There is a concept used in the software development world, called “technical debt”. Coined by Ward Cunningham, it refers to the extra development work that arises when code that’s easy to implement now is used instead of applying the best overall solution. In other words: doing what’s easy right now ends up in a lot more extra work in the future.
Technical debt creeps in slowly. It weighs down the effectiveness of the product, and eventually the company must erase it by rebuilding the underlying code (which is neither easy nor fast). See, for example, the story of Photoshop – first built in 1987, with each new batch of developers it gains a new layer of code. Each version must be carefully written over the former so as not to collapse the entire house of cards, and apocryphal stories states there are parts of the code that nobody knows about. As the features piled on, the underlying code bloated to millions of lines.
If Adobe decided to rebuild Photoshop in 2017, the effort might take over a decade and possibly never be completed. If Paris has its catacombs, Photoshop has its code.
Higher education has its own form of technical debt – the learning design debt. It refers to the limitations of yesterday’s learning design, weighing down the effectiveness of higher education for today’s students. Teaching methods that worked in the 1980s do not work now, yet institutes of higher education are reluctant to change. The accumulating downside, and soon the price we will have to pay, is the degree attainment gap.
In short, what worked for yesterday’s students aren’t working for today’s.
The new gatekeeper of higher education
Traditionally, people have only worried about getting admitted into a college, and it’s true that admittance remains a major factor in the accessibility of higher education. But a newer, previously largely ignored problem is coming to light – many students who have been admitted into universities are failing to thrive. Admissions is no longer the great gatekeeper of higher education.
“Failure to thrive” is a term borrowed from pediatric medicine, when an infant – sometimes for no reason at all – is unable to gain weight as expected. In the last 20 years, more than 31 million students in the United States left college without graduating. To put that into context, approximately one fifth of the U.S population over the age of 25 has some college experience, but no degree.
Initiatives have been taken to combat the gap. The Pell Grant program was launched in 1972 in an effort to achieve degree completion goals. Its effectiveness is debated – a 2014 report by The Education Trust found that Pell Grant recipients lagged behind non-Pell students in graduating by an average of 5.7%. And yet, the same study also found that colleges that serve similar student populations can have “wildly” different outcomes for grant recipients. The question is, why?
One possible answer is that students are going to the wrong colleges. The perceived superiority of STEM (science, technology, engineering, and medicine) has not waned, especially not in a world driven by drama in Silicon Valley. The stigma of a liberal arts degree is as strong as ever. Too many students are being pushed into courses they are uninterested in at best, and incapable of succeeding in at worst.
The second possible answer is that the learning design gap is finally catching up to us. The “code” of higher education is no longer capable of keeping up. We’re still less effective in serving low-income students
. In 1970, 72% of upper- and middle-class students were graduating from college, and 28% of the bottom two quartiles were earning degrees. In 2014, those rates were 77% and 23%. While university admissions have become better at designing for nontraditional students, the numbers keep sinking.
Rewriting the education code
One-third of all students in four-year colleges are clustered in only 25 of the school’s courses. These general education “gateway” courses are the sort of large-enrollment introductory courses that, in person, are typically held in auditoriums rather than classrooms: Introduction to Biology, for example, or Physics 101.
These auditoriums are yesterday’s design solution to the scale problem, and they are where most of the learning design gap in higher education is accumulating. One teacher lecturing to hundreds of students is a cheap way of solving admission intakes that grow every year. Most students who fail or drop out do so within their first two years of college, when they are enrolled in introductory courses. Students who come from lower performing schools already struggle in colleges – putting them in a mass-market class ends up in dismal retention numbers across the board.
In 2006, George State University discovered that students weren’t completing their101 math courses. They then redesigned the course and launched its pilot in 2006
, using adaptive learning technology and a new hybrid format. It more than halved the drop/fail/withdraw rate.
Georgia State’s experiment took place 11 years ago. Since then, advances in edtech and learning science have given us the ability to infuse evidence-based learning design into large-enrollment courses in a scalable way. The goal is not to make the course outcome easier; the rigor of the curriculum must be maintained if we are to have capable graduates. Instead, the goal is to use learning design and the resulting data alongside traditional teaching practices.
Combining learning science with data
In 2009, Dr. John Hattie – a Professor of Education at the University of Auckland published a meta-analysis
of 800 meta-analyses. Dr. Hattie’s study aggregated, correlated, and ranked factors that most improved learning outcomes. The resulting meta-analysis showed that targeted feedback and the trust built by teachers with their students, were the most important factors in learning. Metacognition – a student’s ability to reflect upon their own learning – ranked highly as well.
Personalised learning technology is able to bake that formative practice and metacognition into an online course, giving students goal-directed practice with timely hints and feedback. The truly interesting bit is that this practice and feedback can be individualised to each student based on their own learning process.
As students learn, the learning system learns about them, generating data in real time and feeding it to instructors. And as students complete activities and assignments, so the system collects and analyzes the data, predicting the progress of a student.
And personalised learning at scale is possible.
New authoring tools simplify and accelerate collaborative learning design, enabling institutions to rapidly reiterate (to borrow another term from software development) problem courses. It aligns with other academic support as well, forming a deeply customised educational experience. Learning data generated from student interactions within the online learning environment triggers individualised support and coaching services, solving two problems in one – where students aren’t learning the way they need to, and students are going to colleges they shouldn’t be.
Imagine a world where a student is automatically matched to their best fit college
. Where every lesson plan and every curriculum is customised around the student’s own unique learning methods. A world where “no child left behind” isn’t just lip service. In this, data and machine learning will lead the way.
While we must not allow education to lag behind the rest of society in adopting new tech, do not make the mistake of assuming the role of teachers is obsolete. While technology will, eventually, erase the necessity of mankind in several industries, it’s highly unlikely teaching or counselling will ever be one of them.
Predicting a student’s grades is simple enough. Predicting a student’s ability to thrive relies on a hybrid blend of technology and humanity – the soft touch of a teacher familiar with a young adult’s needs and wants.