#Article: Likelihood of Human Error in FRE

[Boyd Baumgartner]

http://www.dailyindia.com/show/439510.php

Study reveals likelihood of human error when dealing with fingerprints

From ANI

Washington, May 12: In a crime scene television show, many times a perfect fingerprint is taken from the crime scene and then a suspect is quickly identified.


However in real life, these crimes can take weeks, months or even years to solve because the crucial portion of an investigation is often done by humans, not by computers.

And now, a new study has revealed that the human factor in the process of tracking fingerprints could lead to errors and incorrect conclusions.

"We knew from other psychological research that the all-too-human foibles of distraction, rushes to judgment, biases and expectations can't be avoided even by the most diligent professionals, so we were understandably concerned about the potential for error," said lead study author Jason Tangen of the University of Queensland.

Tangen, along with the other authors, set out to find determine the likelihood of human error when dealing with fingerprints.

To conduct the study, the researchers drafted 37 qualified fingerprint experts and 37 novices and gave them pairs of fingerprints to examine side by side.

Participants were asked to judge whether the prints matched or not using a "confidence" rating scale from one to 12, one indicating that they were positively sure they were different and 12 denoting that they were certain the fingerprints were a match.

The novices mistakenly identified 55.18 percent of the similar, non-matching pairs as matches. The experts did exceedingly well overall correctly identifying just over 92 percent of the matching pairs. However, they also misidentified almost 8 per cent of the non-matches. It is these kinds of errors that can often lead to failing to identify a criminal.

The study is upcoming in Psychological Science, a journal of the Association for Psychological Science.


Copyright Asian News International/DailyIndia.com

[Boyd Baumgartner]

This is study number 2 that puts error rates at approximately 10%.

The first being the one out of Minnesota BCI found here: http://projects.nfstc.org/ipes/posters/ ... ummary.pdf

[Boyd Baumgartner]

I stand corrected. I found another article where they give the false positive rate of .68 percent. Seems to be a discrepancy between the two articles. I blame math.......


http://www.psychologicalscience.org/ind ... t-csi.html

Dusting for Fingerprints – It Ain’t CSI
Fingerprints – dozens of crime dramas revolve around them. The investigators find the victim, dust for fingerprints, run them through a computer program and -voilá- the guilty party is quickly identified and sent to prison.

If only it were that easy. The reality is that this common but crucial part of an investigation is done by humans, not by computers. An upcoming study in Psychological Science, a journal of the Association for Psychological Science, reveals that the human factor in the process could lead to errors and false convictions of innocent people.

“We knew from other psychological research that the all-too-human foibles of distraction, rushes to judgment, biases and expectations can’t be avoided even by the most diligent professionals, so we were understandably concerned about the potential for error,” said lead author Jason Tangen of The University of Queensland.

“But despite its 100 year history, there have been few peer-reviewed studies directly examining the extent to which experts can correctly match fingerprints to one another.”

Tangen, along with the other authors, set out to determine the likelihood of human error when dealing with fingerprints. They gave 37 qualified fingerprint experts and 37 novices pairs of fingerprints to examine and decide whether a simulated crime scene matched a potential suspect or not. Some of the print pairs belonged to the “criminal” while others were highly similar but actually belonged to an “innocent” person.

The experts correctly matched just over 92 percent of the prints to the criminal. But, they mistakenly matched 0.68 percent of the prints to the innocent person.

“Expertise with fingerprints provides a real benefit,” said Tangen. “But experts—like doctors and pilots—make mistakes that can put lives and livelihoods at risk.” he added.

Tangen said that experts tended to err on the side of caution by making errors that would free the guilty rather than convict the innocent. Even so, they made the kind of error that may lead to false convictions.

“The issue is no longer whether fingerprint examiners make errors, but rather how to acknowledge them.”

###
For more information about this study, please contact: Jason Tangen at jtangen@psy.uq.edu.au.

[Boyd Baumgartner]

Another paper from the same author, RE: fingerprint matching

http://www.nicta.com.au/research/resear ... w9MQ%3D%3D

[Boyd Baumgartner]

http://people.uleth.ca/~vokey/pdf/bbcs2004.pdf

A poster presentation on some fingerprint research by the same author.

The funny thing here is that these guys are psychology researchers and they used a smiley on the match and a frown on the non match buttons. I wonder if that subconsciously influences people?

[Graham F]

Boyd

What a superb conclusion the authors come up with in the 2nd paper you listed - "It appears that novices find it difficult to remember fingerprints and identify them from a line-up." This is some great groundbreaking research that will no doubt be of great bebfit to the fingerprint community.

It has certainly got my juices flowing. I am now thinking of researching why novice plumbers have more leaks from the joints they have soldered when compared to expert plumbers.

[Pat]

What's the error rate of psychology researchers? Higher than the error rate of LPEs, I would be willing to bet.

[Steve Everist]

What's the error rate of psychology researchers? Higher than the error rate of LPEs, I would be willing to bet.

I'm still waiting for the error rate of science.

[Tazman]

I'm still waiting for the error rate of science.

Wait a minute. Is there an error rate in calculating an error rate? OMG -- what if there's an error rate in the error rate of the error rate?

[David Fairhurst]

Of course there is an inherrent error in error rate calculation, as there is in any sampling procedure.

Ever heard of the opinion poll that put the "Yellow Party" 3% ahead of the "Purple Party" plus or minus 5%. Who's more popular?

Well how about the error rate study that comes up with a 3% error rate with a 5% confidence interval? Where's the lab that has 102% accuracy?

Even when we're wrong, we are sometimes right!

[mgirard]

If I'm reading these articles correctly it appears that the 8% error rate is for erroneous exclusions ("Tangen said that experts tended to err on the side of caution by making errors that would free the guilty rather than convict the innocent."), which is approximately what the Black Box Study had it at also. I don't think this is really anything that most examiners wouldn't already suspect, that "missed ids"/erroneous exclusions are where the majority of our errors are. I also think that there needs to be more training focused on exclusions.

[Neville]

I must say that I was amazed to see on the news that science has discovered rogue planets zooming around the universe for the last zillion years, not controlled by a star to orbit around. This it appears has thrown their science into a scenario unthought-of. Wow are they admitting they do not know it all!! Is there something else out there that science knows nothing about, you can bet your life on it.
The word science may mean knowledge but it does not necessarily mean knowing all knowledge. Yes scientists are just as fellable is artists, lets face it the most famous scientists were also artists.