Aim
to examine inattentional blindness for an event in dynamic scene
- to consider role of task difficulty in detection
- to look at effect of superimposed (1 image on top of other) version of display [transparent] compared to live ver [opaque]
- to see whether unusualness of unexpected event has effect on detection rates (gorilla or umbrella)
Research Method
Lab
✓ high control lvls + standardized ~
demand characteristics
Independent measures design
Participants
- 228 ppts almost all students
- volunteered w ✕ compensation
- or given candy bar
- or paid single fee
IVs
1) transparent/umbrella condition
2) transparent/gorilla conditions
3) opaque/umbrella
4) opaque/gorilla
each condition had 4 task conditions
1) white/easy
2) white/hard
3) black/easy
4) black/hard
easy: total no. passes
hard: total no. bounce passes + aerial passes
DV & Materials
- no. ppts in each 16 conditions who noticed unexpected event aka gorilla or umbrella
- 4 vids w teams of 3 players, events occur at same time (44 + 48s), lasts 5s
Procedure
- 21 experimenters tested ppts individually + used standardized script to deliver instructions + followed protocol outlining how/when to present vid & collect data
- ppts told to watch/pay attention to allocated condition vid w 2 teams of 3 players
- then immediately asked to write down their counts on paper and asked:
1) while counting, did u notice anything unusual in vid?
2) did u notice anything other than 6 players?
3) did u see gorilla/umbrella woman walk across screen?
> yes = asked to provide details
> after, asked if heard of or participated in similar event, if yes then data discarded + replaced
> debriefed included replaying vid on request
Results
IV vs IV - % vs %
- transparent vs opaque - 67 vs 42
- hard vs easy - 45 vs 64
umbrella vs gorilla - 65 vs 44
- gorilla noticed by black (58%) more than white (27%)
- but little difference w umbrella B(62%) & W(66%)
Conclusions
- individuals fail to notice unexpected event if engaged in primary monitoring task
- inattentional blindness occurred more frequently in superimposed displays compared to live action
- lvl of inattentional blindness depends on primary task difficulty
- individuals more likely to notice unexpected events if these r visually similar to events they’re paying attention to (black noticed gorilla more)
COGNITIVE AREA
1) Computer model: Input (video clip & counting passes) processed and manipulated to create output (if they saw unexpected event). Noted that brain doesn’t process all info ; inattentional barrier —> don’t notice unexpected event
2) Internal mental processes eg. Attention. Links to cognitive ; individuals more likely to notice unexpected if visually similar to what they’re paying attention to
Types of data
Quantitative data:
✓ Easy to analyse & compare %
✕ Missing qualitative data: no human insight
Validity
- Low ecological; vids artificial, can’t be generalised to real life BUT inattentional blindness occurs irl
- High internal: successfully tested what it was set out to test (examine inattentional blindness in dynamic scene)
Reliability
- replicable & standardized procedure ; controls
- eg. Length of unexpected event (5s), when it occurred, same instructions
- issue: screen size not same for each ppt
Ethics
- little deception, not told ab gorilla
- debrief ; played vid again
Sampling bias/Ethnocentrism
- ✕ Be generalised ; students hv certain intelligence lvls in order to get to uni & r used to getting tested on
- ethnocentric ; in American uni
Practical applications
- eye witness testimonies may not always be reliable ; ppl don’t notice things they don’t actively pay attention to
- insight on factors affecting attention
- can reduce inattentional barriers factors
