Under what conditions is noncontingent reinforcement more likely to lead to a
    return to baseline levels of performance and why is this the case?
    when praise was previously contingent on a specific behavior 

    because it is not praise itself but the relationship between praise and the bx
  12. In what situations would it be appropriate to start with intervention rather
    than baseline?
    when the base rate of Bx is zero or has never been performed
  13. Explain how you can still meet the requirements of a reversal design without
    returning to baseline.
    by testing the effectiveness of two interventions showing control by one rather than the other at rates significantly different from baseline
  14. Describe potential problems that should be considered when using a reversal design.
    • ethics of a reversal
    • if bx change agents are willing to reverse
    • if bx change is reversable
    • if the client is willing to reverse
Card Set
272 U4
272 U4