1.(This begins Exercise 1 which includes questions 1-11)

The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

- What two groups does the unadjusted mean wage difference (A) compare?

a. Females to males.

b. Females to males of the same age.

c. Females to males of the same age and union membership status.

d. Females to males of the same age, union membership status and job type.

e. Females to males of the same age, union membership status, job type, occupational sector, and marital status.

1point

2.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

-The mean hourly wage for males in the sample is $9.99/hour.What is the mean hourly wage for females in the sample?

a. $ 9.99

b. $ 2.12

c. $ 12.11

d. $ 7.87

1point

3.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

- Compute and report a 95% CI for the unadjusted mean difference in hourly wages for females compared to males.

a. (-$3.00, -$1.24)

b. (-$2.85, -$1.13)

c. (-$2.78, -$1.06)

d. (-$2.87, -$1.15)

1point

4.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

- What two groups mean does the mean wage difference, adjusted for age and union membership (C) compare?

a. Females to males.

b. Females to males of the same age.

c. Females to males of the same age and union membership status.

d. Females to males of the same age, union membership status and job type.

e. Females to males of the same age, union membership status, job type, occupational sector, and marital status.

1point

5.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

-Compute and report a 95% CI for the mean difference in wages for females compared to males adjusted for age and union membership.

a. (-$3.00, -$1.24)

b. (-$3.00, -$1.28)

c. (-$2.85, -$1.13)

d. (-$2.78, -$1.06)

1point

6.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

- What two groups does the mean wage difference, adjusted for age, union membership, job type, sector and marital status (E) compare?

a. Females to males.

b. Females to males of the same age.

c. Females to males of the same age and union membership status.

d. Females to males of the same age, union membership status and job type.

e. Females to males of the same age, union membership status, job type, occupational sector, and marital status.

1point

7.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

- Compute and report 95% CI for the mean difference in wages for females compared to males adjusted for age and union membership, job type, sector and marital status.

a. (-$3.00, -$1.24)

b. (-$3.00, -$1.28)

c. (-$2.78, -$1.06)

d. (-$2.87, -$1.15)

1point

8.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

-What is the maximumdifference (in absolute value) between the unadjusted mean difference in wages and the four adjusted estimates?

a. $0.13

b. $2.12

c. $0.20

d. $0.02

1point

9.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

- How do the 95% CIs for the five mean differences (A-E) compare: do they all overlap, or do at least some of the 5 not overlap with each other?

a. There is no overlap between any of the 5 confidence intervals.

b. All five confidence intervals overlap with each other.

c. All of the confidence intervals for the adjusted associations overlap with each other, but none overlap with the confidence interval for the unadjusted association.

1point

10.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

-Do these results suggest that the original, unadjusted mean difference in salaries was mainly attributable to at least some of the factors used for adjustment in the above table?

a. Yes.

b. No.

c. Yes: specifically, these analyses show that the original, unadjusted mean difference was primarily attributable to age difference between males and females.

d. Yes: specifically, these analyses show that the original, unadjusted mean difference was primarily attributable to differences in the job type distributions between males and females.

1point

11.The following results come from a dataset that contains information on a random sample of 534 U.S. workers surveyed in 1985.The dataset contains information about the hourly wage (in U.S. dollars) and the sex of each of the workers surveyed, as well as information about each worker's age (categorized into 4 quartiles), union membership, job type (collapsed into 8 different categories), marital status (married or not), and occupational sector (3 categories). You are interested in assessing the relationship between hourly wages and sex in the population of workers from which the sample was taken.Forty-six percent of the sample is female, and the overall average hourly salary for the 534 workers is $9.00 per hour (in 1985 dollars).

You are reviewing the following set of salary mean difference estimates comparing female and male workers.The first estimate is the unadjusted association, and the subsequent estimates in the table have been adjusted for the characteristics listed.

- Do these results show that the relationship between hourly wages and sex was modified by age?

a. Yes, because the unadjusted mean differences in wages, and the mean difference in wages adjusted for age are different.

b. No, because the unadjusted mean differences in wages, and the mean difference in wages adjusted for age are not different.

c. Additional information is needed to assess whether relationship between hourly wages and sex was modified by age

1point

12.(This begins Exercise 2which includes questions 12-14)

A 2011 article published in the Journal of Substance Abuse Treatment reports the results of arandomized clinical trialdesigned to assess the effects of acupuncture on anxiety related to withdrawal from psychoactive drugs. As per the abstract: (you may recall this article from Homework 1 of Statistical Concepts1

"Auricular acupuncture (AA) is a widely accepted treatment option for substance abuse that is used in more than 700 treatment centers worldwide. Despite claims of perceived clinical benefits by patients and treatment staff, research efforts have failed to substantiate purported benefits, and the mechanism(s) by which AA serves in the treatment of addiction remain inconclusive. Numerous studies have shown AA to

be an effective treatment for perioperative anxiety. In this study, we hypothesize that AA reduces the anxiety associated with withdrawal from psychoactive drugs. The study used a randomized, controlled design and included a sample of 101 patients recruited from an addiction treatment service. Subjects were assigned to one of three treatment groups (National Acupuncture Detoxification Association [NADA] AA, AA at sham points, or treatment setting control) and were instructed to attend treatment sessions for 3 days. The primary outcome measurestate anxiety was assessed using a pretest-posttest treatment design."The following boxplots (Figure 4 in the article) show the main results from this study: (higher anxiety scores indicate higher anxiety)

- The boxplots show the unadjusted comparisons of the anxiety score distributions before and after the treatment sessions for each of the 3 treatments.The authors do not provide any adjusted comparisons in the manuscript.Is this appropriate, or did they blatantly ignore a high potential for confounding?

a. Because this study is an observational study, these results are likely to be confounded by other participant characteristics.

b. Because this study is an observational study, these results are not likely to be confounded by other participant characteristics.

c. Because this study is an randomized trial, these results are likely to be confounded by other participant characteristics.

d. Because this study is an randomized trial, these results are not likely to be confounded by other participant characteristics.

1point

13.A 2011 article published in the Journal of Substance Abuse Treatment reports the results of arandomized clinical trialdesigned to assess the effects of acupuncture on anxiety related to withdrawal from psychoactive drugs. As per the abstract: (you may recall this article from Homework 1 of Statistical Concepts1

"Auricular acupuncture (AA) is a widely accepted treatment option for substance abuse that is used in more than 700 treatment centers worldwide. Despite claims of perceived clinical benefits by patients and treatment staff, research efforts have failed to substantiate purported benefits, and the mechanism(s) by which AA serves in the treatment of addiction remain inconclusive. Numerous studies have shown AA to

be an effective treatment for perioperative anxiety. In this study, we hypothesize that AA reduces the anxiety associated with withdrawal from psychoactive drugs. The study used a randomized, controlled design and included a sample of 101 patients recruited from an addiction treatment service. Subjects were assigned to one of three treatment groups (National Acupuncture Detoxification Association [NADA] AA, AA at sham points, or treatment setting control) and were instructed to attend treatment sessions for 3 days. The primary outcome measurestate anxiety was assessed using a pretest-posttest treatment design."The following boxplots (Figure 4 in the article) show the main results from this study: (higher anxiety scores indicate higher anxiety)

-Based on the boxplots, what conclusion can you make about the efficacy of acupuncture on anxiety reduced as compared to the other two groups?

a. Acupuncture is clearly better than the other approaches with regards to lowering anxiety.

b. Acupuncture is clearly worse than the other approaches with regards to lowering anxiety.

c. Acupuncture is clearly worse than sham acupuncture, but better than relaxation therapy, with regards to lowering anxiety.

d. There does not appear to any advantage or disadvantages of acupuncture for treating anxiety in this population , compared to sham acupuncture or relaxation.

1point

14.A 2011 article published in the Journal of Substance Abuse Treatment reports the results of arandomized clinical trialdesigned to assess the effects of acupuncture on anxiety related to withdrawal from psychoactive drugs. As per the abstract: (you may recall this article from Homework 1 of Statistical Concepts1

"Auricular acupuncture (AA) is a widely accepted treatment option for substance abuse that is used in more than 700 treatment centers worldwide. Despite claims of perceived clinical benefits by patients and treatment staff, research efforts have failed to substantiate purported benefits, and the mechanism(s) by which AA serves in the treatment of addiction remain inconclusive. Numerous studies have shown AA to

be an effective treatment for perioperative anxiety. In this study, we hypothesize that AA reduces the anxiety associated with withdrawal from psychoactive drugs. The study used a randomized, controlled design and included a sample of 101 patients recruited from an addiction treatment service. Subjects were assigned to one of three treatment groups (National Acupuncture Detoxification Association [NADA] AA, AA at sham points, or treatment setting control) and were instructed to attend treatment sessions for 3 days. The primary outcome measurestate anxiety was assessed using a pretest-posttest treatment design."The following boxplots (Figure 4 in the article) show the main results from this study: (higher anxiety scores indicate higher anxiety)

- What additonal graphics would you need to see in order to ascertain whether the relationship between anxiety scores and acupuncture (as compared to relaxation therapy) is potentially modified by sex?

a. The boxplots of the before and after treatment anxiety distributions for both the acupuncture and relaxtion group, after the distributions have been adjusted for sex differences between the acupuncture and relaxation groups.

b. The before and after treatment anxiety score distributions for the acupuncture and relaxation groups, presented separately by sex.

c. None. The graphics given with this exercise are sufficient to assess effect modification by sex.

1point

15.(This begins Exercise 3which includes questions 15-16)

An October 25, 2012 article in the New England Journal of Medicine[1]reports the results of a study examining aspirin and survival among patients with colorectal cancer.As per the authors:

"We obtained data on 964 patients with rectal or colon cancer from the Nurses'Health Study and the Health Professionals Follow-up Study, including data on aspirinuse after diagnosis and the presence or absence of PIK3CA mutation"

"We used data from two prospective cohort studies ,the Nurses' Health Study (NHS, involving 121,700 women who were enrolled in 1976) and the Health Professionals Follow-up Study (HPFS, involving51,500 men who were enrolled in 1986)."

Additionally, when discussing the aspirin exposure assessment:

"Assessment of aspirin use in the NHS and HPFS cohorts has been described in detail previously. In 1980, NHS participants were asked whether they regularly used aspirin in most weeks, as well as the dose and duration of use; thereafter, this information was updated biennially (except in 1986). Beginning in 1986, HPFS participants were asked whether they regularly used aspirin two or more times each week; after 1992, information on the average dose used each week was requested."

The authors present the following graphic as part of the article.The starting point (time 0) for each patient in the same was year of diagnosis.

- Based on theresults presented, what role does the mutation play does the P1K13CA gene type (mutant or wild type) with regards to the relationship between mortality (colon-cancer related and overall) and aspirin usage?(confounder, effect modifier, neither)

a. The overall relationship between mortality and aspirin usage is confounded by P1K13CA gene type differences between the aspirin and no-aspirin groups.

b. The overall relationship between mortality and aspirin usage is modified by the P1K13CA gene type.

c. Since the P1K13CA gene type is not associated with mortality it can neither counfound, nor modify, this association.

d. Since this is an observational study, P1K13CA gene type can counfound, but cannot modify the association between mortality and aspirin usage.

1point

16.An October 25, 2012 article in the New England Journal of Medicine[1]reports the results of a study examining aspirin and survival among patients with colorectal cancer.As per the authors:

"We obtained data on 964 patients with rectal or colon cancer from the Nurses'Health Study and the Health Professionals Follow-up Study, including data on aspirinuse after diagnosis and the presence or absence of PIK3CA mutation"

"We used data from two prospective cohort studies ,the Nurses' Health Study (NHS, involving 121,700 women who were enrolled in 1976) and the Health Professionals Follow-up Study (HPFS, involving51,500 men who were enrolled in 1986)."

Additionally, when discussing the aspirin exposure assessment:

"Assessment of aspirin use in the NHS and HPFS cohorts has been described in detail previously. In 1980, NHS participants were asked whether they regularly used aspirin in most weeks, as well as the dose and duration of use; thereafter, this information was updated biennially (except in 1986). Beginning in 1986, HPFS participants were asked whether they regularly used aspirin two or more times each week; after 1992, information on the average dose used each week was requested."

The authors present the following graphic as part of the article.The starting point (time 0) for each patient in the same was year of diagnosis.

- The researchers ultimately did separate analyses of the death/aspirin relationship for the two P1K13CA gene type groups (see table below for the colorectal cancer death outcomes).However, theses associations were adjusted for other factors (age, disease stage etc...).Why was this necessary even though the results were computed separately by P1K13CA gene type?

a. This isn't necessary: as this is a randomized study, confounding is not likely.

b. As this is an observational study, it is possible that the unadjusted associations between mortality and aspirin are confounded by other factors that difference between the aspirin and no-aspirin groups for both P1K13CA gene type groups.

c. This is not necessary, as the researchers have already estimated separate mortality/ aspirin association by P1K13CA gene type groups.

1point

17.(This begins Exercise 1 which includes questions 17-20)

The following exercise involves information from the July 2004 AJPH article "Asian/Pacific Islander Adolescent Sexual Orientation and Suicide Risk in Guam" .The authors used survey results based on information collected from 1,381 adolescents in Guam.

A summary of the authors' research motives and results can be found in the abstract:

The following table shows the unadjusted and adjusted associations between suicide attempts (at least one), and same-sex orientation, presented separately for male and female adolescents:

- Can you ascertain, based on these results whether the overall association between suicide attempts and homosexual identity was confounded by sex?

a. No, because the researchers do not give an overall unadjusted association between suicide and homosexuality, both unadjusted and adjusted for sex.

b. Yes, because the researchers provide both unadjusted and adjusted associations between suicide and homosexuality, separately for males and females.

c. Not directly: however, because this is a randomized study, the suicide/sexual orientation association is not likely to be confounded by sex differences between those who identify as homosexual and not homosexual.

1point

18.The following exercise involves information from the July 2004 AJPH article "Asian/Pacific Islander Adolescent Sexual Orientation and Suicide Risk in Guam" .The authors used survey results based on information collected from 1,381 adolescents in Guam.

The following table shows the unadjusted and adjusted associations between suicide attempts (at least one), and same-sex orientation, presented separately for male and female adolescents:

- For males, does it appear the association between suicide attempts and homosexual identity was confounded by at least some of the adjustment factors used to get the unadjusted estimate?

a. Yes, because the unadjusted association is statistically significant, and the adjusted association is not statistically significant.

b. Yes, because the unadjusted association is not statistically significant, and the adjusted association is statistically significant.

c. No, because both the unadjusted and adjusted associations are very similar in value, and have overlapping confidence intervals.

1point

19.The following exercise involves information from the July 2004 AJPH article "Asian/Pacific Islander Adolescent Sexual Orientation and Suicide Risk in Guam" .The authors used survey results based on information collected from 1,381 adolescents in Guam.

The following table shows the unadjusted and adjusted associations between suicide attempts (at least one), and same-sex orientation, presented separately for male and female adolescents:

- By presenting the above associations separately by sex, what phenomenon are the researchers investigating with regards to sex? (confounding,effect modification, neither)

a. Confounding

b. Effect Modification

1point

20.The following exercise involves information from the July 2004 AJPH article "Asian/Pacific Islander Adolescent Sexual Orientation and Suicide Risk in Guam" .The authors used survey results based on information collected from 1,381 adolescents in Guam.

The following table shows the unadjusted and adjusted associations between suicide attempts (at least one), and same-sex orientation, presented separately for male and female adolescents:

- Logistic regression was used to estimate the above unadjusted and adjusted odds ratios (and the 95% CIs).Suppose you were interested in estimating the proportions of males who attempted suicide among those who identified as homosexual, and those who identified as heterosexual.What additional information would you need?

a. The slope of sexual identity (x1 = 1 if homosexual, 0 if not) from the logistic regression relating suicide attempts (yes/no) to sexual identity.

b. The exponentiated slope of sexual identity (x1 = 1 if homosexual, 0 if not) from the logistic regression relating suicide attempts (yes/no) to sexual identity.

c. The intercept (or exponentiated intercept) from the logistic regression relating suicide attempts (yes/no) to sexual identity.

d. The relative risk of suicide attempts for those who identify as homosexual compared to those who identify as not being homosexual,