Lecture 8: Non-random error (bias)

Question 1

What is a non-random error in epidemiological studies?

Answer 1

A non-random error occurs due to poor study design, poor study processes, or poor study measurement, and is also referred to as bias.

Question 2

What is the significance of study validity?

Answer 2

A study with only a small amount of random or non-random error is considered to be a valid study.

Question 3

What does RAMBOMAN stand for in assessing non-random error?

Answer 3

RAMBOMAN stands for:
R - Recruitment
A - Allocation
M - Maintenance
B - Blind
O - Objective
M - Measurement
A - Analyses
N - Analyses

Question 4

What factors are considered in the Recruitment phase of RAMBOMAN?

Answer 4

Factors include whether the participants are a representative sample from a relevant population, if the setting is well described, and if the eligible population is well described.

Question 5

What are the two main ways to allocate study participants to exposure and comparison groups?

Answer 5

The two main ways are by measurement (or observation) and by random allocation.

Question 6

What does the Maintenance phase assess in RAMBOMAN?

Answer 6

The Maintenance phase assesses whether participants remained in their allocated groups and if many participants were lost to follow-up.

Question 7

How does measurement affect the validity of study results?

Answer 7

The validity of study results can be affected by how well exposures and outcomes are measured, including potential biases in measurement.

Question 8

What is confounding in epidemiological studies?

Answer 8

Confounding occurs when the exposure is mixed with another factor that is also associated with the outcome, leading to bias.

Question 9

What does B in RANBOMAN stand for?

Answer 9

Blind
Doctors may not have been blind to the drinking behavior of participants, leading to biased coding of the cause of death.

Question 10

What is a double blind Randomised Controlled Trial?

Answer 10

A study design where neither the participants nor the researchers know who is receiving the treatment or placebo, reducing bias.

Question 11

How can confounding be addressed in a study?

Answer 11

By stratifying the study into sub-studies so that participants with the confounder are all in one sub-study.