The most exciting phrase to hear in science, the one that heralds new discoveries, is not “Eureka” but “That’s funny...”
- Isaac Asimov
Newton’s law of universal gravitation made successful predictions for hundreds of years. Then one odd anomaly in Mercury’s orbit changed everything.[1,2] It eventually led to one of the most successful discoveries in modern physics, namely the theory of General Relativity which was formulated by Albert Einstein over a hundred years ago. It has been one of the most well-tested theories in science and no prediction made by the theory has been falsified so far. GPS and many other things would not exist without it. Yet, it all started with a simple data anomaly.
This year of data anomaly in higher education also provides ample opportunity to learn. It is true that anomalies in data science are a double-edged sword; on the one hand, it could impact current predictions and future modeling, but on the other hand, there is an opportunity to learn from data that would otherwise not have been captured.
For the remainder of this post, it may be helpful to look at this from a couple of different angles, namely the current (now/summer) tactics, the September postmortem diagnosis (after matriculation), and moving forward with future model updates. We will discuss each of these below, starting with current tactics.
It is impossible to know right now how exactly the pandemic is going to affect the accuracy of the current predictions (especially as far as variables that are not captured directly in the data are concerned), although keeping a close eye on indirect indicators such as deposit trends will be very insightful. While we did not choose to be in this current situation, we can certainly determine how we react to it. We want to use all the data and insights that are available to be proactive and consider the following courses of action right now or over the course of the summer (this is not meant to be an exhaustive list):
September is the time when the true impact of the pandemic on yield becomes known and we can start answering questions about the impact with some certainty using data and outcomes. This explanation can provide important insights on how to move forward. Here are some examples of analyses that could be performed (many others exist):
As was pointed out in the previous section, the full impact of the pandemic on the 2020 enrollment year will not be exactly known until September. Depending on the data and insights that arise, several possible scenarios would have to be considered moving forward (Othot investigated all of these in an update to this post.):
The anomaly in Mercury’s orbit provided Einstein with an opportunity to learn something new that was of tremendous value. Similarly, the 2020 anomaly in higher education due to the pandemic can be used to study the data and modeling and gain new actionable insights!
Do you want to know what happened with the Year of Data Anomaly? Read the update, "Five Modeling Lessons Learned From the Pandemic."
2. Coincidentally, Mercury was also the name of the first release of the Othot Platform years ago, and we are currently hard at work on the 3.2 Earth release (4.0 Mars will be next).