Eyewitness identification and change blindness
We’ve written about change blindness (also known as inattentional blindness) before and it’s probably best known as including those experiments with the invisible gorillas. My personal favorite is the one where researchers hid their gorilla in brain scans and had radiologists review the slides. (And social science researchers wonder why professionals like radiologists usually just say NO when asked to participate in their research…)
Today though, we are talking about another (often maligned) area–that of eyewitness identification. It is well-known that eye-witness testimony is often inaccurate even though jurors usually pay close attention to testimony by a witness who “saw it with my own eyes!”. Canadian researchers used 180 undergraduate students (129 women, average age 19.9 years) and had them view two different videos after instructing them to take the role of an eye-witness and pay close attention to what they saw in the videos.
The videos contained footage of a man dressed in what we think of as the Steve Jobs uniform of blue jeans and a black sweater (later referred to as the video innocent) walking up to and entering a building. He then enters two different hallways, shaking the handles of each locked door. Then a second man (also dressed in blue jeans and a black sweater but weighing about 50 pounds LESS than the video innocent) was shown walking through another hallway. He (later referred to as the video culprit) approaches an office door, forces it open, enters the office and finds an iPad. He then exits the office with the iPad in his hand. The video itself was 67 seconds long, the video innocent was on-screen for 33 seconds and the video culprit was on-screen for 31 seconds.
The researchers were interested in whether the research participants would notice there were two different men in the video (one significantly heavier than the other) and whether they would make the correct identification in culprit-present and culprit-absent lineups.
First the participants were asked to recall as much as they could about the video and the researchers used this task to see if the participant had noticed the two different actors (“change detection” versus “change blindness”). Then, they were given black and white photographs of a lineup and asked to circle the photo of the person who stole the iPad, if present. They were told the video culprit may or may not be present in the photograph and they should write “not here” on the photo sheet if the thief was not present.
Only about 1/3 (36.1%) of the participants noticed there were two different men in the video. There was not a significant difference in noticing the two actors based on participant gender (men noted the change 34% of the time, women 36.4%).
Correctly identifying or rejecting lineup photos was lower in the change blindness group (28.7%) than in the change detection group (53.1%). In those situations where the culprit was not present, those who had not noticed there were two actors in the video (i.e., the change blind group) only correctly rejected the lineup 31% of the time (compared with 69% of those in the change detection group). The change blindness group also had higher misidentification rates (i.e., they chose either a filler or the video innocent as the thief). In contrast, none of the change detection group misidentified the video innocent as the thief.
However, being aware of the two actors in the video did not lead to an increase in correct identifications when the culprit was present! The change blind group correctly identified the culprit 26% of the time and the change detection group correctly identified the culprit 34% of the time but this difference was not significant (p = .43).
In short, the majority of participants (64%) did not notice there were two actors in the video (the video innocent and the video culprit). And, while the participants who noticed the two actors in the video did better in avoiding misidentification of the video innocent, they did no better at all in accurately identifying the video culprit when he was present in the lineup. (In research-speak, false positive identification was prominent, but no difference was seen in false negative identification.)
The researchers say that when eye witnesses experience a change blindness error, they are more likely to misidentify a person who is then at risk of wrongful conviction. The researchers clarify the application of their research to a line up task at a police station in that while “fillers should match either a description of the culprit or the appearance of the suspect, archival research suggests that line-ups are often biased towards [a similarity to the appearance of] the suspect”. This is, the researchers say, a “biased lineup” and can lead to a miscarriage of justice. However, if the lineup is properly constructed, the researchers say “change blindness may have few negative consequences for identifying a culprit from a line-up”.
From the perspective of litigation advocacy, you of course want to ensure the lineup was an appropriate one but you also want to consider educating jurors on the well-documented problems with eye-witness identification–particularly in cases of cross-race identification. In our experience, jurors want to do the right thing and if the only strong evidence is an eye-witness identification, that may be enough to introduce reasonable doubt.
Fitzgerald, R., Oriet, C., & Price, H. (2014). Change blindness and eyewitness identification: Effects on accuracy and confidence Legal and Criminological Psychology DOI: 10.1111/lcrp.12044