DSS Lection 11 Cognitive Biases

A set of vocabulary flashcards covering the definitions, types, and impacts of cognitive biases in the context of data visualization and decision-making.

Cognitive bias

A systematic error in thinking that occurs when people process and interpret information in their surroundings, influencing their decisions and judgments.

Heuristics

Rules of thumb that help individuals make sense of the world and reach decisions with relative speed, often used in conditions of uncertainty or complexity.

Cognitive Load Scarcity

The state where human cognitive resources are limited due to factors like limited working memory, limited ability to carry out complex algorithms, and a lack of readily-accessible knowledge.

Task-based Taxonomy (Dimara E. et al.)

A classification of $154$ cognitive biases organized by the experimental tasks in which they occur, such as estimation, decision, and hypothesis assessment.

Availability bias

A bias occurring when people estimate the importance of data that is vivid or visually memorable, or when events are perceived as more probable if they are easy to remember.

Confirmation bias

A tendency where people focus on charts that support what they already believe while ignoring conflicting evidence.

Anchoring effect (Anchoring bias)

The tendency for individuals to base a significant portion of their decisions on the first piece of information they receive, which acts as a cognitive reference point.

Framing effect

A cognitive bias where a person's decision depends more on how the information is presented (e.g., $95\%$ survival rate vs. $5\%$ mortality rate) than on its actual content.

Truncation bias

The manipulation of an axis scale to exaggerate differences, such as starting a bar chart at $90$ instead of $0$.

Recency bias

An error where recent data receives more attention than long-term trends, such as ignoring a five-year growth pattern because of a drop in the current month.

Overconfidence bias

A situation where users assume they understand a visualization better than they actually do, often leading to false conclusions about correlation and causation.

Apophenia (Pattern recognition bias)

The human tendency to see meaningful patterns in random or noisy data where no statistically significant relationship exists.

Survivorship bias

A bias where the visualization includes only successful cases and ignores failures, leading to an incomplete analysis.

Base-rate neglect

A cognitive error in which individuals ignore overall proportions and focus only on highlighted or specific cases.

Salience bias (Color bias)

The phenomenon where bright colors attract attention disproportionately, causing viewers to focus on specific sections regardless of their data percentage.

Size bias

The perception that larger visual elements are more important or influential, even if the underlying value is not proportionally higher.

Selective attention bias

A tendency for users to notice only visually prominent information, such as one large green KPI, while ignoring surrounding metrics.

Loss aversion

A principle where negative changes feel more significant to a viewer than equivalent gains.

False precision bias

The illusion of certainty created by showing exact numbers (e.g., $87.43\%$) that suggest more accuracy than the actual measurement supports.

Decoy effect

A phenomenon in decision-making where the introduction of a third option (the decoy) influences a user's preference between two other options.

Debiasing

Methods and strategies used to reduce subjective or emotional influences and mitigate the risk of wrong decisions by making users aware of cognitive biases.

Rainbow colormaps

Colormaps that can trick people into seeing false patterns, create false associations, or cause data differences to appear artificially smaller or larger.

Sequential colormaps

Recommended colormaps that visually preserve relative data magnitudes and support more accurate insights into smoothly varying datasets.

APA Graph Recommendations

Design standards suggesting a $4:3$ aspect ratio for histograms and bar graphs and requiring that all ordinates start at $0$.

3D effect distortion

A misleading graphical element where projection in a $3D$ pie or bar chart causes certain sections to appear larger than the data supports or occludes internal values.