Nothing compares 2 views: Change blindness can occur despite preserved access to the changed information

Mitroff, Stephen R.; Simons, Daniel J.; Levin, Daniel T. · 2004 · OpenAlex-citations

DOI: 10.3758/bf03194997

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 mechanisms underlying change blindness, specifically testing whether the failure to detect visual changes stems from a lack of internal representation or a failure to compare retained representations. While change blindness is often interpreted as evidence that visual representations are sparse or fleeting, the authors argue that it may result from a comparison failure, where observers retain sufficient information about both pre- and post-change states but fail to align them. To test this, the study examines whether observers retain representations of both the pre-change and post-change objects on trials where they report no awareness of a change. The researchers conducted experiments using a change detection task where participants viewed arrays of line-drawn objects (4, 6, or 8 items) separated by a blank interval. On change trials, one object was replaced. Participants reported whether they detected a change and then answered three two-alternative forced-choice (2AFC) questions identifying which of two options had been present in the display: the pre-change object, the post-change object, or an unchanged object. This design allowed the authors to measure memory for both states independently of change detection reports. The analysis compared performance on "unaware" trials against two chance boundaries: a lower boundary assuming random guessing (25% joint accuracy) and an upper boundary assuming observers could retain only one representation per trial, forcing a guess on the other. The results demonstrated that observers significantly exceeded both chance levels on unaware trials. In Experiment 1, participants correctly identified both the pre- and post-change objects on 42.98% of unaware trials, which was significantly higher than the upper boundary of chance (40.63%, p < .004). This pattern was replicated in subsequent experiments. Furthermore, accuracy on individual 2AFC questions for pre-change, post-change, and unchanged objects was significantly above chance for both aware and unaware trials. These findings indicate that observers retained sufficient visual information for both the initial and final states of the display even when they failed to notice the change. The study concludes that change blindness can occur despite preserved access to the changed information, providing strong empirical support for the comparison failure hypothesis. The findings suggest that internal visual representations are more detailed and stable than previously assumed, but that successful change detection requires a specific process to compare these representations. This implies that the richness of visual experience may rely less on the completeness of stored representations and more on the cognitive mechanisms that align and compare successive views of the world.

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 2026-06-26; verification: verified.

Topics

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

• inattentional change blindness