In simple terms, bias is something that acts as a prejudice either in the favor or against a person, a thing or an opinion. All our lives are biased in some way or the other. You can witness bias in everything of your life, whether it’s your personal or professional life. Mathematical statistics is an important branch of math, as well as science. Statistics also has various different types of bias involved in its calculation and computations. We shall focus on all of these biases in detail here.
Selection Bias: the very first form of bias observed in statistics is the selection bias. When you are doing any form of statistical analyses, you are bound to take samples for references to come to your conclusions. And the sample that you take might have error because of the involvement of a particular type or group of people in it. This can create problems in the end conclusions. The selection bias has to be always taken into account of the statistical analyses to get adequate results.
Bias of an Estimator: the next form of bias occurs when there exists a significant difference between the expectation of the estimator and the true value of the parameter. This type of bias can cause major errors in the entire study because the basis itself becomes deformed. Hence, it is extremely crucial that the difference between the two factors remains zero for accurate predictions and calculations.
Null Hypothesis Bias: when you are doing a hypothesis testing with statistics, there can be situations where the data gets completely biased. One such scenario is occurs when in the case of true null hypothesis, the probability of rejecting the null hypothesis becomes less than or equal to the significance. Such bias is also known to occur when the probability of rejecting of the null hypothesis is greater than or equal to the significance level in situations where the alternative null hypothesis is true.
Detection Bias: there is another common type of bias observed in statistics and that is usually known as the detection bias. At times there are certain factors that affect the study or the characteristics of a particular set of study objects, thus leading to different problems and troubles.
These are some of the major bias observed in different statistical techniques and studies. Other than these, some other common forms of bias like the funding bias, reporting bias and data-snooping bias are also highly prevalent and observed in statistics. These types of bias often lead to certain problems and errors in the entire process of calculation and analysis and hence should be avoided. You students need to ensure that you do everything possible to not allow the bias to creep into your studies.