Robertson2017-analogy

Transfer Studies of Analogical Problem Solving

Transfer of learning often involves analogical transfer, particularly in ill-defined problems.
Analogy: A mechanism for structured comparisons of mental representations (Holyoak, 2012).
Analogical thinking involves transferring knowledge by mapping correspondences between different situations (Gick & Holyoak, 1983).

Duncker's Radiation Problem (1945): A doctor must destroy a malignant tumor with rays without harming healthy tissue.
- High intensity destroys both tumor and healthy tissue.
- Low intensity does not affect tumors.
Solution Structure: Similar to the Fortress problem where an army splits and converges simultaneously.
Solution for Radiation Problem: Use multiple lower-intensity rays to converge on the tumor.

Investigated the influence of previous analogous problems (Fortress problem) on solving the Radiation problem.
Results showed various suggested solutions based on the solution types presented in the Fortress problem.
- Different strategies (e.g., open supply route, tunnel) led to varied responses for the Radiation problem.
Divide and Converge Solution: Effective strategy derived from the Fortress problem, pertaining to reduced ray intensity and multiple machines.
Notable statistic: Without hints, only 10% in control group solved analogically; with hints, 75-80% solved using the provided analogy.

Accessing Past Experience: Major challenge present when problems are from different domains (e.g., military vs. surgical).
Cross-domain analogies can be fruitful but require effective retrieval strategies.
Example: Checklist systems from aviation implemented in healthcare to reduce errors (WHO Surgical Safety Checklists).

Mapping involves establishing a structural alignment between problems.
Requires matching elements that play similar roles (e.g., general to surgeon, army to rays).
Candidate inferences projected from base (source problem) to target (current problem), allowing knowledge transfer.

Adapting a past solution to a current problem can be challenging.
Difficulties can include pre-existing knowledge biases or unnecessary constraints imposed on problem statements.
Example: Misinterpreting resources available (e.g., number of ray machines).

Identities: Elements that are the same in both problems.
Indeterminate Correspondences: Elements yet to be matched or clear inferences.
Structure-Preserving Differences: Surface differences that do not alter the solution structure.
Structure-Violating Differences: Changes that impact the foundational solution.

The more semantically similar the objects in different problems, the more likely an analogy will be utilized effectively.
Different analogies (e.g., beams of light vs. armies attacking) signal different approaches to problem-solving.

Requires deep-level comparison and similarity recognition of goals and relationships.

Recognizing structural similarities and mapping between problems can enhance problem-solving ability.
Analogies assist in both knowledge retention and concept understanding, influencing overall cognitive processes.