Bias is a systematic error in the design of a study that results in a mistaken estimate of an exposure's effect on the risk of disease. There are more than thirty specific types of bias.
Sampling is a statistical procedure that is concerned with the selection of individual observations. Ideally, sampling should be random, double-blinded, and representative of the population being studied.
Berkson's Bias | is a type of selection bias that happens when hospitalized patients are used as the control group.
Channeling Bias | occurs when something about the patient like prognosis or severity of illness dictates which group they are assigned to.
Hawthorne Bias | occurs when subjects change their behavior because they know they are being watched.
Information Bias | occurs when information is collected differently between groups.
Interviewer Bias | occurs when the person interviewing participants solicits, records, or interprets information differently.
Late Look Bias | occurs when the data is gathered or analyzed at an inappropriate time.
Lead-time Bias | occurs when one test detects disease earlier so there seems to be increased survival.
Publication Bias | occurs when only positive or favorable results are published.
Pygmalion Effect | occurs when the outcome of the study is affected because the researcher's belief in the treatment.
Recall Bias | occurs when participants have to recall the past (can also be called response, responder, or reporting bias).
Sampling Bias | occurs when the study population isn't representative of the population as a whole. Sampling bias is generally considered a type of selection bias. The main argument is that Sampling bias affects external validity while Selection bias affects internal validity. *
Selection Bias | occurs when there is a systematic difference in the method of choosing study groups such a systematic difference in the enrollment of participants in a study that leads to an incorrect result (e.g., risk ratio or odds ratio) or inference.
Transfer Bias | occurs when there participants are lost to follow up.
Precision (Reliability) is the degree to which research/instruments produce the same results when repeated. This refers to the lack of random error.
Accuracy (Validity) is the degree to which research/instruments actually measure what it says it is supposed to. This refers to the lack of systematic error.
CONFOUNDING occurs when an observed association is due to three or more factors: the exposure of interest, the outcome of interest, and third factor which is INDEPENDENTLY associated with both the outcome of interest and the exposure.
When a third factor is present it DISTORTS the effect of the exposure of interest on the outcome causing an over or underestimation of the true effect.
EFFECT MODIFICATION occurs when the magnitude of effect of the primary exposure on an outcome differs because of the presence of a third variable.