Errata for Data Analysis for the Behavioral Sciences Using SPSS

Errata for Data Analysis for the Behavioral Sciences Using SPSS

Chapter 1

Chapter 1 Solutions

P. 594, 1.12. b) Ordinal

Chapter 2

P. 26, third paragraph: In Table 2.5, the stem represents the one-hundreds place…

P. 36, Example 2.3: You can also use Frequencies to find any percentile that you want.

P. 39, Second paragraph: It is instructive to examine the boxplots for both self-concept in twelfth grade and expected income at age 30 that appear in Figures 2.16 and 2.17, respectively, and to compare these boxplots to the corresponding histograms of these variables that appear in Figures 2.12 and 2.13.

Chapter 2 Exercises

P. 55, 2.18. b) LAYOFFSP

P. 55, 2.18. c) LAYOFFSP

P. 56, 2.20. a) Create a stem-and-leaf plot for this distribution. Use this graph to answer questions b) – f).

P. 56, 2.20. b) Approximately, what is the highest tenth grade social studies achievement score among student in the NELS data set?

P. 56, 2.20. d) Approximately, what is the lowest tenth grade social studies achievement score among student in the NELS data set?

P. 56, 2.20. g) What is the exact value of the highest tenth grade social studies achievement score among student in the NELS data set?

P. 56, 2.20. h) Exactly how many students earned the highest score found in part h)?

P. 56, 2.20. i) What is the exact value of the lowest tenth grade social studies achievement score among student in the NELS data set?

P. 56, 2.20. j) What is the exact value (as calculated by SPSS) of the 15th percentile?

Chapter 2 Solutions

P. 598, 2.19. j) (ii)

P. 599, 2.20. g) 72.89

P. 599, 2.20. h) 13

P. 599, 2.20. i) 32.91

P. 599, 2.20. j) 46.16

P. 605, 2.25. i) the location of its center is 95, and it is not so widely dispersed in that the middle 50 percent of the data are in a 3 point range. However, because two of the schools have attendance rates in the 40’s, it has a large range.

P 605, 2.26. b) Between 70.2 and 74.8 inclusive.

Chapter 3

P 75, Example 3.7. (Recall that these values are approximately 56.83 and 26.64, respectively.)

Chapter 3 Exercises

P. 89, 3.9. l) 40th percentile

Chapter 3 Solutions

P. 612, 3.5 a) The skewness ratio for the South is .

P. 612, 3.9 The first table was obtained using the SPSS Frequencies procedure. Replace it with the following, which contains the 40th percentile:

P. 613, 3.10 d) (iii). …the distribution is severely skewed.

P. 614, 3.11 Because both distributions are severely skewed, robust statistics are recommended.

P. 614, 3.11 a) According to the median, tenth grade math achievement is higher for students whose families owned a computer when they were in eighth grade (Median = 59.01) than for those whose families did not (Median = 55.76). The means corroborate this conclusion.

P. 614, 3.11 b) According to the interquartile range, tenth grade math achievement is slightly more variable for students whose families owned a computer when they were in eighth grade (IQR = 11.71) than for those whose families did not (IQR = 10.45). However, the standard deviation contradicts this conclusion. For those whose families owned computers, SD = 7.64. For those whose families did not, SD = 7.69.

P. 614, 3.12 e) For rural children the distribution is negatively skewed, although less so,because the box is not evenly centered between the whiskers.

P. 620, 3.30 a) The mean is higher. Because fewer than half of the players earn more than the average, the median is lower than the mean.

Chapter 4

Chapter 4 Exercises

P. 120, 4.20 The mean grade level (GRADE) for female students…

P. 122, 4.27 a) According to percentile ranks? Follow the instructions below to obtain separate frequency distributions for males and females.

Chapter 5

P. 143, Remark. Notice the outlier in the distribution of male science achievement scores in Figure 5.8. checking further, we find that this outlier score belongs to case number 194. (The series of steps needed to obtain this case number is the focus of Exercise 5.29.

Chapter 5 Exercises

P. 158, 5.11 The Interactive scatterplots below…Add at the end: For both of the voting variables, the coding was 0 = Not guilty and 1 = guilty.

P. 159, 5.12. b) What percentage of senators voted not guilty on perjury?

P. 162, 5.16. Nevertheless, the investigator was able to make reasonably accurate predictions of a person’s manual dexterity score on the basis of his or her age.

P. 163, 5.25. e) +.82

P. 164, 5.29. In Example 5.5….

P. 165, 5.30. For each of the following questions about the students in the NELS data set, …

Chapter 5 Solutions

P. 629, 5.26. f) Among the students in the NELS data set, a higher socioeconomic status was associated with a higher science achievement, r = .32.

P. 631, 5.30. b) The contingency table provides the numbers to quantify the relationship: 85 persent, 86 percent, and 87 percent of the urban, suburban, and rural dwellers, respectively, did not smoke cigarettes.

P. 633, 5.30. f) By treating SES as interval and URBAN as ordinal with only a few categories, we compare an appropriate measure of central tendency of SES for the three levels of urbanicity. Because none of the three distributions is severely skewed, we use the means. Students in the NELS from urban environments enjoy the highest overall level of SES (Mean = 20.33), followed by those from a suburban environment (Mean = 19.23). Those from a rural environment had the lowest level of SES (Mean = 15.94).

P. 635, 5.32. a) The Pearson correlation between SES and COMPUTER for the students in the NELS data set is r = .35.

Chapter 6

P. 181, 9. This is the value of the slope of the regression equation as explained in part (4).

Chapter 6 Exercises

P. 189, 6.10. In the blurb above the question, eliminate the last sentence: Recall that in Exercise 4.21 we created the linear transformation Age = Grade + 5, which converts the students grade level to their age.

Chapter 6 Solutions

P. 640, 6.21. Scatterplots may be used to explain the solution.

P. 640, 6.22. a)

Chapter 7

P. 200, middle of the page: (The 4 was not clearly printed).

P. 201, middle of the page: At each stage of the selection process, the probability of selecting a student who never smoked is .

Chapter 7 Exercises

P. 205, 7.2. A crosstabulation of the top 26 schools by region and whether or not they belong to the Ivy League is given on page 206.

Chapter 8

P. 223, bottom of page: For this example, this set of question marks should be replaced by .975, 18.4, and 6.92 so that we now have the following in the Numerical Expression box: IDF.Normal(.975, 0, 1).

Chapter 8 Exercises

P. 228, 8.24. Then, find the actual percentile rank of 72.

Chapter 8 Solutions

P. 646, 8.9. e) .985

P. 646, 8.9. f) .0846

P. 646, 8.9. g) .10

P. 646, 8.9. h) .985

P. 647, 8.24. The actual percentile rank of 72 is 94.2.

Chapter 9

P. 234, near end. The 75 values may be entered manually, or for the more sophisticated reader, may be entered using the syntax window with the program and instructions given in Appendix B.

P. 243, Figure 9.5: m = 8, s = 4.

P. 244, Figure 9.6: = 8.

Chapter 9 Exercises

P. 246, 9.3. d) If we were to select a single sample of size N = 8, what is the probability that the mean of this sample will be two or more points higher than the actual population mean?

P. 246, 9.4. d) If we were to select a single sample of size N = 30 at random from this population of scores, what is the probability that the mean for this sample will be more than two standard errors away from the actual population mean?

P. 246, 9.5. d) If we were to select a single sample of size N = 50 at random from this population of scores, what is the probability that the mean for this sample will be more than two standard errors away from the actual population mean?

Chapter 10

P. 267, First sentence in the section of decreasing the standard error of the mean: If we hold a and d to be the same as in Figure 10.9, but we decrease , the standard error of the means, to , then…Said differently, an increase in the standard error of the means…, will decrease the power of the test.

Chapter 10 Exercises

P. 270, 10.5. f) What can you conclude based on your results in part (e)?

Chapter 10 Solutions

P. 651, 10.9. …and all of these values are larger than 100.

P 651, 10.11. p = .03.

Chapter 11

P. 278, middle of page: Upper limit = … = … = … = 114 + 4.38 = 118.38

P. 287, middle of page: 1. on average, college-bound males from the South who have always been on grade level differ in the number of years of math taken in high school from females from the South?

P. 306, bottom of page: Results of this t-test are given on page 307.

P. 311, there is an omission from the flow chard. The box, “For Levene’s, is p>a?” does not have a “yes” line. If yes, homogeneity of variances is met and we use the equal variances (first) line of the test.

Chapter 11 Exercises

P. 316, 11.1. Delete part e).

P. 317, 11.2. In part a), the question is repeated twice. Delete one instance.

P. 317, 11.3. a) ..If so, is the normality assumption violated? Which is more appropriate, the Equal variances assumed line, or the Equal variances not assumed line?

P. 319, 11.3. Delete part f).

P. 319, 11.3. h) Are the results of parts (e) and (g) consistent? Explain.

P. 319, 11.4. In the blurb, delete the line: Assume that the underlying assumptions have been met.

P. 326, 11.17. In the blurb, replace math comprehension with IQ or intellectual ability.

P. 332, 11.24. In the blurb, delete “from the Northeast”.

Chapter 11 Solutions

P. 652, 11.1. Delete part e)

P. 652, 11.1. Delete part f)

P. 652, 11.1. Change g) to e) and change the answer to: One cannot use the confidence interval given in the SPSS output to answer a question associated with a directional alternative hypothesis.

P. 652, 11.3. a) An independent samples t-test is selected because there are two groups (males and females) and the scores within the groups are independent (not related). Because the sample size is large (N1 = N2 = 80 > 30), the t-test is robust to violations of the normality assumption, so that the validity of the results of the t-test would probably not be compromised even if the normality assumption were not met. Because the two sample sizes are equal, the t-test where equal variances are assumed may be used.

P. 652, 11.3. b) Based on the position of the medians in the boxplots, it is anticipated that no statistically significant differences will be detected between males and females.

P. 652, 11.3. Delete f).

P. 653, 11.4. b) However, the validity of the results of the t-test would probably not be compromised by a violation of this assumption because for both males and females, N > 30.

P. 653, 11.4. e) the perception of stress due to finances by females (M = 3.39, SD = 1.24) is statistically significantly higher than that by males ( M = 2.90, SD = 1.35), ….

P. 653, 11.4. f) …A small to moderate effect size, according to Cohen’s scale….

P. 654, 11.15. … p < .001.

P. 655, 11.17. b) Add: The confidence interval gives a range of plausible values for the difference between the population mean and 100.

P. 655, 11.18. d) Add: Alternatively, can convert to a confidence interval for m by adding 75 to the endpoints. The 95% CI for m is (74.64, 80.62). Because 75 is contained in this interval, the average e reading comprehension score for children diagnosed with learning disabilities in the city is not found to be different from 75.

P. 660, 11.29. e) p = .23.

Chapter 12

P. 355 Solution 1: Table 6 in Appendix C does not list dfw = 27 exactly, and so we will use the more conservative values associated with dfw = 28. In doing so, we find that 8.49 exceeds 7.19, the critical value associated with p = .001 when K = 3. Hence, we conclude that m1 ≠ m2 and p < .001.

P. 355 Solution 2: As in the previous case, 6.38 exceeds 4.57, the critical value associated with p = .01 when K = 3 and dfw = 28. Hence, we conclude that m1 ≠ m3 and p < .01.

P. 355 Solution 3: The value of 2.13 is less than 2.29, the critical value associated with p = .10 when K = 3 and dfw = 28. Hence, we conclude that m2 = m3 and p > .10.

Chapter 12 Exercises

P. 368, 12.1. f) …Tukey HSD Test…

P. 368, 12.2. Delete from blurb: Assume that the underlying assumptions are met. Insert a) Evaluate the underlying assumptions or indicate why it is not necessary to do so.

P. 369, 12.3. Delete the extra “,0” for the sum of squares values in the table. The SSBetween value should be 2,240 and the SSWithin value should be 26, 208.

P. 369, 12.4. c) Evaluate the underlying assumptions or indicate why it is not necessary to do so.

P. 369, 12.4. f) If appropriate, conduct post-hoc analyses using both LSD and Tukey HSD approaches and interpret results. If not, indicate why not.

P. 369, 12.5. f) If appropriate, conduct post-hoc analyses using both LSD and Tukey HSD approaches and interpret results. If not, indicate why not.

Chapter 12 Solutions

P. 668, 12.2. Insert: a) The test is robust to violations of the normality assumption because N > 30 for each group and it is robust to violations of the homogeneity of variances assumption because the sample sizes are equal for all groups.