October 2021 Summaries

Cognitive biases, which are systematic errors in thinking often influenced by cultural and personal experiences, can significantly distort decision-making processes even when working with data. Despite the common belief that data provides an objective foundation for decisions, biases can still influence how data is interpreted, leading to unexpected outcomes. This text highlights various types of data biases, such as confirmation bias, selection bias, historical bias, survivorship bias, availability bias, and outlier bias, all of which can skew perceptions and analyses. These biases can also impact machine learning models, as they inherit the biases of their creators, leading to problematic outcomes. To mitigate these biases, it is crucial to be aware of them, employ strategies like randomization, inclusivity frameworks, and comprehensive data analysis, and remain open to alternative perspectives. Recognizing and addressing these cognitive biases is essential for making more informed and accurate decisions with data.