Power Estimates Kopriva

1
Power Estimates(Kopriva Shaw,
1991)Difference Scores(Peter, Churchill
Brown, 1993)

• Christina Abbott

2
Articles

• Caution in the use of difference scores in
  consumer research
• Power estimates The effect of dependent variable
  reliability on the power of one-factor ANOVAs

3
Difference Scores

• Subtract one measure from another to create a
  construct.
• Example
• Expectations-PerceptionPerceived Usability.

4
Analysis

• Subjective Data
• Ordinal Numbers in Equation

5
Difference Scores

• Benefits
• Surveys are easy to develop
• Surveys are easy to administer
• Allows for alternate uses of the survey data.
• Capitalize on trends in data to illustrate points.

6
Difference Scores

• Correlation Problems
• Correlation
• Reliability issues
• Component score decreases
• Correlation increases
• Discriminant Validity
• Spurious Correlations
• Variance Restriction

7
Difference Scores

• Alternatives
• Direct Comparison Operationalizations
• Reframe research questions
• Author Recommendatoins
• Not Used
• Input to other statistics

8
Power Estimates

• Study quantified and tabulated the degree to
  which reliability affects the power of ANOVA
• Sample
• Significance
• True Score Effect Size

9
Power Estimates

• Introduction
• Results confirm that the impact.
• Low-level of confidence
• Lit Review
• Include details on disagreements

10
Power Estimates

• Kopriva and Shaw found
• Reliability has substantial impact on power
• sample sizes are lt 100 per group
• small or moderate effect size differences are
  expected
• Important in behavioral sciences

11
Takeaways

• Reliability affects power
• Alternative methods to using difference scores
  are recommended
• Proceed with caution when using Difference Scores
• Power Tables for Difference Scores

12
a .05, 2 groups µ1 µ2
13
a .05, 3 groups µ1 - µ2 µ2 - µ3
14
a .05, 3 groups µ1µ2 ltµ3