Fairness in data analysis
The data analysis process is a very delicate one just as data analysis as a whole is also as sensitive. Data-driven decision-making affects all aspects of our lives: in business, finance, health, tech, and so on. If the wrong decisions are taken, adverse effects entail and that is why it is very important to ensure that our data is very fair and does not reinforce bias.
Fairness in data analysis means using data in a way that does not create or reinforce bias.
There are different types of bias and we should be mindful so that it doesn't affect our data analysis process. The types of bias are as follows:
Sampling Bias: This type of bias involves the use of a sample that is not representative of the population. A fault in the way you choose your samples may lead to sampling bias. As a result, you systematically exclude a subset of your data due to a particular attribute. It is important to keep in mind that almost every component of both quantitative and qualitative surveys carry a risk of sampling bias. As a result, both the survey creator and the responders may make easy work of identifying its origins.
Observer Bias: This is the tendency for different people to observe things differently. When a researcher's expectations, viewpoints, or prejudices affect what they observe or record in a study, this is known as observer bias. Studies, where observers are aware of the research aims and hypotheses, are frequently affected by it. Detection bias is another name for observer bias.
Interpretation Bias: This is the tendency to always interpret ambiguous situations in a positive or negative way.
Confirmation Bias: This is the tendency to search for or interpret information in a way that confirms pre-existing beliefs. The tendency to process information by looking for or interpreting information that is consistent with one's previous opinions is known as confirmation bias. This biased approach to decision making is sometimes unintended and results in the rejection of contradictory facts.