I listened to an interesting podcast recently featuring Liz Cohen, author of The Future of Tutoring. Although I intuitively know that tutoring works — after all, I see it firsthand every day — I always appreciate seeing the research that supports that intuition.
Through her research on large school districts implementing tutoring programs to address post-pandemic learning loss, Cohen found something striking but perhaps unsurprising to educators and parents alike: one-to-one tutoring was one of the fastest and most effective ways to improve academic outcomes.
While tutoring has become increasingly common in recent years, the research behind its effectiveness is not new.
Back in 1984, educational psychologist Benjamin Bloom published his influential paper, The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as One-to-One Tutoring. Bloom found that students receiving individualized tutoring dramatically outperformed students in traditional classroom settings — often by as much as two standard deviations.
In simple terms, the average tutored student performed better than nearly all students receiving conventional classroom instruction.
Bloom’s research helped highlight something educators have long understood: personalized instruction matters.
Why Tutoring Works
What makes tutoring so effective is not simply additional practice or homework help. Effective tutoring creates a learning environment that is difficult to replicate in a traditional classroom.
In a one-to-one or small-group setting, instruction can be tailored to:
- A student’s pace of learning
- Specific strengths and weaknesses
- Preferred learning style
- Confidence level
- Gaps in foundational knowledge
Students are also far more likely to ask questions when they feel comfortable with their tutor. In many classrooms, students may hesitate to admit confusion or ask for clarification. Tutoring creates space for mistakes, curiosity, and individualized feedback.
The relationship itself also matters more than many people realize.
One of the most interesting takeaways from recent tutoring research is that there was little difference in outcomes between virtual tutoring and in-person tutoring. The most important factors were:
- A low student-to-teacher ratio
- Consistency
- Alignment with school curriculum
- Frequent feedback
- Building a strong relationship with the tutor
That last point is especially important.
The Rise of AI Tutoring
At the same time, AI is rapidly transforming the educational landscape.
Tools such as Khan Academy’s Khanmigo and other AI-powered tutoring platforms promise highly personalized instruction available anytime and at scale. There is no question that AI can be an incredibly useful educational tool, but can it replace a traditional tutor?
The reality so far has fallen short of the more revolutionary claims.
What AI Still Cannot Replicate
For those of us who have worked closely with students, the limitations are obvious.
Great tutoring is not just about delivering information. It is about understanding the student sitting in front of you.
A skilled tutor can recognize when a student understands a process mechanically but lacks deeper conceptual understanding. They can tell when a student is overwhelmed, discouraged, or simply exhausted after a long school day. They can read the student’s facial expressions to see when the student is lost even when the student hasn’t yet asked a question.
Human tutors also build accountability and trust. Students are often more motivated to stay engaged when they feel connected to someone who genuinely knows and supports them.
AI can explain a math problem. A human tutor can recognize why a student keeps making the same mistake.
AI can generate an essay outline. A human tutor can encourage a hesitant student to believe they are capable of writing something meaningful.
AI can provide answers. Human tutors help students build confidence.
The Future of Tutoring Is Likely Hybrid
Rather than replacing tutors, AI will likely become one more tool that effective educators and students use strategically.
The research continues to point toward the same conclusion Bloom identified more than 40 years ago: individualized human support remains one of the most powerful educational interventions we have.
Technology can enhance learning. Relationships still drive it.