There is a phenomenon that there are differences in characteristics between people who are selected into the study sample and people who are not in the sample.It is caused by different selection conditions.For example, the case group selected by the researcher from the hospital has selection bias. Patients know their case history and suspect that illness is related to it, so they seek medical advice.
Chinese name
Selection bias
Foreign name
selection bias
Frequent period
Design phase of the study
Features
Characteristics of exposure or outcome
Nature
Errors caused by differences in characteristics
Example
Bias in admission rate, bias in detection of symptoms, etc
There is a phenomenon that there are differences in characteristics between people who are selected into the study sample and people who are not in the sample.It is caused by different selection conditions.For example, the case group selected by the researchers from the hospital has selection bias. The patients know their exposure history and suspect that illness is related to it, so they seek medical advice.In this way, the probability that the case group entering the hospital has the feature of exposure history will be greatly increased, which will inevitably lead to the fact that the exposure rate of the hospital case group is systematically higher than that of the control group (other patients in the same hospital who do not suffer from the disease to be studied, that is, the patient group unrelated to the exposure factor).The ideal hospital case group should be an unbiased representative (sample) of all patients in the population, but it is actually impossible.Light cases do not seek medical treatment due to long distance, high cost or other reasons, which objectively causes patients to flow to the hospital unevenly, forming a biased sample.Choice bias is naturally formed, not caused by the intervention of researchers' subjective consciousness (bias), and belongs to systematic error.When studying the relationship between exposure factors and diseases (such asCase control study）Selectivity bias can lead to overestimation or underestimation of the effect of exposure factors (contact strength).[1]
give an example
Announce
edit
Errors caused by differences in some characteristics between selected and unselected research objects.
includeBerkson bias (Berkson bias), current case new case bias (Neyman bias), detection syndrome bias, non response bias, time effect bias, etc.Some people divide selection bias into elimination bias and non elimination bias according to whether it can be eliminated by clinical research design.The former refers to the selective bias that can be eliminated through scientific research design, such as the use of randomized methods.Including the traditional classification: admission rate bias, current case new case bias, detection syndrome bias, no response bias, etc;The latter, such as the health worker effect.[1]
