Guided search: An alternative to the feature integration model for visual search.

Wolfe, Jeremy M.; Cave, Kyle R.; Franzel, Susan L. · 1989 · OpenAlex-citations

DOI: 10.1037//0096-1523.15.3.419

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper challenges the prevailing Feature Integration Model (FIM) of visual search, which posits that searching for targets defined by conjunctions of basic features (e.g., color and form) requires slow, serial attention. The authors argue that FIM cannot account for experimental data showing that unpracticed subjects often perform conjunction searches with reaction time (RT) slopes significantly shallower than predicted, and sometimes independent of set size. To resolve this discrepancy, the authors propose the "Guided Search" model, in which parallel processes analyzing simple features guide the serial spotlight of attention toward likely targets, thereby increasing search efficiency. The study employs a series of visual search experiments where subjects identified targets among distractors while RTs were measured across varying set sizes. Experiment 1 tested conjunctions of color and form (e.g., red O among green Os and red Xs). Results showed shallow RT slopes (average 7.5 ms/item for target trials), far lower than the ~40 ms/item predicted by serial search, with low error rates ruling out speed-accuracy trade-offs. Experiment 2 replicated these findings using color and orientation conjunctions, revealing nonlinear RT functions that became even shallower at larger set sizes. Experiment 3 examined color and size conjunctions, finding that some subjects exhibited RTs virtually independent of set size. To ensure these results were not due to general subject ability, Experiment 4 demonstrated that searching for rotated Ts among Ls produced steep, serial-like slopes consistent with FIM predictions. Further experiments addressed potential confounds. Experiments 5 and 6 ruled out learning effects, showing that shallow slopes persisted even with minimal practice. Experiment 7 identified stimulus properties as a key factor, noting that certain conjunction stimuli naturally facilitate efficient search. Crucially, the authors tested a prediction of the Guided Search model in Experiment 8: searches for triple conjunctions (color, form, and size) should be more efficient than standard conjunctions because three parallel processes can guide attention more effectively than two. Results confirmed that triple conjunction searches produced shallower slopes than standard conjunctions, and in some conditions, were independent of set size. The findings indicate that the standard Feature Integration Model is insufficient because it fails to predict the efficiency of conjunction searches for naive subjects and cannot explain why triple conjunctions are easier to find than double conjunctions. The Guided Search model provides a coherent explanation by integrating parallel feature processing with serial attentional guidance. This framework suggests that visual search is not strictly serial for conjunctions but is instead guided by preattentive feature information, offering a more accurate account of how humans locate complex targets in cluttered visual fields.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-25; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• visual search