One of the difficult challenges for any institution of higher learning is delivering tests to students who, in their eagerness to compete for good grades, might be tempted to cooperate with other students on tests or quizzes or use other means to improve their scores. That concern is certainly felt at Penn State University, a school located in University Park, Pennsylvania but with 19 satellite campuses and a worldwide reach.

In order to instruct students in these far flung places, the school has joined the legion of those offering an online curriculum. Dr. Paul Howell, who teaches an online version of his general education introductory class on material science and engineering, not only wanted to make his course available to a wider audience but also to give those on campus more flexibility in how they took the course.

The strategy paid off. Currently, for example, Howell has 250 online students and 150 who come to class. Whether they listen to a live lecture or get the information on their computer, all students take the course’s weekly quizzes on their computers. Each week of the semester, students log on to the university’s course management system from wherever they are and take the 20-question, end-of-week quiz.

“I try to make each quiz into a positive learning experience,” he said. “All my exams are designed to be open book so the students are allowed to use the course textbook or anything else to get the right answers. I am more interested in their understanding the concepts to get the right answer.”

However, to limit student cooperation, individual enrollees get different exams. Each gets a custom delivered quiz with 20 questions – different from every other student’s – taken from a database of 250 questions for that assessment. The challenge becomes one of ensuring that each student gets questions that are even-handed in both content and level of difficulty.

According to Penn State Senior Measurement Specialist Ralph Locklin, the school relies upon Questionmark Perception’s ability to save response data across semesters so he and others can evaluate each and every question in terms of appropriateness and difficulty.

“We had several instances where we could identify a problem,” he said. “Some of the questions in the group were very, very difficult and others were not nearly so. That helped us reorganize the database in a way that permitted us to be sure that every student would get one of these more difficult questions.”

By accumulating enough response data to evaluate the performance of each of the questions, Locklin’s group was able to make adjustments to the sampling within the database easily. He found that Perception’s ability to provide a complete item analysis enabled his team to make the exams even handed and represent each student’s knowledge level more accurately.

He observed that his students appreciate being able to get instant feedback on the tests. That capability actually further enhances his goal to make the quizzes learning experiences.

“As soon as they submit their quiz they will get immediate feedback,” he said. “The feedback not only indicates whether their answer is right or wrong but also my feedback about why an answer is right or why an answer is wrong. Students tell us that they appreciate being able to read the feedback on the questions that they answered incorrectly.”

The result? Students are not only quizzed using this system; they also learn.