CHAPTER 11
Questions 1–18: You are asked to conduct a study to determine whether there is an association between consumption of milk proteins and levels of serum antibodies in children with autism. You randomly select 50 children previously enrolled in an autism clinical trial, and they are the sample for the study. You select an alpha of 0.05 and a power of 0.80.
1. Your study measures consumption of milk proteins as a yes/no question. It measures serum antibodies as a present/not present question. What level are the two variables?
2. What analysis method do you propose based on this information?
3. Your research partner believes it would be better to measure milk protein consumption as a low/moderate/high question and to quantify the serum antibody level by using the actual amount present. If you take this approach, what analysis method do you propose?
4. The biostatistician in your department recommends that you change your milk protein measurement to one that is the number of servings per day. You continue to use the actual quantity of the serum antibodies. What analysis method do you recommend?
5. Write an appropriate null hypothesis for this study.
6. Write an appropriate alternative hypothesis for this study.
7. Is your sample size large enough to conduct a correlation test? Is it large enough to assume a normal distribution and homoscedasticity?
8. You decide to utilize the measurement variables as recommended by the biostatistician and conduct a Pearson’s correlation coefficient test. You determine r = 0.6. What must your p-value be for this to be statistically significant?
9. Your p-value is 0.08. Is this significant?
10. What do you conclude?
11. Suppose that you were able to refine your measurement tools and repeat the study. This time you determine you had an r of 0.6 with a p-value of 0.049. What would you conclude?
12. This information would let you know that your original conclusion was actually what type of error?
13. What is the strength of the relationship between consumption of milk proteins and serum antibodies?
14. Is the relationship positive or negative? Interpret this in plain English.
15. In this study, what percentage of variance in serum antibody levels is explained by the consumption of milk proteins in children with autism?
16. How much of the variance is explained by other variables?
17. Is this clinically important?
18. If you were going to use this effect size to determine your sample size for another study, would you expect to need a large or a small sample?
Questions 19–24: You develop a screening test to be used for children with autism to detect serum antibodies from milk protein consumption and have the following 2 × 2 table.
Calculate the following values:
19. Sensitivity:
20. Specificity:
21. Positive predictive value:
22. Negative predictive value:
23. Prevalence:
24. If early treatment helps, is this a good screen?
Questions 25–27: You are asked to complete a study for a small school district that is trying to keep as many students at age-appropriate grade levels as possible. You have a measure of grade level and age for the students, as well as the following statistical programming output from a randomized independent sample:
25. What is your sample size?
26. What is the appropriate correlation coefficient, and why?
27. Are age and grade significantly correlated?
Questions 28–30: A nurse researcher conducts a study to determine if taking a new fertility drug is associated with multiple-fetus pregnancies. Her sample includes 500 women who are pregnant in her fertility practice. She selects an alpha of 0.05 and a power of 0.80.
28. If taking the new drug is measured as yes/no, what level of measurement is it?
29. If a multiple-fetus pregnancy is also measured as yes/no, what correlation test is appropriate, and why?
30. The researcher reports that p = 0.044. What conclusion should the researcher make about the null hypothesis? Why?
Question 31–39: A researcher conducts a study to determine if there is an association between time spent in solitary confinement and depression rates in 120 male prisoners. The alpha is 0.05, and the power is 0.80.
31. If time spent in solitary confinement is measured in total hours, what level of measurement is this variable?
32. If depression is measured on a scale with values from 1 to 10, what level of measurement is the variable?
33. What would be the appropriate correlation test, and why?
34. The study reports that r = 0.4. What does this mean in plain English?
35. How much of the variance in depression is explained by hours in solitary confinement?
36. If p = 0.07, is the correlation significant? Why or why not?
37. If the hypothesis decision is incorrect, what type of error could it be?
38. If instead the study measured time in solitary confinement as none, infrequent, or regular, what level of measurement would it be?
39. Would this change the correlation test you would recommend? Why or why not?
Research Application Article
Look at the following research article to see how the statistical techniques you have already learned are used in practice:
Watson, J., Kinstler, A., Vidonish III, W. P., Wagner, M., Li, L., Davis, K. G., . . . Daraiseh, N. M. (2015). Impact of noise on nurses in pediatric intensive care units. American Journal of Critical Care, 24(5), 377–384. doi:10.4037/ajcc2015260
1. What was the purpose of this study?
2. What type of sample was collected?
3. What are the noise-level recommendations from the Environmental Protection Agency (EPA) and the World Health Organization (WHO)?
4. Why does the noise level in the workplace matter?
5. The sample included nurses from what units?
6. What nurses from these units were not included in the study?
7. How was noise level measured? What level of measurement is this variable?
8. How was heart rate measured? What level of measurement is this variable?
9. How was stress measured? What level of measurement is this variable?
10. In addition to the noise-level measurement, the observer log recorded what information about the noise? What level of measurement was this additional information?
11. Look at Table 2 in the article. Which unit, on average, was the loudest?
12. Look at Table 2 in the article. Which unit was the quietest on average? Was this unit within the recommended EPA guidelines for workplace sound levels?
13. Look at Table 2 in the article. What percentage of the time was the noise level of the Cardiac Intensive Care Unit (CICU) over the cutoff level that the National Institute on Deafness and Other Communication Disorders (NIDOCD) indicated can cause physiological damage?
14. If you were a nurse working in the CICU, would this finding concern you?
15. What does this study tell you about the noise level on the weekend? Based on this study, would you be more or less concerned about noise levels on the weekend?
16. Look at Table 2 in the article. Which location on the units was the noisiest on average (sound pressure level [SPL]? How noisy was it on average?
17. Look at Table 2 in the article. Were patient interactions or employee interactions louder?
18. The sample was found to reflect the same gender and age distribution of the population of inpatient nurses within this facility. What does this mean about the representativeness of the sample for this facility? Of a national population of nurses from all inpatient facilities? How does this affect the generalizability of the study results?
19. In the statistical analysis and results section, the authors report that the noise level in the three units was compared. Because they were comparing three groups, what statistical test was utilized, and where were significant differences found? The researchers later offer an explanation for why this may be the case. What explanation is provided?
20. Look at Table 3 in the article. What was the most frequent source of noise?
21. Look at Table 3 in the article. What was the most frequent location of noise?
22. Look at Table 3 in the article. What was the most frequent source of noise that was > 75 dBA?
23. Look at Table 5 in the article. Pearson correlation coefficients were calculated to determine the association between heart rate and noise level (SPL). What do you know about the overall correlation between these two variables in this sample?
24. Calculate the coefficient of determination. What percentage of variance in heart rate is explained by the noise level in the overall relationship?
25. Look at Table 5 in the article. Was this correlation between heart rate and noise level (SPL) seen in the CICU?
26. Look at Table 5 in the article. At what noise location is the correlation between heart rate and noise level the strongest?
27. The researchers indicate that significant positive correlations were found between heart rate and noise levels in patients’ rooms, communicating with staff and between patients and their families and during all nursing activities. They explain that direct care may be a confounding factor for this finding. Explain what this means in your own words.