+1 (208) 254-6996 [email protected]

Read it and complete the assignment Write a 2 page paper (2 pages of TEXT, 1 inch margins, Times New Roman 12 point font, double spaced. You do not need an abstract or a title page) summarizing what the researchers did, what they found to be significant and interesting, any critical evaluations of their methods or findings, and your personal response to the article. The paper should be in your own words. You should not cut and paste any content from the article.


Psychological Science 2019, Vol. 30(3) 436 –443 © The Author(s) 2019 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/0956797618822969 www.psychologicalscience.org/PS

Don't use plagiarized sources. Get Your Custom Essay on
Write 2 Page Paper
Just from $13/Page
Order Essay


A flick of the wrist, and an object vanishes; a wave of the hand, and an object transforms. Magicians are in the business of hiding changes in plain sight, diverting attention to induce change blindness; we almost never spot their method even when it happens in clear view. In his book, Magic by Misdirection, Dariel Fitzkee (1945/2009) describes one such misdirection technique: “A sudden change in the direction of a movement, as from a horizontal path of action to a vertical one, in making a pass, is a distraction” (p. 139). This technique is used in the wave-change card trick, in which one card appears to transform into another at the moment the wave changes direction (see https://www.youtube .com/watch?v=-H5UFMVJslg; the magician in the video mentions that the change is disguised better when the change happens the moment the hand changes direc- tion, rather than at an arbitrary point during the wave). Misdirection in magic can be spatial—a magician can direct audience members’ attention away from the critical action—but more elegant misdirection can occur while audience members look right where the effect occurs. To execute the trick, a magician must use

a technique capable of hiding a change even as it hap- pens in plain sight.

Magicians sometimes mask their actions with other, bigger motions, noting that “a big move covers a small move” (Macknik, Martinez-Conde, & Blakeslee, 2010, p. 67) and “if more than one movement is visible, the eye will tend to follow the larger movement” (Lamont & Wiseman, 1999, p. 41). Attention researchers have capitalized on a similar idea to induce change blindness from one moment to the next, hiding a small local change signal with a large visual disruption. Although many forms of visual disruption have been used to induce change blindness (e.g., Arrington, Levin, & Varakin, 2006; O’Regan, Rensink, & Clark, 1999; Rensink, O’Regan, & Clark, 1997; Turatto, Bettella, Umilta, & Bridgeman, 2003), to our knowledge, none have involved a change in motion direction. The closest example comes from studies in

822969 PSSXXX10.1177/0956797618822969Yao et al.Motion-Induced Change Blindness research-article2019

Corresponding Author: Katherine Wood, University of Illinois at Urbana-Champaign, Department of Psychology, 603 E. Daniel St., Champaign, IL 61820 E-mail: [email protected]

As if by Magic: An Abrupt Change in Motion Direction Induces Change Blindness

Richard Yao, Katherine Wood , and Daniel J. Simons Department of Psychology, University of Illinois at Urbana-Champaign

Abstract Magicians claim that an abrupt change in the direction of movement can attract attention, allowing them to hide their method for a trick in plain sight. In three experiments involving 43 total subjects, we tested this claim by examining whether a sudden directional change can induce change blindness. Subjects were asked to detect an instantaneous orientation change of a single item in an array of Gabor patches; this change occurred as the entire array moved across the display. Subjects consistently spotted the change if it occurred while the array moved along a straight path but missed it when it occurred as the array changed direction. This method of inducing change blindness leaves the object in full view during the change; requires no additional distractions, visual occlusion, or global transients; and worked in every subject tested here. This phenomenon joins a body of magic-inspired work that yields insights into perception and attention.

Keywords attention, failures of awareness, change blindness, motion perception, open data, open materials

Received 3/24/18; Revision accepted 10/22/18http://www.psychologicalscience.org/pshttps://www.youtube.com/watch?v=-H5UFMVJslghttps://www.youtube.com/watch?v=-H5UFMVJslgmailto:[email protected]https://sagepub.com/journals-permissionshttp://crossmark.crossref.org/dialog/?doi=10.1177%2F0956797618822969&domain=pdf&date_stamp=2019-02-07

Motion-Induced Change Blindness 437

which an entire scene moves, with the change intro- duced when the scene changes direction at the end point of its motion (Schofield, Bishop, & Allan, 2006). Such scene motion is perhaps more akin to a large, global visual interruption such as a blank interstimulus interval than a magician’s subtle sleight of hand, in which both the change and the motion masking it are small components of an otherwise static scene.

Whereas studies of change blindness typically employ instantaneous disruptions and changes (but see Simons, Franconeri, & Reimer, 2000), continuous motion can mask continuous, gradual changes. An array of dots whose color (or shape, size, or luminance) continu- ously changes appears to stop changing once that array begins to rotate; faster rotation results in more effective “silencing” of the color changes (Suchow & Alvarez, 2011). This phenomenon also seems to be a case of “a  big motion covering a smaller one” (Macknik & Martinez-Conde, 2012, p. 47) but is distinct from the instantaneous change that Fitzkee sought to mask in his stage magic. Continuous motion can hide a continu- ous change, but can a single change in direction hide a sudden, otherwise visible change to an object?

In a series of experiments, we tested Fitzkee’s claim in a more controlled setting than a magician’s stage, and in so doing, we demonstrate an unusually compelling form of change blindness. A change in motion alone can mask an otherwise obvious change, requiring no other visual transients or global disruptions to the scene.

General Method

The experiments were conducted between 2011 and 2013 and were reported in Richard Yao’s doctoral dis- sertation (Yao, 2013). Here, we report the subset of his experiments demonstrating this new form of change blindness (other experiments are reported separately; for the complete set of experiments, see http://hdl .handle.net/2142/45273). The experiments were not formally preregistered. However, experimental scripts, analysis scripts, and data may be found at this project’s Open Science Framework page (https://osf.io/fyrq8/).


All experiments were conducted using Psychophysics Toolbox (Version 3; Brainard, 1997; Kleiner, Brainard, & Pelli, 2007; Pelli, 1997). Experiments 1 and 2 used MATLAB 2007b, and Experiment 3 used MATLAB 2011b (The MathWorks, Natick, MA). The analyses presented in this article were conducted in the R programming environment (Version 3.4.3; R Core Team, 2017) using the packages dplyr (Version 0.7.2; Wickham, Francois, Henry, & Müller, 2018), purrr (Version 0.2.3; Henry &

Wickham, 2019), and tidyr (Version 0.6.3; Wickham & Henry, 2018). The plots were generated with the pack- ages png (Version 0.1-7; Urbanek, 2013), gridExtra (Version 2.2.1; Augie, 2017), and ggplot2 (Version 2.2.1; Wickham, 2009).


Experiments 1 and 2 were conducted on Apple eMac computers, which had a built-in 17-in. CRT display. The screens subtended approximately 33° (horizontal) × 25° (vertical) of visual angle and were operating at a resolu- tion of 1,024 × 768 with a refresh rate of 89 Hz. Subjects’ eyes were fixed at a distance of approximately 56 to 60 cm from the screen by a custom frame and eyepiece.

Experiment 3 was conducted on a Dell Optiplex GX280 desktop computer. Subjects rested their heads on a desktop-mounted chin and forehead rest posi- tioned 60 cm from the display. Stimuli appeared on a 21-in. ViewSonic CRT display set to 1,024 × 768 resolu- tion and a refresh rate of 85 Hz. The display subtended 43.9° (vertical) × 32.9° (horizontal) of visual angle. A second identical computer controlled an EyeLink 1000 eye tracker (SR Research, Kanata, Ontario, Canada) and used EyeLink II/CL Version 2.32 eye-tracking software running from DOS. The tracker monitored the subject’s left eye (monocularly) at a sampling rate of 250 Hz. Eye gaze was tracked only in Experiment 3.

Experiment 1

Experiment 1 examined whether an abrupt and uniform change in the motion of an entire array of Gabor patches would induce change blindness for the rotation of one element in the array. Six Gabor patches traveled together down the left side of the display and then rightward along the bottom (an L-shaped trajectory). One of the patches, randomly chosen on each trial, underwent an instantaneous 15° rotation the moment the array changed direction. On each trial, subjects reported which patch had rotated. In control trials, a randomly chosen Gabor patch rotated while the array was moving along the horizontal leg of the motion path. These trials tested whether continuous motion alone induces change blindness (as suggested by Schofield et al., 2006). We then compared accuracy rates for each motion condition. (Videos illustrating the control and experimental conditions may be viewed at https://osf .io/vn5py/ and https://osf.io/aucve/, respectively.1)


Subjects. Twenty-one University of Illinois undergradu- ates participated in exchange for psychology coursehttp://hdl.handle.net/2142/45273http://hdl.handle.net/2142/45273https://osf.io/fyrq8/https://osf.io/vn5py/https://osf.io/vn5py/https://osf.io/aucve/

438 Yao et al.

credit. They completed the change-detection task after an unrelated task involving rapid object recognition. Because the extent of change blindness was expected to be large (if it was present) and because of the within- subjects nature of the experiment design, a sample size of 21 was deemed sufficient to detect the effect. We did not conduct a formal, a priori power analysis.

Materials and procedure. Arrays were generated automatically for each trial by evenly distributing six Gabor patches in fixed positions in a ring around a cen- tral fixation cross. The ring had a radius of 75 pixels (2.30°), measured from the center of the cross to the cen- ter of the Gabor patches, which appeared at 30°, 90°, 150°, 210°, 270°, and 330° relative to the horizontal. Each Gabor subtended 56 pixels (1.6° of visual angle), had a spatial frequency of approximately 1.5 cycles per degree, and was oriented randomly. The target change consisted of an instantaneous 15° (clockwise) rotation in the orien- tation of one Gabor.

Every trial began with a 0.3° × 0.3° (10 pixels × 10 pixels) fixation cross positioned at the array’s starting location for 500 ms. An array of Gabor patches then appeared around the cross, and the array and cross immediately began moving downward on the screen for a distance of 13.09° of visual angle (426 pixels) at a constant velocity of 19.14° per second (7 pixels per refresh). When the Gabor array was near the lower left corner of the screen, it immediately changed direction to begin moving 90° to the right at the same speed for 18.99° (618 pixels; see Fig. 1). When it reached the end of its path, the array disappeared and reappeared with postrotation orientations at the center of the display under the words “Click the one that rotated.”

In what we labeled the flexion condition, the target rotated at the point of flexion in the lower left corner of the screen, where the array’s path of motion changed direction. In the control condition, the target rotated halfway along the horizontal path of motion, at the screen midline. The target orientation change occurred equally often (five times) at each possible array location in each condition, for a total of 60 randomly ordered trials. At the start of the experiment, subjects received the following instructions:

In the following task, you will follow, with your eyes, an array of striped patches moving across the screen. On EVERY TRIAL, exactly ONE of the patches will rotate slightly while moving with its neighbors. When prompted, you will have to click on the patch that rotated. We are testing what conditions make that rotation harder or easier to see, so do not be surprised if you did not see any rotation. Just do your best and take a guess if you

are unsure. The task is easiest if you follow the “+” sign that appears at the middle of the array, so follow that with your eyes on each trial.

After completing the experiment, subjects were debriefed and asked questions about their perceptual impressions and any strategies they might have used in performing the task.

Results and discussion

Subjects in the flexion condition accurately located the change only 25.6% of the time (SD = 7.2%), consistent with Fitzkee’s claim that a sudden change in direction can conceal a visual signal. In contrast, when the change occurred during linear motion (control condi- tion), subjects localized the changed Gabor 85.9% of the time (SD = 14.3%; see Fig. 1). Every subject exhib- ited change blindness, with higher accuracy in the con- trol condition than the flexion condition.

As the data indicate, a change in trajectory induces change blindness far more effectively (a 60% difference) than continuous motion alone. Continuous, linear trans- lation did little to hide the target rotation, but introduc- ing a change to an object’s orientation as the array changed direction had a profound effect on the visibil- ity of a change, reducing localization to near-chance levels. Critically, change blindness occurred while the object array was fully visible and in the absence of any additional transients or disruption. Moreover, an identi- cal target change was reliably detected in the control condition, showing that the change itself is otherwise readily perceptible.

Experiment 2

An abrupt change in direction, or an anticipated one, might induce change blindness by triggering a saccade. Subjects regularly miss changes that occur during a saccade (Blackmore, Brelstaff, Nelson, & Trościanko, 1995; Grimes, 1996; Hollingworth & Henderson, 2000; McConkie & Zola, 1979), and if the abrupt change in direction induced a saccade, then the change blindness in Experiment 1 might reflect this familiar saccadic sup- pression rather than a novel mechanism. Given that the array was continuously visible and moved along the same predictable path on every trial, subjects should have been able to smoothly pursue the array across the abrupt change in direction. To test the possibility that the instantaneous change induced an eye movement, we conducted Experiment 2, in which we replaced the abrupt 90° flexion with a gradual 90° arc. People smoothly pursue a magician’s hands more often during curved motion than straight motion (Otero-Millan,


Ex pe

rim en

t 1

M ot

io n

Pa th

a nd

C ha

ng e

Lo ca

tio n

Ex pe

rim en

t 2

Accuracy (%)

Ex pe

rim en

t 3

Change Blindness (%)

025507510 0

Co nt

ro l

Fl ex

io n

025507510 0

Co nt

ro l

Ar c

025507510 0

Co nt

ro l

Fl ex

io n

025507510 0 025507510 0 025507510 0

Accuracy (%) Accuracy (%)

Change Blindness (%) Change Blindness (%)

F ig

. 1 .

M et

h o d a

n d r

es u lt s

fo r

E xp

er im

en ts

1 t o 3

. T h e

le ft c

o lu

m n s

h o w

s th

e p o ss

ib le

p at

h s

o f

m o ti o n f

o r

th e

ar ra

ys , w

it h t h e

p o in

t o f

ro ta

ti o n i n d ic

at ed

b y

a cu

rv ed


ro w

o r

a li n e

se gm

en t. B

lu e

in d ic

at es

t h e

p at

h a

n d p

o in

t o f

ro ta

ti o n i n t

h e

fl ex

io n c

o n d it io

n w

h en

t h e

ar ra

y ch

an ge

d d

ir ec

ti o n ; o ra

n ge

i n d ic

at es

t h e

p at

h a

n d r

o ta

– ti o n f

o r

th e

co n tr

o l

co n d it io

n ,

in w

h ic

h t

h e

ch an

ge o

cc u rr

ed a

lo n g

a st

ra ig

h t

p at

h .

In E

xp er

im en

t 3,

t h e

ar ra

y co

u ld

t ra

ve l

le ft t

o r

ig h t

o r

ri gh

t to

l ef

t (l

ef t

to r

ig h t


sh o w

n ).

T h e

m id

d le

c o lu

m n s

h o w

s ac

cu ra

cy i

n l

o ca

li zi

n g

th e

ch an

ge d G

ab o r

fo r

ea ch

c o n d it io

n ,

w it h e

ac h d

o t

re p re

se n ti n g

a su

b je

ct m

ea n f

o r

th at

c o n d it io

n .

N o te


at t

h e

d es

ig n w

as w

it h in

s u b je

ct s,

s o e

ac h s

u b je

ct h

as o

n e

d o t

in e

ac h c

o n d it io

n .

T h e

b la

ck l

in e

in d ic

at es

t h e

o ve

ra ll m

ea n .

T h e

sh ad

ed g

ra y

ar ea

i s

a sm

o o th


k er

n el

d en

si ty

p lo

t o f

th e

p o in

ts . T h e

ri gh

t co

lu m

n s

h o w

s th

e m

ag n it u d e

o f

ea ch

s u b je

ct ’s

c h an

ge -b

li n d n es

s ef

fe ct

, ca

lc u la

te d b

y su

b tr

ac ti n g

th e

su b je

ct ’s

a cc

u ra

cy i n

th e

fl ex

io n c

o n d it io

n f ro

m t h at

i n t h e

co n tr

o l co

n d it io

n . A

ga in

, ea

ch d

o t re

p re

se n ts

1 s

u b je

ct . N

o te

t h at

t h e

va lu

es f o r

E xp

er im

en t 3

ar e

ca lc

u la

te d a

ft er

e xc

lu d in

g tr

ia ls

o n w

h ic

h a

s u b je

ct m

ad e

a sa

cc ad

e at

t h e

m o m

en t

o f

ch an

ge .

440 Yao et al.

Macknik, Robbins, McCamy, & Martinez-Conde, 2011), so the curved path should facilitate smooth pursuit of the array and reduce the likelihood of a saccade. In Experiment 2, the target Gabor patch could rotate either clockwise or counterclockwise to ensure that the direction of rotation did not drive the effect. (Dem- onstration videos for the control and experimental conditions may be found at https://osf.io/a25xs/ and https://osf.io/h9z7s/, respectively.)


Subjects. Eleven University of Illinois undergraduates participated in Experiment 2 in exchange for psychology course credit. Unlike in Experiment 1, subjects completed this study prior to any other tasks in the lab. Because the motion-induced change-blindness effect was so large and appeared in every subject tested, and because we contin- ued to use a within-subjects design, a sample size of 11 was deemed sufficient to detect the motion-induced blindness effect. We did not conduct a formal, a priori power analysis.

Materials and procedure. Unless otherwise noted, Experiment 2 was identical to Experiment 1. Stimulus velocity was lowered slightly to aid smooth pursuit (6 pixels per screen refresh, or 16.4° per second) and maintained throughout each trial. Gabor patches were positioned at 0°, 60°, 120°, 180°, 240°, and 300° relative to the horizontal because of arbitrary changes in the way the experiment was programmed. The array’s path was identical to that in Experiment 1 except that when the array neared the lower left corner of the display, it tra- versed a 90° arc with a 100-pixel (3.07°) radius until it was moving horizontally to the right.

Given that an arching path has no clear flexion point at which the array changes direction, we made the change to the target Gabor’s orientation gradual as well; the target rotated 10° during the 26 screen refreshes (~292 ms) required to traverse the 90° arc. The target rotation in the control condition was identical to that in the arc condition (10° rotation occurring gradually over 26 screen refreshes, with rotation beginning when the center of the array crossed the screen’s vertical midline while moving horizontally). Thus, as in Experiment 1, the target rotation co-occurred with the change in direction. Unlike in Experiment 1, the curved path allowed for easier smooth pursuit of the fixation point, and the gradual change allowed for possible detection even if subjects made a saccade during part of the rotation.

Subjects completed 72 randomly intermixed trials, with 24 in each of three conditions: clockwise rotation occurring halfway along the horizontal path of motion (control condition), clockwise rotation along the 90°

arc, and counterclockwise rotation along the 90° arc. Each of the six target locations was represented four times in each condition (see Fig. 1). The experimenter instructed subjects to localize the target change that occurred on every trial—guessing when necessary— and to follow the fixation cross with their eyes.

Results and discussion

Subjects exhibited substantial change blindness with the arced path of motion, averaging 25.4% correct local- ization (SD = 12.5%), but were much better at detecting the same change when it occurred during continuously horizontal motion (68.6%, SD = 19.8%; see Fig. 1). With arced motion, performance was comparable for coun- terclockwise (23.1%, SD = 12.6%) and clockwise (28.4%, SD = 12.8%) rotation. Every subject exhibited change blindness, performing more accurately in the control condition than the arc condition. Given that the arced motion should permit smooth pursuit, this finding sug- gests that changes in direction induce change blindness even when subjects do not make a saccade at the moment of the change.

Experiment 3

Experiment 2 was designed to reduce the likelihood of saccades by introducing an arced motion path and a gradual change. Experiment 3 used eye tracking with a variant of the change-detection task from Experiment 1 to rule out saccade-contingent change blindness as an explanation for the observed effect. (Demonstration videos for the control condition and two experimental conditions can be found at https://osf.io/aybt3/, https:// osf.io/brkcp/, and https://osf.io/9qn4b/.)


Subjects. Eleven2 subjects participated in Experiment 3. Eight subjects received course credit as part of the Uni- versity of Illinois subject pool, and 4 subjects were volun- teers (two undergraduates and two graduate students from affiliated labs). Because the motion-induced change- blindness effect was so consistently large and showed up in every subject tested, a sample size of 11 was deemed sufficient to observe the effect (again, we did not con- duct a formal, a priori power analysis). The graduate stu- dent volunteers were aware that the experiment was testing whether saccades occurred at the point of flexion but did not know the theoretical reasons for the experi- ment. The task in this experiment had a 5-s time limit in which to make a response; two trials from 1 subject were discarded because of timing out.https://osf.io/a25xs/https://osf.io/a25xs/https://osf.io/h9z7s/https://osf.io/aybt3/https://osf.io/brkcp/https://osf.io/brkcp/https://osf.io/9qn4b/

Motion-Induced Change Blindness 441

Materials and procedure. Subjects completed a task similar to those of Experiments 1 and 2 but with slightly different stimuli. The array consisted of eight Gabor patches, each subtending approximately 2.66° (62 pix- els), distributed evenly into fixed positions around an invisible circle. A 0.17°-wide (4 pixels) fixation dot was displayed at the center of the array, and the distance from the dot to the center of each Gabor was 3.30° (77 pixels). Each Gabor patch had a spatial frequency of approxi- mately 0.94 cycles per degree with identical phase and a random angle of orientation (see Fig. 1).

Each trial began with a vertically centered fixation dot appearing for 500 ms, 14.62° (341 pixels) to the left or right of the center of the display. Subjects were instructed to keep their eyes on the fixation dot throughout the trial, even as it moved. As soon as the stimulus array appeared, it moved for approximately 500 ms at a constant velocity of 14.57° per second (4 pixels per refresh) toward the center of the screen for 7.37° (172 pixels). As soon as the center of the array reached the midpoint of the screen, the target Gabor instantaneously rotated 30°. Simultaneously, the array either continued moving horizontally in the same direc- tion for another 7.37° (172 pixels) or changed direction to move straight up or straight down for 6.34° (148 pixels). The array disappeared for 100 ms before reap- pearing centered on the display, where subjects clicked on the Gabor they thought had changed.

The two direction conditions (left to right and right to left) were completely crossed with the three angle- of-flexion conditions (0°, 90°, and 270°), and the eight stimulus-array locations could serve as the target location five times during the experiment, yielding a total of 240 randomly intermixed trials. Twelve practice trials consist- ing of random combinations of the conditions and pos- sible target locations preceded the 240 test trials.

While the task was being completed, gaze was tracked using a desktop-mounted EyeLink 1000 eye tracker (SR Research, Kanata, Ontario, Canada). Sac- cades were defined by the default normal sensitivity values of the EyeLink parser: Velocity threshold was 30° per second, acceleration threshold was 8,000° per second2, and motion threshold was 0.1°. The experi- ment began with calibration of the eye tracker, and a

drift correction preceded every trial. Subjects were given breaks every 50 trials, and the eye tracker was recalibrated after 150 trials or more often as needed (i.e., if drift correction failed on multiple attempts).

Results and discussion

As occurred in Experiments 1 and 2, subjects exhibited substantial change blindness in the direction-change condition (50.7% correct identification, SD = 8.5%) com- pared with the control condition (98.8% identification, SD = 1.4%). Accuracy in both conditions was higher than in the prior experiments, likely because the change was larger (30° rotation in Experiment 3 vs. 15° in Experiment 1 and 10° in Experiment 2).

If change blindness in Experiments 1 and 2 resulted from saccades occurring during the point of flexion, then change blindness should have been reduced or eliminated after we excluded all trials in which subjects made saccades at the moment of change (i.e., the sac- cade began before the flexion point and ended after the change had occurred). Across all 11 subjects, the change occurred during a saccade for only 0.72% of trials. After excluding those trials, we found that sub- jects correctly located the change 99.3% of the time (SD = 0.9%) in the control condition but only 51% of the time (SD = 8.55%) when Gabor orientation changed at the moment of flexion (see Fig. 1). Neither the direc- tion of travel nor the turn direction had much impact on change localization (see Table 1). Regardless of whether saccade trials were included or excluded, every subject exhibited change blindness.3

Subjects rarely made a saccade at the moment of change, but they still exhibited substantial change blindness. This finding eliminates the possibility of saccade-contingent change blindness. Instead, the change in direction appears to play a causal role in inducing change blindness.

General Discussion

A technique used in stage magic to divert attention away from the magician’s method is equally effective at hiding a rotation to a Gabor patch, even though the change happens in plain sight. Across all experiments, an orien- tation change to a single element was noticed and local- ized far worse if it occurred when the larger array changed direction than when it occurred during continu- ous, linear motion. This change blindness could not be explained by saccadic suppression (Experiment 3), and it occurred regardless of whether the direction change was gradual (Experiment 2) or abrupt (Experiments 1 and 3). The magnitude of change blindness—the differ- ence in correct identification between the linear control condition and the flexion condition—ranged from 43%

Table 1. Mean Percentage of Change-Detection Accuracy for Each Motion Path in Each Direction of Travel in Experiment 3

Direction Straight Upward Downward

Right to left 98.8 (1.8) 51.7 (12.7) 54.5 (11.9) Left to right 99.8 (0.8) 46.7 (7.8) 51.2 (14.9)

Note: Standard deviations are given in parentheses. Trials on which a saccade co-occurred with the Gabor rotation are excluded.

442 Yao et al.

to 60% across experiments, and every subject exhibited the effect. Collectively, the results confirm Fitzkee’s (1945/2009) claim that changes in motion direction them- selves constitute a distraction that can induce a failure of awareness and join a body of work on attention and perception inspired by magic (e.g., Barnhart, Ehlert, Goldinger, & Mackey, 2018; Macknik et al., 2010; Rensink & Kuhn, 2015).

Few documented methods of inducing change blind- ness allow for objects to remain completely visible dur- ing the change and introduce no large visual disruptions. The rotation of the Gabor was effectively masked only by the change in motion of the array that it was embed- ded in, in contrast to most other methods that involve either a global change to the visual display (flicker, mud splash, saccadic suppression, cuts in a movie) or com- plete occlusion of the object (Simons & Levin, 1998). We found that change blindness can occur even when the distraction is relatively small, local, and entirely predictable, rather than large and global. In other stud- ies showing change blindness in the absence of large disruptions, changes were introduced gradually over time to avoid creating a localizable transient signal (Simons et al., 2000), or gradual changes were embed- ded in a continuously moving context (Suchow & Alvarez, 2011). However, we found change blindness even when the change occurred instantaneously (Experiments 1 and 3).

The sudden change in motion apparently hides the orientation change, even though subjects know that a change is coming and when in the trajectory it will occur. This finding of change blindness, like many oth- ers, is consistent with the coarseness of our visual rep- resentations from moment to moment. However, our task was unique in how little disturbance it required to mask an instantaneous and obvious change.

Constraints on Generality

Change blindness was observed consistently across all three experiments, with large differences in localization accuracy between the experimental and control condi- tions despite changes to the task parameters. Given that these effects occurred for every subject in our experi- ment, that they relied on basic cognitive and perceptual limits, and that other forms of change blindness occurred reliably across people and cultures, we have no reason to expect the effects to be limited to the arbitrary characteristics of our sample (although they might vary in magnitude with different groups of sub- jects). We also expect this form of change blindness to be robust to the type of computer display, viewing distance, and software and hardware, as well as to other arbitrary characteristics of the testing procedures.

Given that all of these experiments used orientation changes to Gabors, we do not know whether direction changes would induce change blindness for other types of objects or other types of changes. Although we observed the effect across small variations in the num- ber of objects in the array, the magnitude of the target rotation, and the speed of motion, we do not know whether change blindness would be moderated by more extreme variation in these parameters. Other means of inducing change blindness (e.g., the flicker task) have concealed a wide range of change types with varying degrees of effectiveness, and continuous motion masks continuous changes on many feature dimensions (Suchow & Alvarez, 2011). Sudden changes in direction of motion, however, might mask only instantaneous orientation changes or other motion-based events (see Yao, 2013). Additional experiments are needed to explore the generality of change in motion direction as a means of inducing change blindness.

Action Editor

Alice J. O’Toole served as action editor for this article.

Author Contributions

R. Yao and D. J. Simons jointly planned and designed the experiments. R. Yao programmed the experiments, oversaw data collection, and wrote the thesis from which this article is adapted. K. Wood programmed the analysis, generated the figure and videos, and drafted the manuscript. All the authors edited the manuscript and approved the final manuscript for submission.


Katherine Wood https://orcid.org/0000-0002-3877-9625

Declaration of Conflicting Interests

The author(s) declared that there were no conflicts of interest with respect to the authorship or the publication of this article.

Open Practices

All data and materials have been made publicly available via the Open Science Framework. Data can be accessed at https:// osf.io/dteh2/ (Experiment 1), https://osf.io/dp2u7/ (Experiment 2), and https://osf.io/spqbw/ (Experiment 3). Materials can be accessed at https://osf.io/y6rcj/ (Experiment 1), https://osf.io/ cpzkj/ (Experiment 2), and https://osf.io/gbm78/ (Experiment 3). The design and analysis plans for the experiments were not preregistered. The complete Open Practices Disclosure for this article can be found at http://journals.sagepub.com/doi/ suppl/10.1177/0956797618822969. This article has received the badges for Open Data and Open Materials. More informationhttps://orcid.org/0000-0002-3877-9625https://osf.io/dteh2/https://osf.io/dteh2/https://osf.io/dp2u7/https://osf.io/spqbw/https://osf.io/y6rcj/https://osf.io/cpzkj/https://osf.io/cpzkj/https://osf.io/gbm78/http://journals.sagepub.com/doi/suppl/10.1177/0956797618822969http://journals.sagepub.com/doi/suppl/10.1177/0956797618822969

Motion-Induced Change Blindness 443

about the Open Practices badges can be found at http://www .psychologicalscience.org/publications/badges.


1. Videos for all experiments are recreations of the experimental conditions but are not captured from the original code and may differ slightly from the original appearance. The code used to make the videos is available on the Open Science Framework at https://osf.io/ug5x6/. 2. It was originally reported that 12 subjects participated in this experiment (Yao, 2013). However, there were systematic differ- ences between 1 subjects’ data and those of the other subjects in terms of variables collected and stimulus timing, suggesting that it was part of a different experiment; because these dis- crepancies could not be resolved or explained, the subject was excluded from the analysis. 3. More detailed analysis of the eye-tracking data can be found in Richard Yao’s thesis (Yao, 2013, pp. 27–29).


Arrington, J. G., Levin, D. T., & Varakin, D. A. (2006). Color onsets and offsets, and luminance changes can cause change blindness. Perception, 35, 1665–1678.

Augie, B. (2017). gridExtra: Miscellaneous functions for “grid” graphics (R package Version 2.2.1). Retrieved from https://CRAN.R-project.org/package=gridExtra

Barnhart, A. S., Ehlert, M. J., Goldinger, S. D., & Mackey, A. D. (2018). Cross-modal attentional entrainment: Insights from magicians. Attention, Perception, & Psychophysics, 80, 1240–1249.

Blackmore, S. J., Brelstaff, G., Nelson, K., & Trościanko, T. (1995). Is the richness of our visual world an illusion? Transsaccadic memory for complex scenes. Perception, 24, 1075–1081.

Brainard, D. H. (1997). The Psychophysics Toolbox. Spatial Vision, 10, 433–436.

Fitzkee, D. (2009). Magic by misdirection. Provo, UT: Magic Box Productions. (Original work published 1945)

Grimes, J. (1996). On the failure to detect changes in scenes across saccades. In K. Akins (Ed.), Vancouver Studies in Cognitive Science: Vol. 5. Perception. New York, NY: Oxford University Press.

Henry, L., & Wickham, H. (2019). purrr: Functional program- ming tools (R package Version 0.2.3). Retrieved from https://CRAN.R-project.org/package=purrr

Hollingworth, A., & Henderson, J. M. (2000). Semantic infor- mativeness mediates the detection of changes in natural scenes. Visual Cognition, 7, 213–235.

Kleiner, M., Brainard, D. H., & Pelli, D. G. (2007). What’s new in Psychtoolbox-3? Perception, 36(ECVP Abstract Suppl.).

Lamont, P., & Wiseman, R. (1999). Magic in theory: An intro- duction to the theoretical and psychological elements of conjuring. Hatfield, England: University of Hertfordshire Press.

Macknik, S. L., & Martinez-Conde, S. (2012). Mind-warping visions: 10 brain-twisters compete to be the best illusion of 2011. Scientific American Mind, 22(6), 46–51.

Macknik, S. L., Martinez-Conde, S., & Blakeslee, S. (2010). Sleights of mind: What the neuroscience of magic reveals about our everyday deceptions. New York, NY: Henry Holt.

McConkie, G. W., & Zola, D. (1979). Is visual information inte- grated across successive fixations in reading? Perception & Psychophysics, 25, 221–224.

O’Regan, J. K., Rensink, R. A., & Clark, J. J. (1999). Change- blindness as a result of ‘mudsplashes.’ Nature, 398, 34. doi:10.1038/17953

Otero-Millan, J., Macknik, S. L., Robbins, A., McCamy, M., & Martinez-Conde, S. (2011). Stronger misdirection in curved than in straight motion. Frontiers in Human Neuroscience, 5, Article 133. doi:10.3389/fnhum.2011.00133

Pelli, D. G. (1997). The VideoToolbox software for visual psy- chophysics: Transforming numbers into movies. Spatial Vision, 10, 437–442.

R Core Team. (2017). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.

Rensink, R. A., & Kuhn, G. (2015). A framework for using magic to study the mind. Frontiers in Psychology, 5, Article 1508. doi:10.3389/fpsyg.2014.01508

Rensink, R. A., O’Regan, J. K., & Clark, J. J. (1997). To see or not to see: The need for attention to perceive changes in scenes. Psychological Science, 8, 368–373.

Schofield, A. J., Bishop, N. J., & Allan, J. (2006). Oscillatory motion induces change blindness. Acta Psychologica, 121, 249–274.

Simons, D. J., Franconeri, S. L., & Reimer, R. L. (2000). Change blindness in the absence of a visual disruption. Perception, 29, 1143–1154. doi:10.1068/p3104

Simons, D. J., & Levin, D. T. (1998). Failure to detect changes to people during a real-world interaction. Psychonomic Bulletin & Review, 5, 644–649. doi:10.3758/BF03208840

Suchow, J. W., & Alvarez, G. A. (2011). Motion silences awareness of visual change. Current Biology, 21, 140–143. doi:10.1016/j.cub.2010.12.019

Turatto, M., Bettella, S., Umilta, C., & Bridgeman, B. (2003). Perceptual conditions necessary to induce change blind- ness. Visual Cognition, 10, 233–255.

Urbanek, S. (2013). png: Read and write PNG images (R package Version 0.1-7). Retrieved from https://CRAN.R- project.org/package=png

Wickham, H. (2009). ggplot2: Elegant graphics for data analy- sis. New York, NY: Springer-Verlag.

Wickham, H., & Henry, L. (2018). tidyr: Easily tidy data with ‘spread()’ and ‘gather()’ functions (R package Version 0.6.3). Retrieved from https://CRAN.R-project.org/package= tidyr

Wickham, H., Francois, R., Henry, L., & Müller, K. (2018). dplyr: A grammar of data manipulation (R package Version 0.7.2). Retrieved from https://CRAN.R-project .org/package=dplyr

Yao, R. (2013). The flick of the wrist and the wave of the wand: Low-level mechanisms for inducing change blindness (Doctoral dissertation). Retrieved from http://hdl.handle .net/2142/45273http://www.psychologicalscience.org/publications/badgeshttp://www.psychologicalscience.org/publications/badgeshttps://osf.io/ug5x6/https://CRAN.R-project.org/package=gridExtrahttps://CRAN.R-project.org/package=purrrhttps://CRAN.R-project.org/package=pnghttps://CRAN.R-project.org/package=pnghttps://CRAN.R-project.org/package=tidyrhttps://CRAN.R-project.org/package=tidyrhttps://CRAN.R-project.org/package=dplyrhttps://CRAN.R-project.org/package=dplyrhttp://hdl.handle.net/2142/45273http://hdl.handle.net/2142/45273

Order your essay today and save 10% with the discount code ESSAYHELP