Bias
- A sample is biased if it’s differs from a population in a systematic way
- That can result in a statistics that’s consistently larger or smaller
Types of Biases
- Selection bias - when you systematically exclude or under-represent a part of population
- Measurement/Response bias - when data is collected with systematic error
- Non-response bias - when responses aren’t obtained from all individuals selected for inclusion in sampling
The bias of an estimator is the difference between an estimator’s expectations and the true value of the parameter being estimated.
In statistical hypothesis testing, a test is said to be unbiased when the probability of committing a type I error is less than the significance level, and that of getting a true positive (rejecting the null hypothesis when the alternative hypothesis is true) is at least that of the significance level.
Data-snooping bias comes from the misuse of data mining techniques.
https://en.wikipedia.org/wiki/Bias_%28statistics%29 https://en.wikipedia.org/wiki/Bias_of_an_estimator http://stats.stackexchange.com/questions/108998/what-is-statistical-bias