Cognitive Biases and Data Visualization

Cognitive Biases and Data Visualization

Billie Anderson, J. Michael Hardin
Copyright: © 2023 |Pages: 13
DOI: 10.4018/978-1-7998-9220-5.ch076
(Individual Chapters)
No Current Special Offers


This article will provide background on cognitive biases related to data visualizations, with a particular interest towards visual analytics in big data environments. Cognitive biases that appear in visualizations will be discussed. A review of recent studies related to designing experiments to mitigate cognitive biases in data visualizations will be presented. Recommendations, using applied frameworks for detecting and mitigating biases in single visualizations and visual analytic systems, will be provided to practitioners.
Chapter Preview


In 1974, the term cognitive bias was coined by Tversky & Kahneman (1974). Since then, there has been an abundance of research in multiple disciplines such as economics, healthcare, social sciences, and psychology that has been impacted by this term (Featherston et al., 2020). Cognitive biases are errors in judgment or decisions that are made by humans when they attempt to process too much information at a time or when they do not have sufficient information to make correct decisions. Cognitive biases can be found in disciplines from accounting to healthcare. Biases such as overconfidence, anchoring, framing, and confirmation can affect an auditor’s judgement and decision-making abilities (Chang & Luo, 2021). Gopal et al. (2021) developed a checklist to assist medical providers to guard against bias such as mental shortcuts that can lead to errors in diagnosing and treating patients.

Researchers have identified several cognitive biases and there is no doubt that these cognitive biases have led to poor decisions in a multitude of disciplines. Cognitive biases are pervasive across many disciplines. For example, 75% of clinical errors in internal medicine clinical settings are thought to be rooted in cognitive biases (D O’Sullivan & Schofield, 2018). Another study found that it is impossible for cancer patients to make informed choices owing to the treatment provider’s cognitive biases and this is likely to lead to overtreatment (Ozdemir & Finkelstein, 2018). Companies that wish to enter international markets can encounter cognitive biases that, if not overcome, can lead to failure of new product launch (Paul & Mas, 2020).

Data visualization can assist in reducing cognitive biases. Data visualization plays a key role in decision-making process. Visualization allows for data to be consumable, that is, interpretable easily. If data is not consumable, there is a tendency to ignore the facts and rely more on biases. Data visualizations should guard against cognitive biases. However, researchers have found that cognitive biases do exist within data visualizations and can affect decision-making abilities (Padilla et al., 2018). Most recently, there has been interest in designing empirical studies that determine whether cognitive biases can be alleviated (Cho et al., 2017; Dimara et al., 2018; Valdez, Ziefle, & Sedlmair, 2017; Xiong, Van Weelden, & Franconeri, 2019). Some of these studies have provided mitigation strategies to guard against cognitive biases in data visualizations. For example, viewing the data from different positions such as simply reordering the values in a visualization and allowing multiple individuals to critique and provide feedback related to the visualizations (Xiong et al., 2019).

With the advent of big data, practitioners and researchers have become more reliant on data visualizations as a decision-making tool. Telecommunication companies are able to amass large amounts of detailed user data to better understand their customers. Utilities use smart meter data to reduce outages, assign crews, measure energy consumption and meter quality. Governments are able to use data from sensors to monitor road conditions and change the duration of red and green signals of traffic lights according to real-time traffic patterns. With large and complex data, associations, insights, patterns, and errors can be more easily understood with graphical representations than using tables and numbers. The human brain can make sense of pictures more rapidly than tables of numbers. To maximize the impact of big data, it is necessary to incorporate visual analysis at all levels of an organization.

Key Terms in this Chapter

Progressive Visualization: Creating visualizations of big data sets in small pieces. The effect is to create visualizations that display partial results while not sacrificing computing speed.

Information Visualization: A discipline from computer science that deals with efficient methods for displaying data where decision-making is the main objective.

Data Visualization: A graphical presentation of data and information. The subject is applied in a variety of areas such as business, social science, healthcare, and sports.

Data Visualization Literacy: The ability to understand and interpret data that is presented in a visual format.

Data Literacy: The ability to explore, comprehend, and communicate with data.

Cognitive Bias: An unconscious effect that causes humans to make errors in decision-making tasks.

Big Data: Data that is too large to be stored using conventional data storage capabilities. Big data also includes non-traditional types of data such as images, text, weblog searches, and social media data.

Human-Computer Interaction (HCI): The intersection of disciplines such as psychology, information visualization, and artificial intelligence that studies how computers can be more usable to humans.

Complete Chapter List

Search this Book: