1. a. Why might a researcher standardize scores?
b. You know that the mean of a distribution of attitude measures is 24.3, and its standard deviation is 4.7. What is the Z score associated with an attitude score of 27.9?
2. a. A researcher wants to compare e-learning to classroom instruction in principles of life insurance. Fifty trainees are assigned randomly to each type of training.
What is the researcher's null hypothesis?
b. What statistical test would you use, and why?
c. Why use a standard error rather than a standard deviation in your test?
d. The researcher wants to use a Type I error probability of p < .05 to test this null hypothesis. Would you recommend a one-tailed or a two-tailed test? Why?
e. Assuming that classroom instruction yields a mean score that is statistically significantly higher than the mean score for the e-learning group, what does that mean?
3. a. A researcher in Hong Kong computes the correlation between the percentage of employee turnover and the local unemployment rate (also expressed as a percentage) over a 20-month period. She obtains a value of minus 0.57. What type of correlation did she most likely compute?
b. Test the value of -0.57 for statistical significance from zero at p <.01. What critical value did you use?
c. What is your interpretation of the value of 0.57?
d. The data were obtained at the level of the SAR. Can the researcher assume that the correlation of -0.57 applies to her own employer?
4. a. A researcher wants to correlate gender with first-year earnings after graduation from college. What type of correlation coefficient would you recommend, and why?
b. The researcher obtains a correlation of 0.198 in a sample of 102 recent graduates. Test it for significance at p < .05. What critical value did you use? c. Interpret the statistical and practical significance of your result.
5. a. A researcher obtains the following correlations between customer satisfaction and retail sales in a sample of 75 customers at a retail store in Kowloon, and 75 other customers from the same-brand store in Central: 0.78 and .91. Are these values significantly different from each other at p < .05?
b. What is the average value of the correlations that the researcher obtained?
6. a. How is reliability different from validity?
b. If a researcher was not too concerned about errors of measurement due to different samples of items, but very concerned about errors of measurement due to the effects of time, what type of reliability coefficient would you recommend that she use?
c. You use split-half reliability to assess the reliability of a 70-item test, and obtain a value of .77. What would the reliability be if the test were twice as long?
7. a. Your friend just received a score of 360 from a Dental-School admission test. The test is scored so that it yields a mean score of 300 and a standard deviation of 75 points. You know from previous research that the reliability of the test is .94. He asks you for advice about whether he should retake the test. Compute a 95 percent confidence interval for his test score, and use it to advise him about his future prospects.
b. You have just read that the Dental-School admission test described above yielded a predictive validity coefficient of .34 with supervisory ratings of potential after completion of the first year at dental school. The reliability of these ratings is 0.67. What is the estimated "true validity" of the Dental-School admission test?
8. a. In regression analysis, how is the total sums of squares (Y - Y-bar) useful to an analyst?
b. What does R2 tell the analyst?
c. What effect does the regression (Y- Y-hat) have on the sums of squares for error (SSE)?
d. In two-variable regression, how does one tell if it is worth it to compute a regression line?
9. a. Your boss chooses only extremely poor (or extremely good) performers to participate in a training program. How are the outcomes of his decision likely to be affected by the phenomenon known as "regression toward the mean"?
b. In another department, a boss allows all employees (that is, a full range of performers) to attend the training. How are the outcomes of his decision likely to be affected by the same phenomenon?
10. You observe a correlation of -0.46 between the use of a child-care program and employee turnover. However, you believe that the attitudes of the immediate supervisor might affect this relationship.
Assuming that supervisory attitudes correlate .27 with use of the child-care program, and -0.56 with employee turnover, partial out the effects of supervisory attitudes on the relationship between the use of a child-care program and employee turnover. What is the value of the partial correlation?