Article archive / April 1, 2014
Originally published: . Volume 5, No. 8. Source pages: 1, 10.
Airport Security? Error Rates 40 Percent Under Some Conditions
Louisiana State University Assistant Professor of Psychology, Dr. Megan Papesh, has authored new research demonstrating that people can’t match photo IDs nearly as well as expected, bringing into question issues of security programs that use photo ID matching to catch false passports and other high- risk situations.
In her recent article, “Infrequent identity mismatches are frequently undetected,” Papesh and co-author, Dr. Stephen Goldinger, Arizona State University, found that error rates for catching a face to photo ID mismatch could go as high as 40 percent under certain conditions.
The article, published in Attention, Perception and Psychophysics, examined the ability of people to match face photos to IDs under conditions where the observer encounters a fake ID only rarely, or when the observer encounters a high number of fake IDs.
When observers encountered a high number of faked IDs, the error rates were about 20 percent. However, during trials where the frequency of faked IDs was less regular and frequent, the error rates jumped to over 40 percent misses.
So in the easier task, the person missed two out of 10 faked ID, but when the context becomes complicated, he or she can miss as many as four out of 10. How serious? If ten people with faked passports try and come through the airport security gate, four might just keep right on going.
Also, these error rates remained high even when the researchers attempted to have observers slow down and work to catch their mistakes. This suggested to the scientists that face matching is difficult under the best of conditions and errors rates can rocket up when the conditions become more difficult, as in most real world situations.
In the study, volunteers observed current photographs of consenting adults and tried to determine if there was a match or not, with a ID photo which could be up to seven years earlier of the same individual.
The results have important applications and the news media picked the story up quickly. LSUverse.tumblr.com, All Voices, News.nom.co, Science News for Writers, and Science Daily were just some of the media outlets carrying the announcement.
Dr. Papesh is the director of the Eye Movements, Memory, and Attention Laboratory at LSU. Her research on face matching has been covered by the Huffington Post, Wall Street Journal, and on National Public Radio.
Dr. Papesh published an essay on “MIND Guest Blog” for Scientific American describing her work and its potential relevance.
“Imagine that you are a bouncer, checking IDs outside a popular bar in a college town,” she wrote. “It is somewhat dark outside the door, there are many distractions: loud music is playing and your job requires you to also keep an eye on the crowd for trouble. And because the patrons are dressed for a night out, many of them look somewhat different than their ID photos. Despite all these challenges, intuition probably tells you that matching faces to ID photos is easy and accurate.”
That intuition is wrong, she tells readers, and goes on to describe situations at airport security stations.
The Times asked Dr. Papesh what she views as some of the most useful applications of this research. “The most useful piece of information to come from our specific study is that performance is drastically affected by contextual statistics (i.e., the frequency with which observers encounter fake IDs). For example, people can typically assume that a small minority of individuals traveling by air are going to present fake/stolen–but, for all other intents and purposes, valid–photo IDs. Unfortunately, those low-prevalence conditions are the ones that give rise to the worst performance, in terms of catching fake IDs.”
She explained that she and her team will be adding to this discovery. “Absolutely! We are currently following up with a number of experiments aimed at determining the cognitive basis of these effects, and whether they can be alleviated by different training protocols. We’re also ever on the lookout to expand our stimulus set to include a greater range of ages and ethnicities. Our current study likely overestimated performance, as our stimuli depicted college students and our observers were also college students.”
At her Eye Movements, Memory, and Attention Laboratory at LSU, she and her team also research topics related to episodic, memory, language processing, and visual search.
“My work is going to continue to focus on the issue of unfamiliar face matching,” she said, “but I also investigate human memory processes, language, and the role of eye movements in learning and retrieval.” She collaborates with researchers at Arizona State University, University of Pittsburgh, and University of Southampton.
“We are investigating eye movements during ‘mindless reading’ — those times when you catch yourself at the bottom of the page and wonder ‘What did I just read?’”
Sample research items constructed for use in Dr. Megan Papesh’s research on the challenges of photo ID and face matching. People can’t match photo IDs nearly as well as expected, bringing into question issues of security programs that use photo ID matching to catch false passports and other high-risk situations. Here we see the photos of Louis Jones, undergraduate research assistant in Dr. Papesh’s lab at LSU, the type used to simulate face–ID tasks.
This article is part of a historical newspaper archive. Read the original issue for its original layout, photographs and graphics.