Module 3

Question 1

What is a confounder?

Answer 1

They must meet two criteria: they differ in levels/categories of X, and they have a casual relationship with Y

Must be related to the predictor and the outcome.

Question 1

What is Simpson’s Paradox?

Answer 1

A statistical phenomenon where an association between two variables in a population emerges, disappears or reverses when the population is divided into subpopulations (when you control for a confounder).

Question 2

How do we account for confounders?

Answer 2

1. Control for confounding variables, conduct the comparison of interest within confounding groups.
2. Design an experiment where the treatment and control groups are “otherwise similar” by design.

Question 3

What are three practices of reproducibility in research?

Answer 3

1. manual of operations: write down exactly what you plan to do before beginning study
2. reproducible analysis: make your data and data analysis plan available so someone else could reproduce your findings
3. archiving of process, data, and findings: give others access to your study protocol, data, and results

Question 4

What’s the acronym for designing and implementing a research study

Answer 4

QuODDID

Question
Outcome
Design
Data analysis
Interpretation
Dissemination

Question 5

What makes a good (Q)uestion of scientific interest?

Answer 5

Relevant: Is the question relevant to the topic of interest?

Precise: Is the population of interest specified? Is the relationship of interest specified?

Testable by the study: Can we actually answer the question with our study?

Question 6

What are desirable characteristics for (O)utome measures?

Answer 6

relevant, precise and accurate, movable

Question 7

What are good characteristics of (D)esign?

Answer 7

Are we using an experimental or observational study? Are we using parallel groups or a cross-over study?

Question 8

What is a parallel group study?

Answer 8

Subjects allocated to receive only one “level: of the treatment being compared. Make comparison groups similar (via randomization)

Question 9

What is a cross-over study?

Answer 9

Each subject gets multiple “levels” of the treatment. Comparison of outcomes is within experimental units (people), and each unit acts as “own control/comparator