Bias

Posted in Statistics, Total Reads: 800
Advertisements

Definition: Bias

Bias is the systematic deviation from the actual values due to improper method of collecting or presenting the data or due to some faulty design of the estimating technique. It is serious problem because unlike random errors which can be reduced by increasing the sample size and averaging the outcomes, these errors cannot be processed in this way.


Bias can be categorized as:

  1. Selection bias: Here the bias is created because of the selection of a subset of the sampling group more often than others for data collection resulting in inaccurate conclusion.Example of selection bias is the selection of people of a certain race for an experiment.
  2. Spectrum bias: It occurs when the investigated population does not show the similar characteristics of the total population. It may lead for differences in outcomes between investigations.
  3. Omitted-variable bias: By eliminating a variable, the estimated parameters in the model are likely to be biased.
  4. Funding bias: Selecting those samples of data which are favourable to the sponsor of the study or experiment.
  5. Reporting bias: The reporting of certain observations of a study will lead to reporting bias.

Advertisements



Looking for Similar Definitions & Concepts, Search Business Concepts