Graded capacity-sharing in dual-task interference?

Pashler, Harold · 1994 · Journal of Experimental Psychology Human Perception & Performance

DOI: 10.1037/0096-1523.20.2.330

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 investigates the underlying cause of dual-task interference, specifically testing whether it results from a discrete central processing bottleneck or from the graded sharing of mental capacity between tasks. The psychological refractory period (PRP) effect, where the second response in a dual-task sequence is delayed, has traditionally been attributed to a bottleneck in response selection. However, previous studies often instructed participants to prioritize the first response, which may have artificially suppressed any potential for resource sharing. Pashler argues that if graded capacity sharing were possible, subjects instructed to place equal emphasis on both tasks should demonstrate a smooth distribution of interresponse intervals (IRIs). In contrast, a bottleneck model predicts bimodality in IRI distributions, reflecting two distinct processing orders (Task 1 then Task 2, or vice versa) or response grouping. To test these competing hypotheses, Pashler conducted an experiment with 24 participants performing two simultaneous choice reaction-time tasks: responding to a tone with the left hand and a letter with the right hand. Stimuli were presented with varying stimulus onset asynchronies (SOAs), including a critical 0-ms SOA condition where stimuli appeared simultaneously. Crucially, instructions explicitly directed subjects to place equal emphasis on both tasks and discouraged prioritizing one response over the other, thereby creating conditions most favorable for graded resource sharing. The study analyzed reaction times (RTs) and the distribution of IRIs to determine the processing strategy employed by participants. The results revealed that subjects did not exhibit the broad, smooth IRI distributions predicted by the capacity-sharing model. Instead, participants clustered into two distinct groups. Approximately 17 subjects ("double-ridge" subjects) showed bimodal IRI distributions, with responses occurring in one of two distinct orders, consistent with the bottleneck model’s prediction of serial processing. The remaining six subjects ("spike" subjects) exhibited IRIs clustered tightly around zero, indicating a strategy of response grouping, where both responses were selected and executed as a coupled unit. No subjects displayed the intermediate, continuous distribution of IRIs that would indicate arbitrary splitting of mental resources. Furthermore, RT correlations and error rates supported the conclusion that preparation and serial processing constraints dominated performance. The findings strongly support the existence of a structural bottleneck in response selection rather than a flexible, graded sharing of capacity. The data suggest that the bottleneck is not merely a strategic choice made by participants to comply with speed instructions, as it persisted even when equal emphasis was mandated. Instead, the inability to smoothly distribute resources indicates a fundamental limitation in the cognitive architecture for selecting responses. This conclusion challenges theories positing that mental capacity can be arbitrarily divided among concurrent tasks, reinforcing the view that dual-task interference arises from discrete, non-overlapping stages of processing.

Key finding

Most participants exhibited bimodal interresponse interval distributions consistent with a structural response-selection bottleneck, rather than the smooth distributions predicted by graded capacity-sharing models.

Methodology

lab_experiment

Sample size: 24

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. Discovered via openalex_abstract on 2026-05-08.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

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

• dual task multitasking
• dual task performance
• capacity resource theory