What is a confounding factor?
An additional factor that obscures the relationship between an exposure and the outcome.
What are the characteristics of confounding factors?
They are associated with both the exposure and the disease.
What is randomisation in the design stage of an experiment?
A method that aims to produce an even amount of potential risk factors in two populations.
Define Simple Randomisation.
Each participant has an equal chance of being assigned to any of the study groups.
What is a potential drawback of Simple Randomisation?
It may lead to unequal group sizes in smaller studies.
Describe Block Randomisation.
Participants are divided into blocks, and within each block, they are randomly assigned to different groups.
What does Block Randomisation ensure?
That group sizes remain balanced throughout the study.
What is Stratified Randomisation?
Participants are first stratified based on specific characteristics, then randomised within each stratum.
What is the purpose of Stratified Randomisation?
To ensure confounding variables are evenly distributed across groups.
What is Cluster Randomisation?
Entire groups or clusters are randomised rather than individual participants.
When is Cluster Randomisation particularly useful?
When interventions are applied at the group level.
Define Adaptive Randomisation.
The probability of assignment to a group changes based on accumulated data during the trial.
What is the goal of Adaptive Randomisation?
To assign more participants to the more effective treatment.
What is Minimisation in the context of randomisation?
An adaptive method that allocates participants to groups to minimize imbalance in confounding variables.
How does Minimisation work?
It considers the characteristics of participants already assigned when determining the group for a new participant.
What does restriction entail in experimental design?
Limiting study subjects to specific characteristics to avoid confounding.
Provide an example of restriction.
Ensuring all subjects are males between the ages of 40-50 in a study about heart disease.
What is matching in experimental design?
An active form of control to ensure study groups do not differ with respect to possible confounders.
How can matching be implemented?
By enrolling participants with similar characteristics, such as age and gender.
What is stratification in the analysis stage of an experiment?
A statistical technique that creates categories where the confounding variable does not vary much.
What are multivariate models used for in experimental analysis?
To control for confounding using methods like logistic regression and ANCOVA.
