Bias Flashcards
Placebo effect
Treatment benefit due to perception of receiving treatment
Nocebo effect
Harm effect due to perception of receiving treatment
Hawthorne effect
Change in participant behaviour from being studied
Observer-expectancy bias
researcher’s belief in efficacy of treatment changes their behaviour/actions
Selection bias
Nonrandom difference in the sample from the population
Sampling bias
The sample studied is not representative of populatin results that are being extrapolated to eg healthy volunteers to sick patients.
Loss to follow up bias
The patients who are lost to follow-up are not randomly distributed (i.e., patients experiencing severe side effects were more likely to drop out of study, leaving only those who tolerated treatment).
Exclusion bias
Certain populations (e.g., children) were excluded from a study limiting external validity of results.
Information bias
Caused by flawed collection of information about exposures and outcomes. Mitigated by blinding researchers and participants (see above) and standardizing data collection
Recall bias
Awareness of disorder alters recall by subjects
Confounding
A variable associated with both the disease and the exposure (risk factor), leading to detection of a false relationship between the disease and exposure if the researchers do not control for the confounder. Can be controlled for by adjustment, matching, and/or blinding, but best controlled for by randomization (e.g., ice cream sales increase violence is confounded by hot weather)
Effect modifier (interaction)
A variable that modifies the observed effect of an exposure on disease (e.g., a new drug is effective in female children but not male children, then sex is an effect modifier). Can be controlled for by stratification
