Peter Bohacek 
Table of Contents
In general, most of us like to look for the most efficient way to work. I know I say it often: work smarter, not always harder. When students view science this way, it’s understandable why they look for shortcuts when learning science. These shortcuts present a challenge for engaging students in learning.
Learning takes effort and time, and students are busy. Why struggle with science homework or tests when it’s so easy to look up answers online? You can ask a friend, pay for an online homework solution provider, or use one of the new AI (Artificial Intelligence) tools like ChatGPT. But these unproductive shortcuts have negative side effects. Neither the student nor the instructor can tell whether anything has been learned.
As a science teacher with 20 years of experience, a science education researcher, and an edtech product creator, I’ve had the privilege of exploring science education from a variety of perspectives with a range of analytical tools. Drawing on my experience, I’ve developed an approach to help students avoid unproductive shortcuts.
Strategies to Guide Students to Engaging Students Without Ineffective Shortcuts
My approach is two-pronged. Firstly, teachers must prevent useful shortcuts. Second, we must provide a more productive alternative for students. Make the engagement and learning the “easy way,” and make the ‘hard way’ as searching to find answers without gaining any learning.
Simple Randomization
Some edtech systems allow teachers to randomize questions. For example, teachers can use Google Forms and Sheets to create questions where the numerical values provided in the questions are randomized. When teachers randomize values in questions, students can’t directly share answers. However, they can share simple methods for solving problems. For example, a student can say: “take the first number in the problem, square it, and divide by two times the second number.” Similarly, pasting the question text directly into a language model like ChatGPT often returns a correct response. Although randomized numeric values can provide some disincentive for students to look up answers, it’s not enough.
Deep Randomization
More sophisticated systems deeply randomize questions, by randomizing images, graphs, even the question text, in addition to numerical values. For example, if the teacher assigns questions about a specific graph, such as “on what day does the population reach carrying capacity?” Students must read and interpret the answer from the graph. If the graph has different values for each student if students share information, that will not be useful. Even homework services and AI won’t provide answers to questions that include randomized images.
Pivot Interactives is an example of a system that has pre-made assignments that include deep randomization, with students presented with different scenarios that test the same objective. Pivot Interactives is highly resistant to unproductive student shortcuts.
Question Pools for Engaging Students
Some edtech systems present students with a random selection of similar questions pulled from the same pool of questions. This means that students won’t all have the same questions. While this makes it more difficult for students to share answers directly, it doesn’t prevent students from using online help services nor AI-based tools. Online tools like Quia, and learning management systems like Canvas or Blackboard, allow instructors to use question pools.
Lockdown Browsers and Web Filters
Lockdown browsers can make it more difficult for students to use shortcuts by limiting what web resources are available during assessments. But if the student has a second device, such as a phone, they can skirt this by using the phone to search for answers. Similarly, web filters that prevent access to AI tools and other resources in schools can easily be bypassed by students who own multiple devices outside the school’s network. Educators have raised equity and privacy concerns about the use of these solutions.
Provide Simple and Adaptive Feedback for Engaging Students
Simply letting students know whether they are on the right track can have enormous benefits. In the past, students consistently reported that solving problems correctly gave them the confidence to tackle the next challenge. Most online systems offer this, including Discovery Education’s Science Techbooks.
Education researchers concluded long ago that working one-on-one with a tutor is the most effective way to learn. The tutor can remind students of previously learned concepts and point out useful connections in students’ progress. The tutor can also show how students’ work can be used to answer subsequent questions. Tutors notice mistakes in reasoning and can provide encouragement. Sophisticated edtech services often offer students adaptive feedback that acts like a skilled tutor.
In Summation
New technologies, those that enable students to search for answers without learning and those that seek to encourage students to persevere and learn, will continue to improve. For now, deep randomization and adaptive feedback provide both obstacles to the use of unproductive shortcuts and an appealing alternative for students who want to learn.
