When we were in the Next 36, my fellow participants and I all had shared a very similar experience — applying to university.
We sent over our transcripts, and we waited, fingers and toes crossed, hoping for an acceptance based on the limited information we’d been asked to provide.
The current admissions process typically doesn’t allow applicants to tell their story. Excellent students are missed by admissions teams every year because of test scores and transcripts alone, even if they have a great deal to contribute to the classroom.
That’s why we built Kira. We wanted to bring visibility to the applicants who otherwise might be missed. We combine asynchronous, timed video and timed written assessments with the rest of the application to help showcase the real people behind the application and not just the numbers.
As we started talking to admissions teams, we realized how we could build a product that would genuinely improve the process for both applicants and reviewers.
In 2012, the majority of colleges and universities around the world were still transitioning from paper applications to digital systems while trying to manage an increase in applicants each year. For U.S. schools, this rise in applicants has only been further bolstered by an increased demand for enrollment from abroad. In 2016 alone, 1.2 million international students attended American colleges and universities, nearly triple the number of international students in the 1990s. This year, with recent political developments, we’re seeing a jump in international applicants in Canada as well.
With more applicants, comes more competition.
Admissions decisions will always require human judgment, but data-driven technology can enhance human capabilities and help inform decisions.
Fraud is one problem: Some students are willing to do whatever it takes to get in, including paid essay services, plagiarism, and fake transcripts. Bias is another: The more criteria an application includes, such as essays or interviews, the more subjective the assessment becomes, opening each applicant up to reviewer bias.
With all of these challenges, admissions teams are understandably overwhelmed.
Holistic assessment, which means evaluating the whole student across a range of criteria, presents a compelling solution to current admissions woes. Those conducting holistic assessment have found it leads to enrolling a more diverse, more engaged, and generally stronger student cohorts.
However, schools are hesitant to fully adopt this method because of the time and resources needed to conduct such thorough assessments. In a survey by the Council of Graduate Schools last year, 58% of all survey respondents reported time as a barrier to conducting a holistic review.
On one side, there are hundreds of thousands of applicants trying desperately to understand what they can do to stand out to the admissions committee. And on the other, there’s the admissions committee, trying to give a fair and balanced review to hundreds of thousands of applicants, each attempting to stand out.
Smarter technology can bridge the gap between efficient assessment and effective, fair reviewing.
To start, automating the tedious, administrative components required to make holistic assessment successful will save teams an immense amount of time. Using asynchronous, timed video, schools no longer need to coordinate and schedule video calls amid different time zones. In Kira, evaluations for each applicant are tracked digitally, using pre-established rubrics to ensure consistency, and then averaged and weighted in the platform based on the program’s specific admissions criteria.
One school we work with cited they shaved an estimated 73 days out of their admissions cycle, while another said they think they’ve cut their interview time in half. This is time they can spend better-focused on selecting the right applicants.
Looking ahead, artificial intelligence will become an integrated part of the education system, from admissions decisions to curriculum development.
We will be able to use machine learning to understand the effectiveness of admissions criteria in predicting student outcomes.
For example, technology can augment the decision-making process and lead to better decisions. In 2014, researchers at the University of Minnesota found that algorithms beat human judgment when it came to hiring decisions. Humans are great at conducting interviews but terrible at weighing results. They found that a simple equation, when fed with evaluations from different reviewers, actually outperformed human decisions by at least 25 percent.
Humans can be so deeply susceptible to biases. Factors like a comment an applicant made on Twitter, the colour of her top, or her perspective on a topic irrelevant to the job at hand, can subconsciously influence how they evaluate the candidate.
In college admissions, this problem only multiplies due to the sheer volume of applicants. Decision-makers choose from hundreds, sometimes thousands, of unique individuals, for only a few seats in the classroom.
Using our platform, multiple members of a team can review an application and enter their feedback. Kira serves as the unbiased committee member that presents the average, weighted performance back to the team to make a decision. The algorithm helps schools see how an applicant performed across different reviewers and on different criteria, leading to fairer decisions.
Next, we will be able to use machine learning to understand the effectiveness of admissions criteria in predicting student outcomes. Currently, 81% of graduate school staff respondents call for more data that demonstrate the link between admissions criteria and student success in graduate school.
Today, noncognitive assessment has been adopted by progressive organizations, such as the Bill & Melinda Gates Foundation, and has proven to yield better results for student success and retention than standardized tests.
We hope to further demonstrate this link to schools by tapping into the data of hundreds of thousands of admissions decisions, and connecting that information to student outcomes. This will not only reveal student success criteria, but will help inform curricula that will have the greatest impact on each individual student’s development.
Admissions decisions will always require human judgment, but data-driven technology can enhance human capabilities and help inform and improve decisions without human error or bias.
We’re excited to be at the forefront of this shift towards machine learning in admissions, but we’re also so thrilled to create opportunities for the applicants, waiting anxiously on a decision that could influence the rest of their lives, to tell their story and be heard.
Discover more predictions on the next 150 years of tech at whatsnowwhatsnext.ca.