Where does heuristic collide?

In the study of judgment, a heuristic is a quick mental rule that people fall back on to form a judgment or make a choice when they lack the time or information to reason it through fully. Tversky and Kahneman's 1974 Judgment under Uncertainty: Heuristics and Biases described how everyday judgments lean on shortcuts such as representativeness and availability, which usually work but can push an answer in a predictable wrong direction. The referent here is an automatic inference process in a human mind, studied through the errors it produces, not a cost estimate inside a search program, a way of teaching, or a checklist for inspecting a screen.[1]

Unit

One judgment strategy a person applies, named by the cue it relies on

Recovery means

Not applicable

Measured by

Studied experimentally through the systematic biases and error patterns it produces in people's judgments

In artificial intelligence, a heuristic is a computed estimate of how far a state is from a goal, used to steer a search toward promising options without checking every path. Hart, Nilsson, and Raphael's 1968 A Formal Basis for the Heuristic Determination of Minimum Cost Paths set out the A* algorithm and showed that when the estimate never overshoots the true remaining cost, called an admissible heuristic, the search still finds a lowest-cost route. The referent is a numerical function evaluated by a machine, judged by whether it speeds the search and preserves the best answer, and it has nothing to do with a person's snap judgment, a teaching approach, or a usability guideline.[2]

Unit

One estimate of the remaining cost from a state to the goal

Recovery means

Not applicable

Measured by

Assessed by whether it is admissible and by how many states the search must expand to reach the goal

In education, the heuristic method is teaching that leads learners to discover a principle for themselves through guided problem-solving rather than being told it directly. Polya's 1945 How to Solve It laid out a set of general problem-solving moves, such as working backwards or solving a simpler related problem, that a teacher can prompt students to try so they find their own way to a solution. The referent is a way of teaching and a set of consciously applied strategies for discovery, judged by what learners come to understand, not an automatic mental shortcut, a machine's cost estimate, or a rule for inspecting an interface.[3]

Unit

One guided problem-solving move a learner is prompted to try

Recovery means

Not applicable

Measured by

Judged by learners' problem-solving progress and by what they come to understand for themselves

In usability engineering, a heuristic is a broad principle of good interface design that an evaluator uses to inspect a product and flag where it falls short. Nielsen and Molich's 1990 Heuristic Evaluation of User Interfaces introduced a method in which a few reviewers check an interface against a short list of such principles, for example that a system should keep the user informed of its status, and record the problems they find. The referent is an explicit design guideline applied in expert review, counted by the usability problems it uncovers, and it is not a person's judgment shortcut, a search program's cost estimate, or a way of teaching.[4]

Unit

One design principle checked against an interface

Recovery means

Not applicable

Measured by

Counted by the number and severity of usability problems reviewers uncover when applying it