Custom-written exams by robots to fight student cheating
Tailored exams containing unique data sets for each student tested at a UK university could significantly reduce cheating in online assessments, the researchers believe.
With some studies suggesting that cheating has massively increased after switching to remote assessments, researchers have started looking at how they can design exams that prevent students from getting along or plagiarizing the work of their classmates.
In a new approach, chemists at the University of Exeter use computer coding to generate 60 different sets of data for a single class – one for each student – passing a data analysis test worth 20 percent of the total score of the module.
THE Campus resource: using multimedia as a tool to enrich and transform the assessment
The script – which models lab equipment to produce realistic data but introduces some randomness so that each dataset is different – could with very little labor be used to generate unique datasets for the thousands. exams in different disciplines, according to a study published in the Journal of Chemical Education.
“If you have an exam that counts a lot for a degree and it travels online, that offers an opportunity to cheat and even entices it,” said Alison Hill, senior lecturer in biosciences who co-wrote the article with his colleague from Exeter, Nicholas. Harmer.
“This [online cheating] doesn’t just happen in academia – my husband is a Devon chess champion, and when chess moved online during the pandemic there were reports of people using computers to help them win.
One method of avoiding collusion on data analysis questions is to limit exams to one hour. However, this brief exam window unfairly penalized students with poor internet connections or those based in different time zones who often had to start their tests at 3 a.m., explained Dr Hill.
But the student’s preferred option of a 24-hour review window was an invitation for students to share their answers, she argued.
“We have seen in other territories how once the diary goes live, a WhatsApp group is set up immediately – people just see this type of sharing as a good investment of their time,” said Dr Hill, who argued that relying on the academic honor of codes to stop cheating would be “completely naive.”
“We can’t completely stop cheating, but if each student has their own set of data – with the same question – the cost-benefit of cheating is no longer in the student’s favor as they will have to redo the work. . [for a classmate] with a different set of information, ”said Dr Hill.
The design of cheating by creating different datasets could be applied to most data-rich exams – with answer sheets generated automatically for each article – the newspaper article claims.
This type of test design would be far more effective than some pilot-driven exam monitoring techniques during the pandemic, such as webcam monitoring of these in-session exams, which some students easily bypassed, Dr Hill said.
“Students in confinement will find a way around these kinds of rules,” she said.