ST. PAUL -- Minnesota students completed more than 1 million standardized math, reading and science tests last spring, but school districts reported only one case of obvious cheating to the state Department of Education.
If more students or school staff are cheating - and test security experts say some certainly are - state officials have no way to find out.
The department relies on schools to police themselves when it comes to test security. It establishes procedures and asks districts to report any irregularities, but the state does not investigate those reports or use forensic tools to sniff out cheating.
That leaves the Minnesota Comprehensive Assessments, the accountability tests that all public school students take, susceptible to the kind of cheating scandal that in 2009 implicated 178 Atlanta teachers and principals in 44 schools and landed several in prison.
“They were the fox in the hen house. Of course they weren’t going to investigate themselves,” said John F. Olson, who co-wrote a test security guidebook for the Council of Chief State School Officers.
Two years ago, the Education Department convened a committee to rethink its approach to test security. The spring tests had migrated from paper to computers, raising new concerns about how to preserve the integrity of the results.
In July 2015, that group produced 48 recommendations - including monitoring social media for test questions that students may have posted, establishing consequences for those students, using forensic tools to uncover likely cheating and determining a process for investigating it.
But as schools prepare for the next round of testing, the state continues to trust that students and teachers will do the right thing.
Assistant Education Commissioner Kevin McHenry said that the department wants to become “proactive rather than reactive in regards to test cheating,” but that it would take years to get there.
The procedures manual schools use for test security was updated, he said, but there are no near-term plans to use investigative tools to seek out cheating.
Jennifer Dugan, the department’s research and assessment director, said the trouble with data forensics is the tools can’t say with certainty that someone cheated. They can only raise red flags for someone to follow up on - and doing that would require a bigger budget.
“It’s a much more complicated conversation,” she said.
After the Atlanta Journal-Constitution in 2009 scrutinized improbable gains in that city’s test scores, said test security expert Olson, states were forced to examine their own practices. When the nation’s top state education officials hired him, Olson was surprised by how little they were doing to uncover unethical behavior.
“They weren’t doing anywhere near enough, sort of going with the keep-your-fingers-crossed approach,” he said.
In Minnesota, the incentive to cheat is arguably weaker than it was in Atlanta and other places where significant cheating was uncovered.
Atlanta teachers earned bonuses of $500 to $2,000 for high test scores.
Minnesota’s optional Q Comp program also pays teachers for student test performance, but half the state’s students attend schools that don’t participate. Further, districts spend most of their Q Comp dollars on teacher evaluation and professional development, leaving little for test-based performance bonuses. In Minneapolis Public Schools, for example, teachers earn just a token $1 if their students reach the test target for the year.
Minnesota also differs from other states in how it uses student test scores to retain, dismiss and promote individual teachers. In 23 states, student assessments account for half of a teacher’s performance evaluation, but in Minnesota, it’s just 35 percent.
“I’m not saying that it couldn’t ever happen, but I don’t think that we’re in that mindset of, ‘Raise test scores at all costs,’ ” said Gretchen Chilkott, assessment specialist for South Washington County Schools, who served on the state committee.
“It doesn’t sound like Minnesotans or the Upper Midwest to me.”
Still, parents often equate test scores with school quality, and the scores have real consequences in the metro area, where multiple public school districts, charters and private schools compete for students.
States that search for cheaters typically find them. In his guidebook, Olson said every state that has used extensive data forensics has found evidence of cheating.
“It is not an issue of whether one will find cheating in your state assessment program - it is a virtual certainty that some level of cheating … will be found,” he concluded.
Without those tools, and relying on school officials to self-report, Minnesota rarely gets wind of cheating on the annual MCAs.
In response to a Pioneer Press records request, state education officials recently released 61 “security notifications” that school officials filed during testing last spring.
None pointed to coordinated cheating. In the handful of cases where a proctor seemed to be viewing test questions improperly or helping a student take a test, further investigation by the district found otherwise.
The reports suggested just one clear case of cheating by a student: A 10th-grader at Hmong College Prep Academy in St. Paul passed a note to another student asking for an answer to a question on the science MCA.
In that case, the student’s test was not counted. All told, schools invalidated 1,246 math, reading and science tests - about 1 percent - last spring for various reasons.
In one such case, a student took out his cellphone after completing a reading test; in another, fifth-grade math tests were thrown out because students were mistakenly shown and helped with actual test questions instead of samples.
Reports from other schools included:
-- Various students using cellphones, listening to music or viewing YouTube or other websites during testing.
-- Teachers breaking test protocol, such as by giving students tips on how to take the exams or by failing to supervise the class the entire time.
-- Numerous reports of lost test materials.
-- A third-grader who had written test-taking strategies on a small piece of paper.
Absent from the reports were the social-media posts that plagued the department in 2015. That year, several Minnesota students posted photos of test questions, which state officials said forces them to discard the questions.
Education Department spokesman Josh Collins said the state does not punish students for such behavior, but he suggested schools should help them “understand that there is a tremendous amount of money that goes into developing the test questions.”
2015 also saw two teachers get into trouble for apparent cheating: A Cannon Falls teacher resigned after helping fifth-graders with their math tests, and a St. Paul teacher was investigated for helping seventh-graders with their math exams at Linwood Monroe Arts Plus.
In the days of paper-and-pencil tests, erasure analysis became a fairly common method for uncovering possible cheating. If a group of students had an unusually high number of wrong answers changed to right answers, it suggested adults were correcting the answer sheets.
That’s what happened in Atlanta, where educators held erasure parties so they and their schools would meet ambitious achievement targets and win bonuses.
Minnesota’s online tests don’t allow students - or anyone else - to return to previous sections to change answers, but forensic tools can raise other red flags:
-- Did someone quickly click through a test as if he had all the answers?
-- Did students in the same room answer all the questions the same way?
-- Did students get the hard questions right and easy ones wrong?
-- Did students make major improvement one year and take a big step back the next?
Assistant Education Commissioner McHenry said calls from lawmakers, teachers and parents for a reduction in testing has taken the department’s attention away from test security. If the department were to take on an investigative role, “we would need corresponding funding to go with that,” he said.
Sen. Chuck Wiger, DFL-Maplewood, who chairs the Education Committee, said cheating concerns haven’t come up as lawmakers have worked on other issues related to school testing.
“I think that we need to have a hearing on this. We put a great deal of trust on the local level,” he said.
For now, the department is taking a closer look at data that can be gathered easily. Research and assessment director Dugan said they can see, for example, whether test proctors have completed the required training.
Jim Bartholomew, a test security committee member and education policy director for the Minnesota Business Partnership, said that if the department needs more money for test security, he thinks education lobbyists would support them.
“I don’t know anyone who wouldn’t be supportive of bringing improvements to our current system - because it is so critical,” he said.
The MCAs provide important information to schools and the general public, he said.
“If that gets undermined, it’s not a good thing.”