Hey there, Reader,

In this edition, we explore new approaches to college admissions and the power of predictive models....

Novel Approaches to Admissions with Loper: An Interview with Sam Bernstein

In my recent conversation with Sam Bernstein, Co-Founder of Loper, we delved into how his company is revolutionizing the college admissions process....

Loper is rethinking how students find their ideal schools by creating personalized matches between students and institutions - using large data sets.

The platform offers features like fast-track admissions and gamification, making the process more engaging for students while helping colleges better target the right student in a crowded marketplace.

Sam’s insights highlighted how Loper not only enhances the student experience but also empowers institutions to stand out and reach the right audience.

[Excerpts from Conversation Here]

The Power of Predictive Models in College Access: How It Works and What It Predicted

In a recent project, we built a predictive model using ChatGPT to forecast student success based on key factors like test scores, demographics, and socioeconomic status.

The best part?

The model we built in days outperformed some of the fanciest software on the market...

Here’s a breakdown of how we did it and what it revealed:

1. Data Collection & Preparation:​
   We gathered student data (e.g., math, reading, and writing scores) and cleaned it to ensure accuracy. ChatGPT helped prioritize key features such as test scores and parental education.
2. Model Selection & Training:​
   Using logistic regression, we trained the model on historical student outcomes, which allowed us to predict academic challenges with high accuracy.
3. Testing & Validation:​
   We evaluated the model’s performance with metrics like a confusion matrix and ROC curve, confirming its ability to identify at-risk students early.

Key Insights

• Math scores were a strong predictor of success, with lower scores linked to academic struggles.
• Parental education influenced performance, with students whose parents had higher education levels performing better.
• Test preparation participation improved outcomes, highlighting the need for expanded access to prep resources.

Takeaway:​
Predictive models give institutions a powerful tool to make data-driven decisions, improve student success, and allocate resources more effectively.

That said, they're only as good as the data leveraged - and the model itself...

As always, we hope these insights help you navigate the challenges and opportunities facing higher education today.

If you’d like to discuss any of these topics in more depth, don’t hesitate to reach out.

-Mike O’Connor

O'Connor & Associates

​​

unsubscribe here.