Question 1: Which statistic is the best expression of error variance in an ANOVA?
Question 2: The magnitude of any sum of square value reflects which of the following?
- The treatment effect
- The impact of the dependent variable
- The number of scores included
- The effect size
Question 3: With fewer degrees of freedom and larger critical values to exceed, how can the dependent samples t be more powerful than the independent t?
- The larger critical value makes it easier to find statistical significance.
- The dependent groups tests tend to have larger treatment effects.
- The dependent groups tests can be repeated without affecting type I error.
- The larger critical value is offset by lower amounts of error variance.
Question 4: In an analysis of significant differences, any variance in scores not related to the independent variable is represented where?
- In any ancillary analysis that may be conducted
- Error variance
- In the differences between groups
- As a portion of the treatment effect
Question 5: What measure of data variability did Fisher rely on for ANOVA?
- The F ratio
- The modified standard score
- The squared difference between particular values and the relevant mean
- The standard error of the difference between multiple groups
Question 6. 6. The treatment effect is a component of which of the following in the within-subjects F?
- The measure-to-measure differences
- The subject-to-subject differences
- The measure times subjects differences
- The residual differences
Question 7: The validity of results from independent t or one-way ANOVA rests on which assumption?
- The dependent variable is at least ratio scale.
- All groups receive the same level of the treatment.
- The groups are equivalent before the treatment.
- The independent variable is at least ordinal scale.
Question 8: When the dependent groups tests use matched pairs, what is it that is matched?
- Subjects are matched on the degree of error variance each one manifests.
- Subjects are matched so that each pair receives the same treatment.
- Subjects are matched on variables other than the IV that affect the DV.
- Subjects are matched so that there are no differences in the level of the DV.
Question 9: Why is matching uncommon when using the within-subjects F?
- With several measures matching is usually unnecessary.
- Multiple before/after t-tests will provide the same data.
- Matching won't control the error variance in more than two subjects.
- It's difficult to find enough subjects that have the same characteristics.
Question 10: Repeated measures designs reduce error variance as long as the scores are correlated.