simon & chabris

• Neisser et al (1979) composed a divided-visual-attention task

  • observers watched superimposed videotapes
  • two teams playing pass with a ball
  • each person asked to attend to one team - press button when one made a pass
  • woman carrying an umbrella walked across the screen
  • 6 out of 28 saw the woman with the umbrella - when asked not to focus on a team, they always noticed the umbrella
• inattentional amnesia - Wolfe (1999) proposed the unexpected event (umbrella woman) is perceived consciously but then immediately forgotten
• aim

  to examine variables that affect inattentional blindness in naturalistic, dynamic events

  • looking at the role of superimposition compared to live events
  • measuring the impact of task difficulty
  • whether unusualness of the unexpected event had an impact on detection rates
• experimental design

  laboratory experiment

  independent measures design

  ppts only took part in one of the 16 conditions
• sample

  228 ppts

  volunteer sampling

  most were undergraduate students - offered a chocolate bar or a one off payment

  data from 36 ppts discarded so results used from 192 ppts - equally distributed across 16 conditions
• discarded data

  1. knew about the experiment or similar experiments - 14
  2. reported losing count of the passes - 9
  3. passes were incompletely or inaccurately recorded - 7
  4. answers not clearly interpreted - 5
  5. ppts total pass count was more than three standard deviations away from the mean of other ppts in the condition - 1
• independent variable

  unusualness of event - umbrella or gorilla

  style of the video - transparent or opaque

  team colour - asked to follow black or white team

  task difficulty - easy or hard
• dependent variable

  number of ppts in each of the conditions who noticed the unexpected event (umbrella woman or gorilla)
• transparent condition - the white team, the black team and unexpected event were all filmed separately, and the three video streams were rendered partially transparent and then superimposed by using digital video-editing software
• opaque condition - all seven actors were filmed simultaneously and could occlude one another and the basketballs (required some rehearsal before filming to eliminate collisions and other accidents to achieve a natural-looking patterns of movement)
• previous research did not systematically consider the role of task difficulty and no direct comparisons were made between performance with a superimposed version of the display with a live version - for this study several video segments with the same set actions, in the same location, on the same day were filmed
• procedure 1

  4 video tapes, each 75 seconds long

  each tape showed two teams of three players, one wearing white shirts and the other black shirts

  players moved around in a relatively random fashion in an open area in front of three elevator doors

  team passed a standard orange basketball to one another in a standardised order: player 1 -> player 2 -> player 3 -> player 1

  players would dribble the ball, wave their arms and make other movements consistent with their overall pattern of action
• procedure 2

  21 experimenters tested the ppts

  standardisation of procedure was ensured by a written protocol that was devised and reviewed with the experimenters before data collection was begun

  experimenters followed the script to deliver the same instructions to all ppts - told them how and when to present the video and collect the data for each trial

  videos were presented on TV monitors of different sizes, between 13-36 inches

  ppts tested individually and gave informed consent in advance
• procedure 3

  before viewing the video tape, ppts were told they would be watching two teams of three players passing basketballs and they should pay attention to either the team in white (white condition) or team in black (black condition) - told to keep a silent mental count of the total number of passes made (easy condition) or separate silent mental counts of the number of bounce and aerial passes made (hard condition)

  ppts either in transparent or opaque condition

  after viewing and performing the monitoring task, ppts were asked immediately to write down the counts on paper
• procedure 4

  after 44-48 seconds of action

  • the umbrella-woman condition - a tall woman holding an open umbrella walked from off camera on one side to the other, left to right
  • gorilla condition - a shorter woman wearing a gorilla costume that fully covered her body walked the same way

  both unexpected events lasted 5 seconds, players continued their actions during and after the event
• questions asked following the task

  1. while you were doing the counting, did you notice anything unusual in the video
  2. did you notice anything other than six players?
  3. did you see anyone else appear on the video?
  4. did you see a gorilla [woman carrying an umbrella] walk across the screen?
• details of any 'yes' responses were noted.

  observers asked whether they had previously heard of or participated in an experiment similar

  ppts were debriefed and given the opportunity to rewatch the video

  • if ppts said yes, they were replaced and their data was discarded
  • each testing session lasted 5-10 minutes
• to test this further, they investigated whether a longer and more obvious event would be noticed

  separate opaque-style video recording, the gorilla walked from right to left into the live basketball-passing event, stopped in the middle of the players as the action around it continued, turned to face the camera, thumped its chest, and resumed walking across the screen

  12 new observers watched the video while attending the white team and engaging in the easy task - only 50% noticed -> roughly the same that noticed the normal opaque/gorilla walking event (42%) under same task conditions
• results (percentage) who noticed in easy task

  transparent:

  white, umbrella - 58

  white, gorilla - 8

  black, umbrella - 92

  black, gorilla - 67

  opaque:

  white, umbrella - 100

  white, gorilla - 42

  black, umbrella - 58

  black, gorilla - 83
• results (percentage) who noticed in hard task

  transparent:

  white, umbrella - 33

  white, gorilla - 8

  black, umbrella - 42

  black, gorilla - 25

  opaque:

  white, umbrella - 83

  white, gorilla - 50

  black, umbrella - 58

  black, gorilla - 58
• overall level of inattentional blindness

  54% noticed the unexpected event

  46% failed to notice
• effect of type of video

  67% noticed in the opaque condition

  42% noticed in the transparent condition
• difficulty of task

  easy condition - 64%

  hard condition - 45%

  transparent condition - easy 56%, hard 27% per condition

  opaque condition - easy 71%, hard 62% per condition
• nature of the unexpected event

  65% noticed umbrella woman

  44% noticed gorilla

  regardless of video type, monitoring task or attended team - image was most closely associated with the attended task as seeing a gorilla would be much less likely in real life
• conclusions

  individuals have a sustained inattentional blindness for dynamic events - fail to notice an outgoing and highly salient but unexpected event if engaged in a primary monitoring task

  level of inattentional blindness depends on difficulty of primary task

  more likely to notice unexpected events if these events are visually similar to the events they are paying attention to

  no conscious perception without attention
• research method - strength

  number of controls limited extraneous variables - timings of presentation of video were identical for each ppt, moves in the opaque condition were carefully rehearsed so videos for black and white teams were the same
• sample - weakness

  mainly undergraduate students - difficult to generalise as predominantly young people are likely to be more vigilant than average

  basic cognitive processes are universal but not always - Masuda and Nesbitt (2006), cultural differences in change blindness -> asian ppts more likely to detect changes in the context than in the focal aspects of an image
• sample - strength

  participant variables were controlled such as knowledge of the phenomenon investigated - removes some bias from ppts familiar with what they had to do
• quantitative data - strength

  form of yes/no responses to questions following the video - information that is easy to analyse statistically and not open to interpretation

  calculate the percentage who noticed the unexpected event - can compare a number of conditions
• ethics - strengths

  individual's consent was obtained before participation

  unlikely to cause distress despite ppts not knowing the true aim of the task

  ppts were fully debriefed after completing the experiment and were allowed to view the video again
• reliability - strength

  controlled through the use of a standardised script to consistently brief and question ppts
• reliability - weakness

  21 different experimenters, range of television screen sizes used from 13 to 36 inches which could have introduced inconsistencies
• validity - weakness

  low ecological validity - ppts completed the attention task watching a video within a controlled situation -> in real life when we concentrate carefully on tasks requiring our attention there would be a number of other environmental distractions

  inattentional blindness is not a universal visual experience - roughly half of ppts noticed the unexpected event
• validity - strength

  findings were consistent with previous research, including computerised trials, increasing likelihood that findings can be generalised to other situations
• practical applications

  explains why we may fail to notice certain events in real-world situations when attention is focused on a different goal

  can be used to identify situations that may increase likelihood of inattentional blindness - undertaking a particularly challenging task

  leads us to question when 'blindness' occurs - could be we fail to perceive unexpected event altogether or it is quickly forgotten due to its irrelevance to the main task