Advanced Epidemiology (PUB 695) – Fall 2021

HOMEWORK 3

Due: Sunday, Nov. 21 at 11:30pm Eastern

NAME____________________________________

Instructions: Please type your responses to each of the questions and then save your responses as a Word document called, “YOUR NAME HW3.doc”. Upload your homework on Moodle in “HW3 Assignment”. Only the student who uploads the document and the instructor will have access to the file. Moodle will automatically include the time the file was uploaded.

Instructions:

Please read all instructions and questions carefully and, for the short answer questions, make sure you use your own words if you want full credit for your answers.

As the instructions state below, the multiple choice questions only have ONE right answer. Select only ONE for each answer.

Please do not delete anything from the homework sheet. Just add your answer to the sheet. For example, don’t delete the instructions, the question numbering, any graphs, the choices the mult choice questions, the point totals, the grading page, etc.

Please type your name onto the homework.

Please type your responses to each of the questions and then save your responses as a Microsoft word document called, “YOUR NAME HW2.doc”. Upload your homework onto Moodle in the “HW2 Assignment” folder.

Lateness policy: The HW is due Sunday, Nov. 21 at 11:30pm Eastern. If you miss this deadline, you can submit the homework up to three days late (up until Wednesday, Nov. 24 at 11:30pm Eastern) without penalty this time only. But I prefer that you meet the Sunday, Nov. 21,11:30pm Eastern deadline. After 11:30pm Eastern on Wednesday, Nov. 24, the HW cannot be accepted and a zero grade will be assigned which cannot be dropped. The lowest HW grade can only be dropped if all three HWs are fully completed and submitted by the deadline in a Microsoft Word file that I can open.

Submit only Microsoft Word files. Do not submit pdf files, pages files, RTF files or any other files. If I am unable to open the file you submit and the deadline has passed, I cannot give you credit for the assignment.

Always log back in and open the file you submitted to ensure you submitted a correct file and check email daily in case I need to reach you.

If Moodle is down when you submit, email the HW to rocosta@adelphi.edu before the deadline,

No two HWs between students should be similar, and it is very obvious when HWs are similar even if students attempt to disguise it. When students do not submit their own work, Adelphi’s rule about Academic Integrity is violated and it has serious repercussions. Do your own work.

Please do not google answers. Use the lecture slides instead. Info posted on websites may be inaccurate and/or the info may not be described in the way this department has chosen to teach these concepts. This is the reason we take courses in degree-granting institutions instead of just learning from material that is publicly accessible online.

Again, after submitting, be sure to log back onto Moodle and open the HW file you submitted to ensure you submitted the correct file. Please see syllabus for additional details.

Be sure to read each question very carefully! In the field of public health, we have to learn to be very detail-oriented.

Grade:

Part 1: ___ out of 15

Part 2: ___ out of 12

Part 3: ___ out of 7

Part 4: ___ out of 16

Extra Credit ___ out of 3

Total: ___ out of 50

PART 1: Multiple Choice (Questions 1-5): Unless otherwise indicated, bold or highlight the ONE best answer. (3 points each; 15 points total)

1. A 95% confidence interval around an odds ratio is typically calculated to rule out the likelihood that this measure of association was a result of (select the BEST answer):

Selection bias

Confounding

Information bias

Random error

2. If your 3rd variable is a mediator you report the:

a. adjusted OR

b. crude OR

c. stratum-specific ORs

3. Researchers conduct a study on adults over the age of 45 to determine whether an association exists between meditation sessions and heart disease. They report that those who engage in daily meditation sessions have a risk ratio of 0.6 in developing heart disease over 10 years as compared to those who do not engage in daily meditation sessions. The 95% confidence interval had a range from 0.2 to 1.2. Select the correct statement solely based on the information provided in this scenario:

a. The authors should fail to reject the null hypothesis.

b. A Type I error could have occurred.

c. The study had sufficient power to detect a true difference of at least 0.6.

d. The researchers are confident that daily mediation sessions would protect against heart disease in at least 95 out of 100 studies, if conducted similarly.

e. The authors should have reported a 90% confidence interval to obtain statistical significance.

4. Select the correct statement concerning confounding:

If the authors report that a variable is a confounder in their study, they should adjust for that variable.

Randomization is the best way to address confounding in a retrospective cohort study.

If confounding is present, it suggests that the association between exposure and disease differs according to the levels of a third factor.

If the crude risk ratio in a study is 2.0, and the stratum –specific risk ratios for the third variable are 1.4 for one stratum and 11.3 for the other stratum, that third variable is probably an important confounder in the study.

5. What is the best definition of a Type I error?

a. False negative

b. The probability your study results can be replicated.

c. Probability of failing to reject the null when the null is, in fact, wrong.

d. Probability of rejecting the null when the null is, in fact, true

PART 2: Name that Measure of Association/Measure of Frequency (12 points)

For questions 6-9, match the measure of association that could potentially be calculated for each study design type by typing the letter of the appropriate measure of association on the lines provided below. Be sure to solely base your answers on what we learned in this course. (3 points each)

a. Rate ratio

b. Risk ratio

c. Odds ratio

d. all of the above

e. None of the above

6. Cohort Study ____________

7. Case Series ______________

8. Case-Control Study __________

9. Randomized Controlled Trial __________

PART 3: Short Answer (7 points total)

(Questions 10a and 10b): Answer the following questions using full sentences. Be as specific as possible for full credit and answer the questions in your own words.

10. In a case control study of the association between chewing tobacco and oral cancer, you

think that individuals with and without oral cancer are equally as likely to misreport that they chewed tobacco.

a. What type of misclassification is this? Be sure to state whether it is differential or non-differential, AND whether it is misclassification of the exposure or the disease in your

response. Justify your response. (3 points)

b. How would the misclassification type, you mentioned above, bias the OR of your study findings? Be sure to state whether you think it would bias the OR toward or away from the null value of 1 in your response. (2 points)

Provide your own example of non-differential or differential misclassification of the exposure. This is a totally separate question from the previous questions. (2 points)

PART 4. SPSS (16 points)

Use the SPSS output provided below to answer the following questions.

To obtain full credit on this SPSS section, be sure to do as follows:

Express your responses in complete sentences. Be sure to answer all the questions.

Highlight, bold, or underline the actual values on the SPSS output provided to indicate your answers under each question.

The following SPSS output has been provided for you to determine whether two stratum-specific odds ratios for a third variable are truly different from each other to confirm the possible presence of effect modification.)

Which statistic should you use in this table below to possibly confirm the presence of effective modification? Highlight the statistic on the output below and then write it below. (2 points)

Tests of Homogeneity of the Odds Ratio

Chi-Squared

df

Asymp. Sig.

(2-sided)

Breslow-Day

.001

1

.982

Tarone’s

.005

1

How can I interpret this result from the previous table? (Select the one correct answer) (3 points)

The stratum specific odds ratios are statistically significantly different from each other indicating that effect measure modification is present.

The stratum specific odds ratios are not statistically significantly different from each other indicating that effect measure modification is not present.

Since the result equals the null value of 1, there is no association.

If effect measure modification were present, which odds ratio would you report? (select one) (3 points)

Crude odds ratio

Stratum-specific odds ratio

Adjusted odds ratio

Your brand new research assistant hands you the following SPSS output, some of this output you do not need for this analysis and some of this output you will need to answer this analysis. It is your job to find the applicable statistics to report from the output she handed to you to answer these questions. What is the internal consistency reliability for the 6 items on this safety scale? First list the NAME of the statistic you report and this list the actual statistic (value). Also, highlight it on the output below. (3 points)

Symmetric Measures

Value

Asymp. Std. Errora

Approx. Tb

Approx. Sig.

Measure of Agreement

Kappa

.097

.092

1.131

.258

N of Valid Cases

44

a. Not assuming the null hypothesis.

b. Using the asymptotic standard error assuming the null hypothesis.

Reliability Statistics

Cronbach’s Alpha

N of Items

.778

6

Item-Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

Corrected Item-Total Correlation

Cronbach’s Alpha if Item Deleted

New employees quickly learn that they are expected to follow good safety practices

13.95

10.044

.783

.666

There are no significant compromises or shortcuts taken when worker safety is at stake

13.25

13.169

.681

.720

Where I work, employees and management work together to ensure the safest possible working conditions.

13.64

11.865

.691

.703

Employees are told when they do not follow good safety practices

13.07

14.763

.409

.771

The safety of workers is a big priority with management where I work

13.57

11.879

.720

.697

I feel free to report safety violations where I work

13.55

14.905

.110

.867

16. What is your interpretation of this alpha? Be specific in terms of what this means for this scale. (2 points)

17. Use the output above to determine what if, anything, you can do to increase the internal consistency reliability of your scale now that the data have already been collected. Describe what you can do and why. Be sure to use the statistic(s) to support your statement and, as you were asked to do for all questions, please be sure to highlight the statistic(s) that you used to answer this question. (3 points)

Be sure that you highlighted or bolded your answers on the SPSS output in the questions above from the previous part.

PART 5 – OPTIONAL EXTRA CREDIT (3 points total)

A Kappa statistic of 0.38 has been calculated for the results between two different observers.

Fully report and then interpret this result. Your answer should include two sentences.