18 Oct Failing to account for age cohort effects in smoking prevalence may:

Failing to account for age cohort effects in smoking prevalence may:

 

  obscure the fact that there had been a shift in the age of onset for lung cancer toward earlier ages as well as the differences in smoking prevalence for males and females, and by level of education.

 

  obscure the fact that older cohorts had higher prevalence of smoking in comparison to younger cohorts.

 

  obscure the fact that there had been a shift in the age of onset for lung cancer toward earlier ages.

 

  obscure the differences in smoking prevalence for males and females, and by level of education.

 

Question 2A five-year prospective cohort study has just been completed. The study was designed to assess the association between supplemental vitamin A exposure and mortality and morbidity for measles. The RR for incidence of measles was 0.75, and the RR for measles mortality was 0.5.

 

Which statement is correct?

 

  One of the problems that this study may have faced is individuals lost to follow-up during the five-year period.

 

  Supplemental vitamin A increases the incidence of measles.

 

  A cohort study is not an appropriate study design in this case because the association between one exposure and two different outcomes is being considered.

 

  A cohort study is not a good design to study this association because measles is a very common disease.

 

  None are correct.

 

Question 3An example of passive follow-up in cohort studies is: 

 

  telephone calls to study participants.

 

  collection and maintenance of data by federal and state governments.

 

  use of written invitations to return to the study site.

 

  use of follow-up mailings.

 

Question 4Subjects for an exposure-based cohort study of lead exposure would be selected most appropriately from:

 

  certain occupational groups such as workers who manufacture batteries.

 

  male Harvard alumni from 1916 to 1950.

 

  All are correct.

 

  the residents of a large U.S. county.

 

Question 5Cohort study is to risk ratio as: 

 

  ecologic fallacy is to cross-sectional study and case-control study is to odds ratio.

 

  genetics is to environment.

 

  case-control study is to odds ratio.

 

  ecologic fallacy is to cross-sectional study.

 

Question 6A major advantage of cohort studies over case-control studies with respect to the role of a suspected factor in the etiology of a disease is that: 

 

  they permit direct determination of risk of disease in those exposed to the suspected factor.

 

  it is easier to obtain controls not exposed to the suspected factor.

 

  they take less time and are less costly.

 

  they can utilize a more representative population.

 

  they can be done on a double-blind basis.

 

Question 7A case-cohort study differs from a nested case-control study by:

 

  not using controls.

 

  not using matching.

 

  None of these is correct.

 

  inability to study occupational exposures.

 

Question 8Practical considerations in the design of cohort studies do not usually include:

 

  availability of exposure data.

 

  size and cost of the cohort.

 

  follow-up issues.

 

  scientific justification.

 

  age of the investigator.

 

Question 9As an epidemiologist, you are going to investigate the effect of a drug suspected of causing malformations in newborn infants when the drug in question is taken by pregnant women during their pregnancies. As your sample you will use the next 200 single births occurring in a given hospital. For each birth, a medication history will be taken from the new mother and from her doctor; in addition, you will review medical records to verify use of the drug. (Note: These mothers are considered to have been followed prospectively during the entire course of their pregnancies, because a complete and accurate record of drug use was maintained during pregnancy.) The resultant data are: Forty mothers took the suspected drug during their pregnancies. Of these mothers, 35 delivered malformed infants. In addition, 10 other infants were born with malfunctions.

 

The relative risk between exposure to the drug and malformations is: 

 

  12

 

  18

 

  10

 

  14

 

  16

 

Question 10As an epidemiologist, you are going to investigate the effect of a drug suspected of causing malformations in newborn infants when the drug in question is taken by pregnant women during their pregnancies. As your sample you will use the next 200 single births occurring in a given hospital. For each birth, a medication history will be taken from the new mother and from her doctor; in addition, you will review medical records to verify use of the drug. (Note: These mothers are considered to have been followed prospectively during the entire course of their pregnancies, because a complete and accurate record of drug use was maintained during pregnancy.) The resultant data are: Forty mothers took the suspected drug during their pregnancies. Of these mothers, 35 delivered malformed infants. In addition, 10 other infants were born with malfunctions.

 

Suppose that a relative risk of 0.5 has been found. Which interpretation might follow? 

 

 Use of the drug is a protective factor for congenital malformations.

 

  The risk of congenital malformations is half as high among users of the drug as among nonusers, and use of the drug is a protective factor for congenital malformations.

 

  The risk of congenital malformations is twice as high among users of the drug as among nonusers of the drug.

 

  The risk of congenital malformations is half as high among users of the drug as among nonusers.

 

  The risk of congenital malformations is equal among users and nonusers of the drug.

 

Question 11Which of the following is least affected by outliers?

 

  The standard deviation

 

  The mean

 

  The range

 

 

 

  The median

 

Question 12Which of the following tests whether variances are homogeneous?

 

  Bartlett’s test

 

  Neither Levene’s nor Bartlett’s is correct

 

  Levene’s test

 

  Both Levene’s and Bartlett’s are correct

 

Question 13Which of the following are assumptions underlying the use of parametric tests (based on the normal distribution)?

 

  The data should be normally distributed.

 

  The data should be at least interval level.

 

  All are correct.

 

  The samples being tested should have approximately equal variances.

 

Question 14Which of the following is not a transformation that can be used to correct skewed data?

 

  Log transformation

 

  Reciprocal transformation

 

  Square-root transformation

 

  Tangent transformation

 

Question 15What does independence of data mean?

 

  That scores in one condition are free from influences from other conditions.

 

  That independent researchers must collect the data.

 

  That scores from one participant are free from influences from other participant.

 

  That we must never collect two sets of data from one person.

 

Question 16I collected some data about how much buyers of my book liked it (on a scale of 1 = it’s utter rubbish to 10 = I never read anything else). I ended up with a sample of 15,467 people. When I looked at the distribution, I found a skew of 1.23 (SE = 0.65). The mean rating was 4.78.

 

Which of the following would be the best way to decide whether the skew is problematic?

 

  None are correct because of the large sample size..

 

  See if the z-score is bigger than 1.96 or smaller than –1.96.

 

  Use the Kolmogorov–Smirnov test.

 

  See if the skew is significant at p < .05.

 

Question 17 If a distribution is multimodal, what does this mean?

 

  It will not be a normal distribution.

 

  The data have been entered incorrectly.

 

  It will be a normal distribution.

 

  It will have to be checked with a Levene’s test.

 

Question 18Which of the following is not an assumption of a parametric test?

 

  Homogeneity of variance

 

  Independence

 

  Dependence

 

  Normally distributed data

 

 Question 19Why are z-scores used to check for outliers?

 

  They standardize scores in order to convert them to values closer to the mean.

 

  They standardize scores for a known mean and standard deviation, allowing comparison.

 

  They allow you to allocate letters for missing values.

 

  A z-score is an outlier.

 

Question 20FCEs by age for snoring (R06.5) in 2006-2007

 

‘Finished consultant episodes (FCEs)’ on y-axis. ‘Age (years)’ on x-axis.

 

Source: NHS Digital UK (Links to an external site.)

 

What do you deduce about this distribution?

 

  The incidence of snoring is greatest among females during middle to old age.

 

  Snoring is equally distributed between women and men.

 

  The incidence of snoring is greatest in early childhood and again during middle age.

 

  None are correct.

 

Question 21The covariance is:

 

  An unstandardized version of the correlation coefficient.

 

  Dependent on the units of measurement of the variables.

 

  A measure of the strength of relationship between two variables.

 

  All are correct.

 

Question 22Correlational studies allow the researcher to:

 

  Identify the relationship between two variables.

 

  Make causal inferences about the relationship between two variables.

 

  Predict the effect of one variable upon another.

 

  Test for differences between two variables.

 

Question 23The correlation between two variables A and B is .12 with a significance of p < .01.

 

What can we conclude?

 

  That variable A causes variable B.

 

  That there is a substantial relationship between A and B.

 

  All are correct.

 

  That there is a small relationship between A and B.

 

Question 24Which of the following is not a transformation technique in SPSS?

 

  Reciprocal transformation

 

  Log transformation

 

  Unequal transformation

 

  Square-root transformation

 

Question 25When interpreting a correlation coefficient, it is important to look at:

 

  All are correct.

 

  The sign of the correlation coefficient.

 

  The magnitude of the correlation coefficient.

 

  The significance of the correlation coefficient.