Question 142pts

Research Question	Independent Variable	Dependent Variable
Do physicians in regions of high healthcare spending order more magnetic resonance imaging (MRI) exams for low back pain than those in low spending regions?	[ Select ] ["A. Physicians", "B. Regional health care spending", "C. Low back pain", "D. Number of MRI exams ordered"]	[ Select ] ["A. Physicians", "B. Regional health care spending", "C. Low back pain", "D. Number of MRI exams ordered"]

Flag question: Spacer

Multiple Choice

Choose ONLY ONE RESPONSE per question.

Flag question: Question 15Question 152pts

Which of the following non-parametric statistical test(s) is/are used for paired data?

Group of answer choices

A. Sign test

B. Mann-Whitney test (Wilcoxon Rank Sum test)

C. Wilcoxon Signed-Rank test

D. A and C

E. None of the above

Flag question: Question 16Question 162pts

Which of the following is FALSE regarding the Spearman correlation coefficient?

Group of answer choices

A. It utilizes ranking the data to correct for outliers

B. It tells you the strength and direction of a linear relationship between two continuous measures

C. A value of zero indicates no relationship

D. It is interpreted differently than a Pearson correlation coefficient

E. It ranges from -1.0 to +1.0

Flag question: Question 17Question 172pts

In a study that examined dietary fat intake, a random sample of 10 patients was drawn from a population. The mean grams of fat per day in the sample was 82 grams/day and the standard deviation of the sample was 26. Which is the correct equation to determine the 95% confidence interval (two-tailed)?

Group of answer choices

A. 82 1.96 (26)

B. 82 1.96 (26/10)

C. 82 2.262 (26/10)

D. 82 2.228 (26/10)

E. 82 1.833 (26/10)

Flag question: Question 18Question 182pts

Which of the following statistical tests can be used to evaluate the relationship between a categorical variable (an independent variable with >2 categories) and a dependent variable?

Group of answer choices

A. Chi-squared test

B. McNemar's test

C. One-way ANOVA

D. A and C

E. All of the above

Flag question: Question 19Question 192pts

While exploring a dataset, you plot two continuous variables and arrive at the scatterplot below. Based on this plot, what do you expect regarding the correlation coefficient between the two measures?

Group of answer choices

A. It will be strong and the sign will be negative

B. It will be weak and the sign will be negative

C. It will be strong and the sign will be positive

D. It will be weak and the sign will be positive

Flag question: Question 20Question 202pts

Which of the following is correct regarding the Sign test?

Group of answer choices

A. It compares the count of observed positive signs to those expected by chance alone

B. For small samples (n < 20) it uses the binomial equation to estimate the p-value

C. It is a non-parametric test

D. A and C

E. All of the above

Flag question: Question 21Question 212pts

Which song was played in the survival analysis video?

Group of answer choices

A. "Eye of the Tiger" by Survivor (. . . a great song, by the way)

B. "I Will Survive" by Gloria Gaynor

C. "Hearts on Fire" by John Cafferty & The Beaver Brown Band

D. "Burning Heart" by Survivor

Flag question: Question 22Question 222pts

Based on the linear regression output below, what is the coefficient of determination?

Group of answer choices

A. 0.77

B. 0.31

C. 0.23

D. Cannot determine

Flag question: Question 23Question 232pts

In survival analysis, which of the following are potential reasons a subject may be censored?

Group of answer choices

A. The subject is lost to follow-up

B. The subject experiences the outcome

C. The study concludes

D. A and C

E. All of the above

Flag question: Question 24Question 242pts

Which of the following is FALSE regarding survival curves?

Group of answer choices

A. It always starts at 1.0 (100% survival)

B. If survival improves it can increase

C. It can only decrease or stay the same

D. If all the study participants don't experience the outcome it may never reach 0

Flag question: Question 25Question 252pts

Which of the following is NOT an assumption of linear regression?

Group of answer choices

A. The relationship between the independent and dependent variable is linear

B. The residuals are homoscedastic

C. The dependent variable is normally distributed across levels of the independent variable

D. The relationship between the independent and dependent variable is U-shaped (i.e., non-linear)

E. Each observation is sampled independently

Flag question: Question 26Question 262pts

Which of the following can affect thereliabilityof a diagnostic test?

Group of answer choices

A. Intrasubject variation

B. Intraobserver variation

C. Interobserver variation

D. All of the above

Flag question: Question 27Question 272pts

Your local middle school, in an effort to combat the increase of vaping among students, has received grant funding to study the relationship between vaping and fitness; you are asked to assist in the analysis. You are provided with data on 561 students aged 10-14 years. The independent variable that identifies vaping is dichomotized - Yes (coded as "1") if the student currently vapes, otherwise No (coded as "0"). The outcome is measured as VO2 max, an individual's maximal oxygen uptake during exercise (variable VO2max), measured in liters per minute. Which of the following statistical tests would you use to test the relationship between vaping (yes versus no) and fitness (VO2max)?

Group of answer choices

A. Chi-squared test

B. McNemar's test

C. Paired t-test

D. Two-sample (unpaired) t-test

E. Wilcoxon Signed-Rank test

Flag question: Spacer

Information for Questions 28-29: Researchers examined survival following aortic valve replacement surgery among diabetics (n = 84) versus non-diabetics (n = 460). The following Kaplan-Meier survival curve estimates survival of subjects over 15 years of follow-up.

Flag question: Question 28Question 282pts

Overall which group has higher survival?

Group of answer choices

A. Diabetics

B. Non-diabetics

C. Cannot determine

Flag question: Question 29Question 292pts

What is the median survival and 95% Confidence Intervalamong diabetics?

View keyboard shortcuts

EditViewInsertFormatToolsTable

12pt

Paragraph

Flag question: Question 30Question 302pts

Which of the following is NOT true regarding the kappa statistic?

Group of answer choices

A. It is a measure of interobserver reliability regarding a diagnostic test

B. It ranges between -1.0 and +1.0

C. It evaluates agreement beyond that of chance

D. A kappa value of 0.2 suggests strong interobserver agreement

Flag question: Question 31Question 312pts

Which of the following is the correct definition of a Bonferroni correction?

Group of answer choices

A. A mechanical modification to Italian Vespa scooters that results in enhanced motor performance

B. A traditional Italian technique where a small hand tool is used to create the undulating patterns on the edges of homemade Ravioli

C. A statistical method used to counteract the problem of multiple comparisons that reduces the critical alpha level (to a more conservative level thereby reducing the potential of Type I errors)

D. A popular Chef Boyardee pasta product sold in the United States

E. A variation of the two-sample t-test that accounts for unequal variance in the dependent variable across the two groups of the independent variable

Flag question: Spacer

Information for Questions 32-34: A researcher is interested in examining the relationship between diet and high density lipoprotein (HDL) levels. She obtains detailed data on a variety of daily dietary measures including: protein, carbohydrate, sugar, fiber, fat, and cholesterol intake. The following is the correlation matrix of the dataset:

Flag question: Question 32Question 322pts

Which dietary intake measure has the strongest association with HDL?

Group of answer choices

A. Protein intake

B. Carbohydrate intake

C. Sugar intake

D. Fiber intake

E. Total fat intake

F. Cholesterol intake

Flag question: Question 33Question 332pts

Based on the your answer for the previous question and an of 0.05, do you accept or reject the null hypothesis?

Group of answer choices

A. ACCEPT the null hypothesis

B. REJECT the null hypothesis

Flag question: Question 34Question 342pts

What percent of the variation in HDL is explained by the dietary measure you just identified (in question 32) as being most strongly associated? (Hint: R-squared)

Group of answer choices

A. -11.9%

B. 12.8%

C. 1.6%

D. 5.0%

E. Cannot determine

Flag question: Spacer

Information for Questions 35-37: You have been given a dataset that contains information on 100 women who were admitted to Mountain Hospital to deliver their first child during a one-year period. Measurements include salivary cotinine (an indicator of exposure to tobacco smoke), systolic blood pressure, maternal age, gestational age (number of weeks of pregnancy at delivery), number of hours in labor, and whether the delivery was vaginal or by Cesarean section.

You examine the data using Stata and find the following relationship between systolic blood pressure and salivary cotinine:

Flag question: Question 35Question 352pts

Which of the following conclusions is correct?

Group of answer choices

A. Higher cotinine levels are associated with lower systolic blood pressure

B. Cotinine levels are not correlated with systolic blood pressure

C. Higher systolic blood pressure is associated with higher cotinine levels

D. Variation in cotinine level explains 23.57% of the variation in systolic blood pressure

Flag question: Question 36Question 362pts

You are subsequently able to obtain comparable data, including age in years, from all 5 hospitals in town on all women admitted during the same one-year period to deliver their first child. Before proceeding further with any analysis, you wish to make sure that women are comparable across the 5 hospitals. You are specifically interested in the ages of women delivering at the different hospitals. What statistical test would be best for this analysis?

Group of answer choices

A. Pearson correlation

B. Two-sample (unpaired) t-test

C. Paired t-test

D. One-way analysis of variance

E. Simple linear regression

Flag question: Question 37Question 372pts

The p-value from the appropriate test in the previous question is 0.02, leading you to reject the null hypothesis.What is the null hypothesisregarding women admitted to deliver their first child?

Group of answer choices

A. The mean age at the 5 hospitals are not correlated

B. The greater the number of women delivering at a hospital, the higher the mean age

C. The mean age is different at each hospital

D. The mean age at one hospital is different from the rest

E. The mean age is the same at each hospital

Flag question: Spacer

Information for Questions 38-42: My research lab has been examining national healthcare data and we could use your help with interpreting some analyses. We're writing a paper investigating the use of nurse practitioners (and physician assistants) as a usual source of primary care. First, we're interested in understanding whether the patients who see NPs or PAs are different than those who see physicians. The dataset contains information on 112,401 patients.

Here is some information on the two variables we're examining:

Flag question: Question 38Question 382pts

What type of variable is "provider"?

Group of answer choices

A. Ratio variable

B. Nominal variable

C. Ordinal variable

D. Dichotomous variable

E. Cannot tell without more information

Flag question: Question 39Question 392pts

My PhD student ran the following statistical test to examine the relationship between provider type and obesity status:

What is the statistical test in the above output?

Group of answer choices

A. Two-sample (unpaired) t-test

B. One-way analysis of variance

C. Chi-squared test

D. McNemar's test

E. Mann-Whitney test

Flag question: Question 40Question 402pts

Which of the following conclusions is correct?

Group of answer choices

A. We REJECT the null hypothesis that obesity is the same across different provider types

B. Approximately 38% of NP/PA patients are obese

C. The test statistic is 229.8

D. The percentage of patients who are obese (i.e., the prevalence) is 29.9%

E. All statements are correct

Flag question: Question 41Question 412pts

We decided to look at the number of prescription medications of the patients according to usual source of care provider type. The variable rxtot is a continuous variable that is the total number of prescription medication refills per year for each patient.

Using the rxtot variable my PhD student produced the following:

Based on the output above, which of the following is INCORRECT?

Group of answer choices

A. We REJECT the null hypothesis that the number of medication refills is the same across different provider types

B. We ACCEPT the null hypothesis that the number of medication refills is the same across different provider types

C. The F statistic for the test is calculated as: 103,966.1 / 653.3

D. The output is the result of a one-way analysis of variance test

Flag question: Question 42Question 422pts

With your savvy HS700 skills you remember that one-way ANOVA has certain assumptions that must be met in order to produce valid results. You urge the PhD student to look at the distribution of total medical prescriptions. Here is the histogram of rxtot:

Whoa! - good call, you notice that the data are highly skewed. These data violate the assumptions of the parametric one-way ANOVA test.

Which non-parametric test should be to examine the relationship between the four provider types and number of medication refills?

Group of answer choices

A. Sign test

B. Mann-Whitney test (Wilcoxon Rank Sum test)

C. Wilcoxon Signed-Rank test

D. None of the above

Flag question: Spacer

Information for Questions 43-45: There is some evidence that low levels of alcohol consumption has protective effects on health. Using a dataset of 1,000 adults you examine the effect of alcohol consumption (independent variable) on high density lipoprotein level (dependent variable).HDL is measured in mg/mL and alcohol consumption in mL per day. You decide to run a simple linear regression and observe the following:

Flag question: Question 43Question 432pts

If the variable for HDL is named hdl and the alcohol consumption variable is called etoh, what is the command in Stata that was used to perform the linear regression above?

Flag question: Question 44Question 442pts

What is the correct interpretation of the effect of daily alcohol consumption on HDL?

Group of answer choices

A. A 1 mL increase in alcohol per day is associated with a 0.05 mg/mL increase in HDL

B. A 1 mg/mL increase in HDL is associated with an increase of 0.05 mL of alcohol consumption

C. A 1 mL increase in alcohol per day is associated with a 50.8 mg/mL increase in HDL

D. Cannot determine

Flag question: Question 45Question 452pts

Which of the following is correct?

Group of answer choices

A. The association between daily alcohol consumption and HDL is statistically significant

B. The average residual of the model is 15.6

C. Daily alcohol consumption explains approximately 1% of the variation in HDL level

D. A and C

E. All the above

Flag question: Spacer

Information for Questions 46 -48: You decide that because of the likely large number of non-drinkers in your dataset, conceptually it would make more sense to create a categorical variable. A single alcoholic beverage contains on average 18mL of pure alcohol. Using the continuous measure of alcohol consumption you create the following variable called drinks:

Flag question: Question 46Question 462pts

What type of variable is the drinks variable?

Group of answer choices

A. Ratio variable

B. Nominal categorical variable

C. Ordinal variable

D. Dichotomous variable

Flag question: Question 47Question 472pts

What percentage of people consume 2 or fewer alcoholic drinks per day?

Group of answer choices

A. 6.1%

B. 88.5%

C. 11.5%

D. 82.4%

E. Cannot determine without more information

Flag question: Question 48Question 482pts

The following table compares HDL level at each level of alcohol consumption to that of non-drinkers using a two-sample (unpaired) t-test. For instance the difference in the mean HDL among those who drink between 1 to 2 drinks per day compared to non-drinkers is -2.90 (95% CI for the difference -7.0, 1.2).

Alcohol drinking status	Difference in means (95% CI), mg/mL
Non-drinker	0.0 (reference)
1 drink or less	0.8 (-3.2, 4.8)
1 to 2 drinks	-2.9 (-7.0, 1.2)
More than 2 drinks	-4.6 (-7.8, -1.5)

At what level of daily alcohol consumption is there a statistically significant association with HDL level?

Group of answer choices

A. 1 drink or less per day

B. 1 to 2 drinks per day

C. More than 2 drinks per day

D. Cannot determine without more information

Flag question: Question 49Question 492pts

Using a study population of older women, a researcher examines the relationship between depression in childhood and depression in adulthood. Here are the data in a 2x2 table:

		Adulthood depression
		Yes	No	Row total
Childhood depression	Yes	17	14	31
No	22	48	70
Column total	39	62	101

Basedon the appropriate statistical testfor these data, what is the value of the chi-squared statistic for the association between childhood and adulthood depression? (Hint, are these data paired or unpaired?)

Group of answer choices

A. 1.78

B. 4.97

C. 2.45

D. 5.89

Flag question: Spacer

Information for Questions 50-51: A population-based evaluation was conducted of annual screening for lung cancer with low-dose computed tomography (CT). The participants underwent baseline evaluations in 1996 and repeated annual examinations for two years. Confirmation of a lung cancer diagnosis was based on cellular pathology. Here is a summary of the study results:

		Lung cancer confirmation
		Yes	No
CT Result	Positive	56	532
Negative	4	13,194

Flag question: Question 50Question 502pts

What is thesensitivityof CT screening for lung cancer?

Group of answer choices

A. 55%

B. 64%

C. 93%

D. 98%

Flag question: Question 51Question 512pts

What is thepositive predictive valueof the test? In one sentence, explain how to interpret it.