Seventy-seven Mount Holyoke College students participated
in mixed groups (2x2) study that examined the effects of an imageís familiarity
on oneís ability to detect visual changes within images. Half of subjects
were asked to detect the changes while engaged in conversation to additionally
examine the effect that divided attention has on change blindness. An analysis
of variance was run on the resulting change detection reaction times. Results
show that participants are able to detect changes significantly more quickly
in familiar images than they are able to detect changes in unfamiliar changes.
Further examination of results led us to examine the difference between
the two set of images used. Significantly different reaction times were
found between the two sets of images. There was no main effect for the
distraction variable, although it did not have a significant effect on
the second picture set. Possible reasons for these results, and ideas for
further study, are discussed.
Subjects will be able to detect changes within familiar images
more quickly than they will be able to detect changes within unfamiliar
Distracted subjects will take longer to detect changes than
non distracted subjects.
Those subjects who are distracted and looking for
changes within unfamiliar images will have the slowest reaction times,
while those subjects who are not distracted and are looking for
changes within familiar images will have the fastest reaction times.
Change Blindness is the cognitive phenomenon in which
individuals are unable to detect small or large visual changes within an
Change Blindness is often examined by having subjects look
at an image on
a computer screen while the computer quickly alternates
between the image
and a modified version of the image (one with the change
it). There is sometimes a brief interruption, called
a flicker, between
the two images. This has been shown to cause a
higher rate of Change
Familiarityís role in Change Blindness:
Krueger (1975) concluded that familiar images are processed
more quickly by the brain than unfamiliar images.
Werner & Thies (2000) conducted an experiment that examined
Blindness using images that dealt with football.
They found that those
individuals with expertise in football were less likey
to be change blind
than individuals who were football novices.
Distractionís role in Change Blindness:
Studies have shown that recognition processes require a relatively
amount of attention (Hicks & Marsh 2000).
Strayer & Johnston (2001) conducted an experiment in
were asked to perform a simulated driving test while
conversing on a cell
phone. Their results showed that individuals subjected
distraction failed to notice, and were slower to react
to, changes within
their environment than individuals who were not distracted.
The purpose of our experiment:
To examine if individuals are more likely to be able to detect
changes within familiar images (The Mona Lisa and The Scream)
than within unfamiliar images (Portrait of Ginevra de Benci and
Murderer in the Lane).
To examine whether distracted individuals (those asked to
answer a series of questions) would take longer to detect changes than
77 Mount Holyoke College students,
both traditional and non-traditional participated in the study.
PC computers containing a program,
Presentation (www.neurobs.com), a program designed to implement experiments
involving different kinds of stimuli.
Images included, the Mona Lisa,
Portrait of Ginevra de' Benci , The Scream, and Murderer in the Lane (www.allposters.com)
A pre-set list of questions to
engage participants in conversation (the distraction group).
A post test questionnaire ensured
that participants located the correct change and that the Mona Lisa and
The Scream were accurate measures of familiarity.
Participants sign a consent form.
Randomly assigned to either the
distraction and non-distraction groups.
Each participant is given written
instructions: "When you press the return key a visual image and its modified
version will alternate; when you detect a change, press the return key
All participants view a practice
trial to ensure their confidence with the program.
Participants view first set of
images and fill out first half of the questionnaire. The same procedure
is allowed for the second set of images. The distraction group performs
same task while answering researcher's pre-set list of questions.
Participants receive debriefing
A mixed groups design ANOVA was used
to determine if the familiarity of a picture (familiar and unfamiliar)
and the task (distraction or non-distraction) and picture set (Mona Lisa/Ginevra,
Scream/Murderer) had significant effect in reaction time to changes within
the images. There was a significant main effect for picture set, significantly
slower reaction times occurred using the Scream/Murderer set (M = 9662.668),
F (1,28) = 20.375, MSE = 10479116.469, p = .000.
There was also a significant main effect
for familiarity, significantly faster reaction times occurred when using
familiar images (M = 7204.037) than unfamiliar images (M = 9353.292), F
(1,28) = 9.584, MSE = 13431743.687, p = .004.
A significant interaction occurred
between task and picture set. When participants were exposed to a distraction,
they took significantly longer to detect a change within the Scream/Murderer
set (M = 12910.627) than within the Mona Lisa/Ginevra (M = 8612.377), F
(1,28) = 6.227, p = .019. There were no other significant interactions.
Figure 1. Mean reaction times for familiar and unfamiliar images.
Figure 2. Mean reaction times of distraction
and non-distraction groups in each picture set.
Our hypothesis for the effect of familiarity was supported.
Subjects noticed a change more quickly in a familiar image than in an unfamiliar
image. This supports previous research suggesting that change blindness
is decreased by oneís familiarity with the stimulus.
Our hypothesis for the effect of distraction was not supported.
There was no difference in the amount of time needed to notice changes
between the distraction and non-distraction groups.
Subjects noticed changes more quickly in the Mona Lisa/Ginevra
picture set than in the Scream/Murder picture set.
In the distraction group, subjects noticed changes more quickly
in the Mona Lisa/Ginevra picture set than in the Scream/Murder
Suggestions for follow-up studies
While each set included two pictures by the same painter,
the sets differed in style. The Mona Lisa/Ginevra set used
detailed images, while the Scream/Murder set used images
with thicker brushstrokes. Subjects may have visually interpreted these
two styles differently, causing different rates of change blindness.
The changes within the sets differed qualitatively. The
Scream used the perspective of the bridge to make the change blend
in with the rest of the picture and Murder on the Lane used the
edges of the trees. The changes in the Mona Lisa/Ginevra
set did not use existing qualities of the picture and may have attracted
more attention than the Scream/Murderer set.
The post-test questionnaire indicated that subjects had problems
with both images in the Mona Lisa/Ginevra set. When asked
to name the change in the Mona Lisa, several reported that her smile
changed. Subjects may have concentrated on her smile because it is the
most famous part of the picture and delayed looking at other parts of the
image. When asked to name the change in the Portrait of Ginevra deíBenci,
many subjects said that she blinked or something about her eyes changed.
When we reviewed the two images, we found that the resolution on the original
and the modified image was different and she did appear to shut her eyes
in the modified version. We coded this data as an acceptable change. Therefore,
subjects were distracted from the change in the familiar image and had
two acceptable changes to choose from in the unfamiliar images. This would
decrease the chances of finding a significant difference between the two.
While answering the questionnaire, several subjects in the
distraction group reported seeing a change, but when asked what changed,
they reported "I donít know" or "I donít remember." We discarded this data.
However, this may have been an indication that they were actually distracted,
and does not necessarily mean they did not locate the correct change.
A follow up study should control for style and continuity
throughout the experiment.
Familiarity could be better assessed by using a between groups
design in which half of the subjects study an image before trying to locate
a change in that image on a computer. The other half would simply look
for changes on the computer without studying.
Subjects could be shown the original, but not modified, version
of each image they viewed while completing the questionnaire. This would
allow them to remember the place they saw the change without allotting
them extra time to look for it.