Change Blindness: The Effects of Cohesive Relatedness

and Pictorial Vs. Textual Presentation on Recall

By: Sophorn Chip, Tulani Gonzalez,

Janice Harmon, Jessica Rodrigues

The purpose of this study was to examine detection of change in stimuli (picture or text) and the effect of cohesive relatedness (related or unrelated) on memory recall. A 2x2 mixed groups design was used to measure the percentage of changes (overall and correct) recalled. Forty female college students were randomly assigned to picture or text conditions, which had an equal number of both related and unrelated objects or words. There were two main effects and no interaction between the variables. Participants detected more changes in picture and more correct changes in the related condition. Cohesive relatedness had a significant effect on the detection of change; unrelated pictures or text were more likely to be detected and correctly recalled.


Main Effect 1: Change is more easily detected in text than in pictures.

Main Effect 2: Change is more easily detected in unrelated stimuli than related stimuli.

Interaction: Change is more easily detected in unrelated text than related text and related/unrelated pictures.


Subjects: Materials: Procedure:

To measure the detection of change in pictorial and textual stimuli, each subject was presented with either four pairs of pictorial stimuli (four original pictures and their modified versions), or four pairs of textual stimuli (four original paragraphs and their modified versions). A researcher provided the first stimulus and the subject was allowed to study it for a period of 60 seconds. The researcher then removed the stimulus and the subject was asked to work on a simple crossword puzzle for 60 seconds. The crossword puzzle was used to prevent rehearsal of the original stimulus before the modified stimulus was shown. After the 60 seconds the researcher then provided them with the modified version of the stimulus for another 60 seconds. After the 60 seconds, the researcher removed the stimulus and the subject was instructed to flip to their recall sheet and fill out section A. The sheet informed the subjects to record any changes they may have noticed. Subjects were allotted 1 minute to complete this part of the recall sheet. This process was repeated until the subject had been exposed to all levels of the stimuli (related and unrelated) within their specified condition (pictures or text). After the final section of the recall sheet had been filled out, subjects were thanked for their participation and debriefed.


A 2 (picture or text) X 2 (related or unrelated) between/within mixed groups ANOVA was used to determine whether type of stimuli and degree of cohesive relatedness would have a significant effect on the number of changes detected. The percentage of any detected changes and the percentage of correctly detected changes were the dependent variables.

In our study, our results did not support our hypotheses. Our first hypothesis, that change would be more easily detected in textual presentation than in pictorial presentation, was not supported. Instead, we found that change was more easily detected in pictorial presentation than in textual presentation. Contrary to our previous research in our pilot study, we found that change was more easily detected textually than pictorially. This discrepancy may be due to subjects reading four sets of text, as opposed to one set of text in our pilot study. This increase in the number of texts with 5 color changes created a maturation effect, possibly causing fatigue in participants and reducing memory recall. Our results are supported by Shore’s (2000) research, which found that changes in color were easily detected in pictorially stimuli.

Our second hypothesis that change would be more easily detected in unrelated items, than in related items was not supported. It is important to note that we divided our results into overall changes detected by subjects, and correctly detected changes by the subjects. We found no significant difference between the number of overall changes detected for our cohesive relatedness independent variable. However, we did find a significant difference between the number of correctly detected changes. Our results are supported by Hollingworth’s, et al. (2000) research, which found that change was more easily detected with unrelated items in a scene than with related items in a scene.

Some suggestions for future improvement might be a change in the time allotted per stimulus, different degrees of the complexity of the stimulus, and different types of changes.


Hollingworth, A., Williams, C. C., and Henderson, J. M. (2001). To see and remember: Visually specific information is retained in memory from previously attended objects in natural scenes. Psychonomic Bulletin & Reviews, 8(4), 761-768.

Shore, D. I., and Klein, R. M. (2001). The effects of scene inversion on change blindness. The Journal of General Psychology, 127(1), 27-43.