Bias In Research Flashcards
What is bias?
Systematic error in results
What is imprecision?
Random error
What is selection bias?
Systematic differences in selecting groups for comparison
For example, A study investigating treatment effect on
risk of diabetes
– Individuals with greater body mass index (obesity)
given placebo
– Individuals with lower body mass index given
treatment
What is performance bias?
Systematic differences in the treatment of groups, other than the interventions of interest
Individuals in a trial are made aware of whether they got treatment or placebo, also treated differently by the researchers depending on the group they’re in.
What is detection bias?
Systematic differences between groups in determining outcomes
In a trial, the researchers determine the outcome are aware of which participants received treatment and placebo.
What is attrition bias?
Systematic differences between groups in withdrawal from a study.
For example in a clinical trial, those participants for whom the intervention does not work are more likely to withdraw, so the ones remaining in the study are more likely to have had a positive response.
What is reporting bias?
Systematic differences in the reporting of results from a trial.
For example in a clinical trial only the positive findings are reported, null or negative effects are omitted.
Describe a way of limiting selection bias in clinical trials.
Using a random number generator to ensure participants and ‘randomised’
How can performance bias be limited in clinical trials?
‘Blinding’ of participants and researchers
How can detection bias be limited?
‘Blinding’ of researchers
How can attrition bias be limited in clinical trials?
Clarify reasons for loss to follow up and include these data
How can reporting bias be limited in clinical trial?
Report all results outlined in the pre-specified analysis plan
How can bias be dealt with in research?
implanting what?
Randomisation
Blinding
Controlled groups
What are the potential bias’ in randomisation procedures?
Selection bias - when allocation is predictable
Accidental bias - imbalance of risk factors
Which types of randomisation are used in research?
Simple randomisation
Blocked randomisation
Stratified randomisation
