Epidemiological Research Methods

Epidemiological Research Methods 150 150 Affordable Capstone Projects Written from Scratch


Epi Research Methods I






  1. Explain the difference between age and cohort effects.




  1. List the important (and commonly cited) limitation of ecological studies, and explain how it may arise.




  1. Describe the control selection process in two types of case-control studies within a defined cohort.



  1. Discuss one similarity between case-control studies and cross-sectional studies.


  1. Discuss one advantage and one disadvantage of matching.





Fill-in-the-blank Questions

  1. In a ________ study, controls are selected when cases occur; while in a ________ study, controls are selected at baseline.
  2. ­­ A ________ study is especially appropriate to study acute exposures that vary over time and that produce a transient change in risk of an acute condition within a short latency period.
  3. ­ In ________ matching, for each case, one or more controls with the relevant characteristics matching those of the cases were selected from the pool of eligible individuals.
  4. ­­ Matching on (naturally) ________ variables is relatively straightforward, while matching on ________ variables can be difficult and may result in residual confounding.


Matching Questions

  1. For each of the studies below, put an “A” if it is an example of analytic epidemiology or a “B” if is an example of descriptive epidemiology.


  1. Analysis of US mortality data to determine the rates of colorectal cancer deaths for white (non-Hispanic), black (non-Hispanic) and Hispanics in 2012.
  2. Analysis of survey and laboratory data to determine if individuals with a high salt diet have higher systolic and diastolic blood pressure.
  3. Case-control study to determine if patients with a current diagnosis of breast cancer had ever received hormone replacement therapy.
  4. Analysis of medical records to determine if diagnosis of hypertension is more or less common among adults age 65 or older compared to those younger than age 65.


  1. For each of the ecological study variables listed below, put an “A” if it is an aggregate measure, “B” if it is an environmental measure, and “C” if it is a global measure.


  1. Median daily salt intake.
  2. Population density.
  3. Daily high/maximum temperature.
  4. Existence of handgun registration and licensing laws.
  5. Proportion of insured individuals.
  6. Total inches of rainfall.


True or False Questions

  1. True or false? Birth cohort effects are primarily due to the circumstances surrounding the time of birth of a given cohort of individuals.
  2. True or false? Ecologic studies are imperfect surrogates for studies in which individuals are the observation units, and ecologic studies should only be seen as preliminary studies.
  3. True or false? Nested case-control studies, rather than case-cohort studies, utilize risk-set sampling to select controls.
  4. True or false? In case-crossover studies, all fixed individual characteristics that might confound the association of interest are controlled for.
  5. True or false? The practice of matching is particularly common and useful in the context of case-control studies when trying to make cases and controls as similar as possible with regard to potential confounding factors.

Short Answer Questions

  1. Explain the difference between incidence and prevalence.




  1. Discuss the assumption that explains why one subtracts one half of the total number of censored observations from the denominator when calculating cumulative incidence.




  1. List and explain two assumptions needed for the estimation of cumulative incidence based on survival analysis.





  1. The calculation of incidence based on grouped data is typically used to estimate mortality or incidence for a population or an aggregate defined by residence in a given geographic area over some time period. Name and define the term commonly used for these aggregates, and provide an example.


  1. To use the person-time approach to take into account changing exposures involves an assumption akin to that used in crossover clinical trials. Describe this assumption and discuss how if this assumption is valid when studying smoking as an exposure.


Fill-in-the-blank Questions

  1. For both prevalence and incidence, it is necessary to have a clear definition of the outcome as an ________, typically defined as the occurrence of any phenomenon of disease or health that can be discretely characterized.


  1. ­­ Rather than individuals, the denominator for the ________ is formed by time units (t) contributed to the follow-up period by the individuals at risk (n).


  1. ­ As a descriptive measure, ________ prevalence is a useful index of the magnitude of current health problems and is particularly relevant to public health and health policy.


  1. ­­ When estimating the cumulative incidence based on survival analysis, the necessary assumption of ________ may not be met in the presence of birth cohort and period (calendar time) effects.



Matching Questions

  1. For each of the following situations, put an “A” if it is likely to result in the overestimation of the cumulative incidence estimate, a “B” if is likely to result in the underestimation of the cumulative incidence estimate, a “C” if the direction of the bias is variable, or a “D” if no bias is likely to result.


  1. Study subjects withdrew because they were sicker over time, leading to a higher risk for censored than for noncensored observations.
  2. Healthy subjects were more likely than sick subjects to refuse to allow subsequent follow-up contacts.
  3. Study subjects migrate out of the study.
  4. After five years of follow-up, all remaining study participants are administratively censored.


  1. Fill in the following equation for calculating point prevalence, where I = Incidence, D = Duration, and P = Point Prevalence.


Point Prevalence = (a) x (b) x (1 – c)


True/False Questions

  1. True or false? The censored observations represent an identified event when calculating cumulative incidence based on the Kaplan-Meier approach.


  1. True or false? The assumption that events and losses occur uniformly within intervals is required for the Kaplan-Meier approach.


  1. True or false? When withdrawals (and additions in an open population or dynamic cohort) and events occur uniformly, rate (based on grouped data) and density (based on individual data) are virtually the same.


  1. True or false? The upper limit of values for a rate or a density may exceed 100%, whereas values for probabilities cannot be greater than 100%.


  1. True or false? The cumulative probability of an event (or the cumulative survival) has no time period intrinsically attached to it: time must be specified.


Short-Answer Questions


  1. Explain why it is necessary to use survival analysis techniques to estimate a measure of association in a case-cohort study.



  1. Discuss “built-in-bias” and explain how it affects the relationship between the odds ratio and the relative risk.





  1. Define “attributable risk” and explain the interpretation of this measure of association.






  1. Explain the “rarity assumption” and why it is necessary to estimate the relative risk in case-control studies.





  1. Discuss one (analytical) advantage of using a case-cohort approach to measure an association.




Fill-in-the-blank Questions

  1. When the condition of interest has a high incidence and when prospective data are available, the common practice is to report the ________ because it is a more easily understood measure of association between the risk factor and the outcome.


  1. Measures based on ________ differences are often preferred when public health or preventive activities are contemplated. In contrast, etiologic studies that are searching disease determinants (causes) usually rely on ________ differences in the occurrence of discrete outcomes.


  1. The ________ of an event is the exact reciprocal of the ________ of the nonevent.