Statistics Review in 3 Parts
This free, 3part problembased tutorial is intended to help all anesthesiologists gain a better understanding of basic statistics. Learn at your own pace by actively working through questions. Just click on “Register Now” at the bottom of page and start learning.
Schedule
Session 1  Descriptive Statistics, Probability Distributions, Normal/Gaussian Distributions, Types of Data, Odds Ratio, Relative Risk, Absolute/Relative Risk Reduction, NNT, Mean, Median, Mode, Variance, Standard Deviation, Standard Error. 
Session 2  Inferential Statistics, Hypothesis Testing, Confidence Intervals, Precision, Accuracy, CaseControl, Types of Studies (CrossSectional, Cohort, Randomized Control), Statistical Testing (Chisquare, Student’st, MannWhitney, ANOVA, Linear Regression). 
Session 3  Sensitivity, Specificity, PPV, NPV, Alpha, Beta, Power, Rule of 3s, Bias, Receiver Operator Curve, BlandAltman Analysis. 
Questions
1. Probability distributions plot

 A. Outcome vs time
 B. Outcome vs trial
 C. Outcome/trials vs trials
 D. Outcome/trials vs outcomes
2. Which of the following statements regarding descriptive statistics is true?

 A. A statistic describes a parameter
 B. A parameter describes a sample<
 C. A parameter approximates a statistic
 D. A statistic approximates a parameter
3. Which of the following is best classified as ordinal data?

 A. Gender
 B. Serum troponin
 C. ASA classification
 D. Weight
4. A study is performed in which the effect of esmolol on the number of myocardial segments found to be hypokinetic after aortic cross application is measured by transesophageal echocardiography. The experimental data obtained can best be described as

 A. Nominal
 B. Ordinal
 C. Discrete
 D. Continuous
5. The number of times laryngoscopy is performed prior to successful intubation in Mallampati 3 airways is measured in two groups of CA1 residents at the end of the CA1 year: One for which and airway simulation course was provided the first week of training and the other for which it was not. The results were reported as an average of 1.2 and 1.4 number of attempts, respectively. The results

 A. Cannot be presented this way because the data is ordinal
 B. Cannot be presented this way because the data is nominal
 C. Can be presented this way because ordinal nominal data is also continuous
 D. Can be presented this way because it is meaningful
6. The possibility of an increased risk of postoperative nausea and vomiting in pediatric patients undergoing strabismus repair is studied by performing a prospective study of 100 patients strabismus repair compared to 100 patients undergoing other surgeries and determining how many had PONV. Comparative data should be analyzed as

 A. Odds ratio
 B. Relative risk
 C. Ttest
 D. Analysis of variance
7. The data for the above experiment is as follows:
What is the odds ratio?

 A. (10/90) / (1/99)
 B. (10/100) / (1/100)
 C. (10/11) / (90/189)
 D. (10/1) / (90/99)
8. The possibility of an increased risk of postoperative nausea and vomiting in pediatric patients undergoing strabismus repair is studied by performing a retrospective chart review of 100 patients with PONV compared to 100 patients without PONV and determining how many had strabismus surgery. Comparative data should be analyzed as

 A. Odds ratio
 B. Relative risk
 C. Ttest
 D. Analysis of variance
9. The data for the above experiment is as follows:
What is the odds ratio?

 A. 95% of observations
 B. 99 % of observations
 C. 68% of observations
 D. None of the above
10. Administration of Mikoxaban, a new directacting factor 10a inhibitor, produced a decrease in the rate of markers of postoperative hypercoaguability from 50% to 30%. The relative risk reduction is:

 A. 15%
 B. 20%
 C. 40%
 D. 50%
11. Administration of Mikoxaban, a new directacting factor 10a inhibitor, produced a decrease in the rate of markers of postoperative hypercoaguability from 50% to 30%. The number needed to treat is

 A. 2
 B. 4
 C. 5
 D. 10
12. In the set 1, 2, 2, 3, 3, 3, the mean, median, and mode, respectively, are

 A. 2, 2.5, 3
 B. 2.3, 2.5, 3
 C. 2.5, 2.5, 3
 D. 3, 2.5, 2.5
13. In a distribution that is highly positively skewed

 A. Mean > median > mode
 B. Median > mean > mode
 C. Mode > median > mean
 D. Mode > mean > median
14. Which of the following is the best definition of standard deviation?

 A. The absolute difference between minimum and maximum values
 B. The sum of the squares of the differences between observations divided by n1
 C. How clustered data points are around the mean
 D. The square root of the standard error
15. In a study looking at the average weight in kg of 100 patients admitted to the ICU of a hospital, which of the following correctly describes the variation around the mean in kilograms?

 A. Standard deviation
 B. Standard error
 C. Variance
 D. Median
16. Preoperative betanatriuretic peptide are measured in 6 patients scheduled for cardiac surgery every day for one week. The mean BNP value for each day is calculated. What value best describes the dispersion around the average of these means?

 A. Standard deviation
 B. Standard error
 C. Variance
 D. Mean
17. In a normal distribution, +1 to 0 one standard deviation contains

 A. 95% of observations
 B. 99 % of observations
 C. 68% of observations
 D. None of the above
18. Which of the following statements about confidence intervals is correct?

 A. They must include at least 95% of observed values
 B. Nonrandom distributions increase the size of confidence intervals
 C. They describe the variation of sampled data
 D. They describe the precision in measuring a population parameter
19. The mean preoperative hemoglobin values measured on 100 individuals prior to surgery was 12.2 with a 95% confidence interval of 0.4 means all of the following except

 A. 95% of time, a preoperative Hg value obtained from this population will be between 12.6 and 11.8
 B. It is assumed that the values of Hg values in this population is normally distributed
 C. A Hg value above 12.6 or below 11.8 belongs to a different population
 D. The confidence interval was probably calculated using CI = 1.96 x standard deviation
20. A study using drug treatment yields an odds ratio of 0.7 for death with a 95% confidence interval of 0.90.5. Based on these findings, one can conclude the drug

 A. Decreases mortality
 B. Does not decrease mortality
 C. Increases mortality
 D. Does not increase mortality
21. A test for brain oxygenation that produces values consistently 2040% higher than actual values is

 A. Precise and accurate
 B. Precise but not accurate
 C. Accurate but not precise
 D. Neither accurate nor precise
22. In inferential statistical testing, it is typical to assume

 A. The null hypothesis is true
 B. The null hypothesis is false
 C. The alternate hypothesis is true
 D. The alternate hypothesis is false
23. In inferential statistical testing, the null hypothesis

 A. Can be proven true
 B. Can be proven false
 C. Can never be proven false
 D. If frequently proven true
24. In inferential statistics, hypothesis testing usually assumes all of the following except

 A. The null hypothesis is true
 B. The test statistic need not conform to a normal distribution
 C. If the null hypothesis is false, the alternate hypothesis must be true
 D. If the null hypothesis cannot be disproved, the two groups are the same
25. To identify the risk factors for postoperative blindness following robotic prostatectomy, the most likely study you will find in a literature search to answer your question will be a

 A. Crosssectional study
 B. Casecontrol study
 C. Cohort study
 D.Randomized control trial
26. A study looking at the upcoming 6month incidence of postoperative cognitive dysfunction in patients who received propofol TIVA and comparing them to agematched controls who did not receive propofol is a

 A. Caseseries
 B. Casecontrol study
 C. Cohort study
 D. Crosssectional study
27. You are trying to determine whether a rapid sequence induction decreases the incidence of aspiration among a randomized group of 100 individuals requiring emergency surgery. The most appropriate statistical test to perform is a

 A. Chisquare test
 B. Student’s ttest
 C. Analysis of variance
 D. Linear regression
28. In order to determine whether a new drug has a significant renoprotective effect on renal function as determined by the serum creatinine level, the most appropriate statistical test to perform is a

 A. Chisquare test
 B. Student’s ttest
 C. Analysis of variance
 D. Linear regression
29. The most appropriate statistical test to determine whether patients who received sevoflurane, desflurane or propofol for maintenance of anesthesia had different wakeup times is

 A. Chisquare test
 B. Student’s ttest
 C. Analysis of variance
 D. Linear regression
30. Which of the following does not require the assumption of data in a normal distribution?

 A. Chisquare test
 B. Student ttest
 C. MannWhitney
 D. KruskalWallis AOV
31. 500 matched subjects were randomized to take preoperative betablockers before surgery and compared with respect to the incidence of postoperative MI. What is the best statistical test to perform to determine whether preoperative betablockers decrease the rate of postop MI given the following data?

 A. ANOVA
 B. tTest
 C. Linear regression
 D. Chisquare
32. Which of the following applies for a Student’s ttest?

 A. The data must be ordinal
 B. The data must be nonparametric
 C. The null hypothesis can never be proved
 D. It must compare two independent groups
33. Sensitivity is the ability to

 A. Find disease if disease were present
 B. Rule out disease if no disease were present
 C. Rule in disease if the test were positive
 D. Rule out disease if the test were negative
34. The sensitivity of the halothanecaffeine contracture test is 97%. This means (MHS = MH susceptible)

 A. With a (+) result, 97% of subjects will be MHS
 B. With a () result, 97% of subjects will not be MHS
 C. 97% of subjects with disease will test (+)
 D. 97% of subjects without disease will test ()
35. A new genetic test for MH was tested in a group of rate of subjects, some of whom were biopsypositive for MH and the rest of whom were biopsynegative. The following results were obtained:
The sensitivity for detecting MH with the new test is:

 A. 100%
 B. 23%
 C. 25%
 D. 0%
36. A new genetic test for MH was tested in a group of rate of subjects, some of whom were biopsypositive for MH and the rest of whom were biopsynegative. The following results were obtained:
The negative predictive value of the test is closest to:

 A. 100%
 B. 75%
 C. 50%
 D. 25%
37. The positive predictive value of a test most reliably increases with an increase in

 A. Sensitivity
 B. Specificity
 C. Prevalence
 D. False positive results
38. The ability of a statistical test to detect a difference if a difference exists is expressed as

 A. Alpha
 B. Beta
 C. 1beta
 D. 1alpha
39. The ability of a statistical test to detect a difference if a difference exists is expressed as

 A. The ability to find a difference if one exists
 B. The error rate in finding a difference
 C. The ability to find no difference if there is none
 D. The error rate in not finding a difference
40. The authors of a study erroneously conclude that a new drug's effects on blood pressure placebo is no different than a placebo, when in fact, they are different (the new drug actually does lower BP more than placebo). Which of the following is necessarily true?

 A. Bias occurred during the study
 B. There was a type 1 error
 C. There was a type 2 error
 D. Alpha was set too low
41. All of the following increase the power of a study except

 A. Increase the sample size
 B. Increase the difference between groups
 C. Decrease the variance
 D. Decrease alpha
42. The relationship between increasing doses of heparin and ACT levels was examined in the presence of dabigatran. The best statistical test to address this question is

 A. Analysis of variance
 B. ChiSquare
 C. Linear regression
 D. Student’s ttest
43. In a trial of the use of a new videolaryngoscope, the success rate of the first 100 attempted intubations was 100% on the first try. Based on this information, there is still a 95% chance that 1 out of ____ times the device will fail.

 A. 100
 B. 50
 C. 33
 D. 0
44. In a trial of the use of a new videolaryngoscope, the success rate of 100 attempted intubations on the first try was 100%. Based on this information, there is a 95% chance that in the future, the success rate for firsttime intubations using the device will be

 A. 100%
 B. 97%
 C. 3%
 D. 0%
45. A recent randomized controlled trial of patients admitted to VA hospitals for total hip arthroplasty showed a beneficial effect of a pneumatic compression device (PCD) on the incidence of postoperative venous thromboembolic disease (VTE). VTE was diagnosed by residents instructed on how to perform bedside venous ultrasonography (VUS). The article concluded that PCDs should be made standard of care to prevent VTE in hospitalized patients. The least likely source of bias in this study is

 A. Selection
 B. Observer
 C. Response
 D. Publication
46. Data collected on the ability of betanatriuetic peptide (BNP) to predict the likelihood of major adverse cardiac events in patients undergoing vascular surgery was plotted in the following receiver operating curve. What are the correct Y and Xaxes, respectively?

 A. Y: sensitivity, X: specificity
 B. Y: specificity, X: sensitivity
 C. Y: 1 sensitivity, X: specificity
 D. Y: sensitivity, X: 1specificity
47. A new cardiac output device is being introduced to the market. Validation studies comparing it to the pulmonary artery catheter have been performed. The best statistical method to reveal difference between the new device and PACderived cardiac outputs is

 A. Pearson correlation coefficient
 B. BlandAltman
 C. Chisquared
 D. Student’s ttest
48. In comparing the results of two different tests, A and B, measuring the same variable, the BlandAltman analysis plots the

 A. Values of A vs values of B
 B. Difference between A and B vs average of A and B
 C. The correlation between A and B
 D. Frequency of A minus B differences
49. Two new spirometers, A and B, are compared in a study of whether they are equivalent in measuring expired tidal volumes. Based on the following BlandAltman plot summarizing the data, differences between A and B

 A. All fall outside of the 95% confidence interval
 B. All fall within the 95% confidence interval
 C. Show a bias in the results of A vs B
 D. Are clinically significant