Wrongfully Accused: Where did the system go wrong?

• Tuesday: The case - or lack thereof - against Joe Halvorson.
• Wednesday: Halvorson fears his teaching career is done.
• Today: Where did the system go wrong?

RED WING, Minn. - When prosecution and investigators moved forward with charges against a Rochester man accused of exposing himself to a busload of teenage girls near Cannon Falls, eyewitness identification appeared to be their linchpin.
The suspect, Joe Halvorson, was singled out in 2008 by two girls who independently identified him in a six-man photo lineup.

Halvorson would ultimately be cleared of the charges in 2009, but not before he lost his job as a teacher and spent thousands in attorney fees for his defense.

While First District Judge Kevin Mark later criticized other elements of the investigative process in his order for dismissal, he did not admonish the Minnesota State Patrol trooper's handling of the photo lineup.

According to experts in the field, the Halvorson case illustrates elements of what should - and shouldn't - be done when conducting photo lineups.

University of Wisconsin-Whitewater psychology Professor Elizabeth Olson said police must treat eyewitness testimony as carefully as they would any other evidence. Like other pieces of evidence, it can become contaminated and misconstrued, she said.

"The courts have traditionally thought eyewitnesses are like video-recorders," she said. "People don't work that way."

Eyewitness identification cases can be riddled with flaws from the outset, according to the Innocence Project, a national organization dedicated to exonerating the wrongly accused.

The organization, along with experts in the field of misidentification, supports advances in the field - some of which are already being practiced in Minnesota and around the country.

One reform under way at the Bloomington, Minn., Police Department is what's known as the "double-blind sequential method" for photo lineups. The method includes small - but what experts say are critical - changes to the traditional photo lineup process.

Instead of throwing a smorgasbord of photos to witnesses, authorities present pictures one by one - a process that proponents say forces witnesses to utilize their memory over comparative analysis.

That portion of the method was used by Minnesota State Patrol trooper Troy Siems when he conducted photo lineups with the two eyewitnesses in the Halvorson case. In his report, he describes how he presented the photos sequentially to the girls.

Olson, who has conducted extensive research into photo lineups, said the sequential method means it's less likely for witnesses to think, "Well, I'll just pick the guy who looks closest" to their recollection.

The lineup form Siems had the witnesses sign also included a phrase favored by experts of the alternative method that reads "the suspects(s) involved in the case may or may not be included in the photographs."

But Siems left out the other portion of the method that experts and law enforcement officials say can tip the hand of the investigator.

That element, the "double blind" portion, calls for the lineup administrator to have no knowledge of the faces presented in the lineup.

Supporters of the method say it eliminates the possibility of the administrator giving witnesses intentional or unintentional clues.

"There's no way you can give anybody hints," Bloomington Police Cmdr. Mark Stehlik said of the "double blind."

Minnesota State Patrol Lt. Jeffrey Westrum said the State Patrol does not have a standing policy regarding photo lineups.

Though a proponent of the method, Stehlik pointed out the double-blind method isn't always practical for some law enforcement agencies.

Sometimes it becomes "unwieldy and time-consuming" to locate a second detective to conduct the interview.

Other circumstances

Halvorson admits it's completely possible the girls picked him from the lineup for a simpler reason: they probably did see him.

After the actual perpetrator flashed the girls, he decelerated and dropped back behind the bus. The girls were instructed to get vehicle and suspect descriptions the next time they saw it.

Halvorson, on his way home from work at Zumbrota-Mazeppa High School, passed the bus in his car near Rochester, unaware that apparent similarities led the girls to believe his car was the suspect vehicle.

But even given those circumstances, Halvorson believes that if Siems had conducted the complete double-blind sequential method during the photo lineup, he might have been spared the grief.

"It would have given me a better chance," he said.

Siems did not respond to multiple requests seeking inquiry into the case.

Objections raised

Bloomington police adopted the double-blind sequential process after participating in a pilot program headed by the Hennepin County Attorney's Office. Stehlike said it has been an effective addition to the investigative process.

"It's certainly not any less effective by any means," Stehlik said.

The method is not universally supported, however.

"At the present time, there is no definitive sense that one form of lineup presentation is superior to the other," University of Texas-El Paso professor Roy Malpass said in a 2007 National Institute of Justice report.

Nowhere were those sentiments more scornfully expressed than in a 2003 program conducted in Illinois that tasked its state police with implementing the double-blind sequential method. The results from the program concluded the method produced a higher rate of false identifications.

A panel of social scientists later called that study "flawed" after learning of inconsistencies in its methodology.

"The only way to sort this out is by conducting further studies," the panel opined.

The Innocence Project also criticized the Illinois study, calling it a "seriously flawed, non-scientific field study."

Longaecker writes for the Red Wing Republican Eagle