Difference between revisions of "Bias"
| Line 5: | Line 5: | ||
==Information bias== | ==Information bias== | ||
| − | Information bias results from errors in measurement or classification of exposures or outcomes of interest amongst the individuals included in the study. In the case of analytic studies, this may be differential or non-differential:<br> | + | Information bias results from errors in measurement or classification of exposures or outcomes of interest amongst the individuals included in the study. In the case of analytic studies, this may be classified as differential or non-differential:<br> |
* Differential bias occurs when the chance of bias is different for the different groups being compared. For example, in a study investigating whether Boxer dogs have a higher incidence of mast cell tumours than other breeds, Boxers may be more likely to be diagnosed by vets as having these, due to a postulated breed disposition. | * Differential bias occurs when the chance of bias is different for the different groups being compared. For example, in a study investigating whether Boxer dogs have a higher incidence of mast cell tumours than other breeds, Boxers may be more likely to be diagnosed by vets as having these, due to a postulated breed disposition. | ||
* Non-differential bias occurs when the chance of bias is not affected by the group the individuals belong to. | * Non-differential bias occurs when the chance of bias is not affected by the group the individuals belong to. | ||
| + | Non-differential bias will tend to reduce the strength of any association present, and will increase the probability of a [[Random variation#Hypothesis testing and study power|type II error]], whereas the effect of differential bias cannot be predicted. | ||
| + | |||
[[Category:Veterinary Epidemiology - Introduction|I]] | [[Category:Veterinary Epidemiology - Introduction|I]] | ||
Revision as of 17:21, 10 December 2010
Systematic error, or 'bias' is of particular importance in any epidemiological investigation, and should be avoided wherever possible. Biases will reduce the validity or accuracy of any results obtained. There are a number of types of bias, which may be classified broadly as either selection bias or information bias, and which will differ in the case of different study designs.
Selection bias
Selection bias results in the selection of a study sample which is not representative of the source population, and so affects inclusion of individuals in the study.
Information bias
Information bias results from errors in measurement or classification of exposures or outcomes of interest amongst the individuals included in the study. In the case of analytic studies, this may be classified as differential or non-differential:
- Differential bias occurs when the chance of bias is different for the different groups being compared. For example, in a study investigating whether Boxer dogs have a higher incidence of mast cell tumours than other breeds, Boxers may be more likely to be diagnosed by vets as having these, due to a postulated breed disposition.
- Non-differential bias occurs when the chance of bias is not affected by the group the individuals belong to.
Non-differential bias will tend to reduce the strength of any association present, and will increase the probability of a type II error, whereas the effect of differential bias cannot be predicted.