How to choose a teacher: what the research says actually matters

UsSalam-o-AlaiKum.
In 2013 I wrote "10 advices for students to choose the best teacher". It was honest, and it came from experience. But experience is one man's memory. Since then I have read what researchers found when they followed thousands of classrooms and measured what the students actually learnt. Some of my old advice survives. Some of it does not. This is the new version, with every claim tied to its source at the bottom.
One thing first. I am a teacher, so I am one of the people you are choosing between. Use this list on me as well.
The short version
- The teacher matters more than the school building, the class size or the fee.
- Fame, results posters, thick notes and years of experience tell you very little.
- What tells you a lot: how the teacher explains, questions, checks and gives feedback.
- You can see most of this in one demo class if you know what to look for.
- Tuition helps when the group is small and it follows what school is teaching.
- Online teaching works. Distance is no longer a reason to settle.
- No teacher can do the learning for you. The best one makes you do it well.
Why the choice matters
Economists have measured how much a year with one teacher is worth compared with a year with another. Eric Hanushek's review found that some teachers produce a year and a half of learning in one school year while others, with similar students, produce half a year. [1] A large American study that followed more than a million children found that students of the more effective teachers were more likely to reach university and earned more as adults. [2]
The same research carries a surprise. Very little of that difference is explained by the things printed on a prospectus. A higher degree showed no consistent link with how much students learnt, and after the first few years, extra experience did not either. [1] [3] So you cannot choose from a CV. You have to look at the teaching.
What does not tell you much
These are the things families usually go by. The evidence is not kind to them.
- Fame and popularity. The largest review of university course ratings carries its finding in its title: student ratings of teaching and student learning are not related. [4] A teacher can be loved and still not teach much. A teacher can be liked less and teach a great deal.
- Results posters. A wall of A* faces shows who enrolled, not what the teaching added. A review of private tutoring across Asia for the Asian Development Bank warns that tutoring providers in many countries misrepresent how effective they are. [5] Ask instead what happened to students who started weak.
- Thick notes. Reading and highlighting notes is among the least effective ways to study. [6] A teacher who hands you three hundred pages has given you paper, not understanding.
- Years of experience and degrees. See above. Knowledge of the subject and of how students get it wrong matters a great deal. The certificate does not measure it. [7]
- A small class, and nothing else. The Education Endowment Foundation rates reducing class size at about one extra month of progress in a year, and only if the teacher then teaches differently. [8]
- Crash courses and guess papers. Cramming is one of the weaker ways to learn, and spaced practice over weeks is one of the strongest. [6] My 2013 post called one or two day revision classes "absolutely necessary". I was wrong. They help only on top of steady revision.
- Big promises. A review of 96 randomised studies of tutoring found a real but modest average gain, about a third of a standard deviation. [9] None of them found the miracle some adverts suggest. [10] Be careful with anyone who guarantees a grade.
Prefer to listen? The EduTalk above covers this post in a few minutes.
Twelve signs you can check in a demo class
Three large reviews of what good teaching looks like agree with each other closely: the Sutton Trust's "What makes great teaching?", the Great Teaching Toolkit and Barak Rosenshine's "Principles of Instruction". [7] [11] [12] Cambridge's own Teacher Standards ask for the same things. [13] Put together, they give a list you can take into any classroom.
- Knows the syllabus cold. Can explain one idea in more than one way when the first way fails.
- Knows your mistakes before you make them. In a study of 181 science teachers, the ones who could predict their students' most common wrong answer got bigger gains than those who only knew the right one. [14]
- Starts by reviewing. A few minutes on last week's work before anything new.
- Teaches in small steps. A little new material, then practice, then the next step.
- Shows a worked example. Solves one in front of you and says the thinking aloud.
- Asks a lot of questions. And checks the answers of the whole class, not only the three who always raise a hand.
- Gives feedback you can use. Not "good" or "work harder" but what to change. Feedback is among the most effective things in the Education Endowment Foundation's toolkit, at about six extra months of progress. [8]
- Sets homework that follows the lesson, and marks it.
- Tests often and returns to old topics. Practice testing and spacing are the two study methods with the strongest evidence. [6]
- Teaches you to check your own work. Planning an answer, then judging it against the mark scheme.
- Runs an orderly room. Clear rules, applied the same way to everyone, little time wasted.
- Expects a lot from every student and is respectful to all of them. Not strict for the sake of it. Not soft either.
Rosenshine adds a useful number. In a good lesson students get about 80 percent of the practice questions right. [12] Much lower and the class is lost. Much higher and nobody is being stretched.
Seven questions to ask a current student
Do not ask "is he good?". You will get an opinion about personality. The Gates Foundation's Measures of Effective Teaching project, with 3,000 teachers, found that students can identify effective teaching when they are asked about specific things that happen in the room. [15] So ask these.
- When you do not understand, does the teacher notice?
- Does the teacher explain it another way, or repeat the same words louder?
- Do you get your work back with comments that show how to improve?
- Do you learn to correct your own mistakes?
- Is class time used, or does a lot of it drift?
- Does the teacher push you to think, or only to copy?
- Are weak students treated with the same respect as strong ones?
If you are choosing tuition
Paid tuition is now ordinary in Pakistan. The ASER Pakistan 2025 survey found 44 percent of private school children and 24 percent of government school children taking it. [16] The evidence says tuition can work, under conditions.
- Small. One to one tuition is rated at about five extra months of progress and small group tuition at about four. The benefit falls off once a group passes six or seven. [8] A "tuition" of sixty students is a second school, not tuition.
- Linked to school. It works best when it is additional to normal lessons and tied to what they are covering. [8]
- Short and regular. Several sessions a week over some weeks beat one long sitting.
- Taught by a trained teacher. In the tutoring studies, teachers got better results than volunteers. [9]
- Online is fine. Randomised trials in Spain and Italy found that fully online tutoring raised mathematics scores by about a quarter of a standard deviation. [17] [18] In 2013 I told you to choose a centre near home. Today a good teacher far away beats an average one next door.
One warning from the Asian Development Bank study. Where school teachers also tutor their own students for money, some hold back in the regular lesson to keep the demand alive. [5] If your child's class teacher is urging you towards their own evening batch, think about that.
What my 2013 post got wrong
- It told you to bargain on the fee. Price tells you nothing about teaching, so I have dropped it.
- It praised teachers with many online videos and notes. Resources are not learning. Practice with feedback is.
- It called last-minute revision classes "absolutely necessary". The evidence favours spacing.
- It told you to ask about "most and least appearing topics". That is topic prediction, and it fails students every year.
- It treated past results as a good indicator. Ask about improvement from a starting point instead.
What still stands: take a demo class, ask how to reach the teacher when you are stuck, and do not sit through back-to-back evening classes when you are tired.
The part no teacher can do
The best teacher in the world gives you clear explanations, good questions and honest feedback. The learning still happens in your own head, when you close the book and try to recall, and when you sit a past paper and mark it properly. Choose a teacher who makes you do that. Then do it.
The past papers with mark schemes, and the practice with marking, are free on cswithzak.com.
ZAK
References
This post was published on 14 November 2026, and every source was checked before it went out. Most of this research comes from schools in the United States and Europe. It is the best evidence available, and it should be applied to our own classrooms with judgement.
- Hanushek, E.A. (2011). The economic value of higher teacher quality. Economics of Education Review, 30(3), 466 to 479.
- Chetty, R., Friedman, J.N. and Rockoff, J.E. (2014). Measuring the Impacts of Teachers II. American Economic Review, 104(9).
- Rivkin, S.G., Hanushek, E.A. and Kain, J.F. (2005). Teachers, Schools, and Academic Achievement. Econometrica, 73(2).
- Uttl, B., White, C.A. and Wong Gonzalez, D. (2017). Meta-analysis of faculty's teaching effectiveness: Student evaluation of teaching ratings and student learning are not related. Studies in Educational Evaluation, 54.
- Bray, M. and Lykins, C. (2012). Shadow Education: Private Supplementary Tutoring and Its Implications for Policy Makers in Asia. Asian Development Bank.
- Dunlosky, J. and others (2013). Improving Students' Learning With Effective Learning Techniques. Psychological Science in the Public Interest, 14(1). Summary for teachers: Strengthening the Student Toolbox, American Educator, 2013.
- Coe, R., Aloisi, C., Higgins, S. and Elliot Major, L. (2014). What makes great teaching? Review of the underpinning research. Sutton Trust.
- Education Endowment Foundation. Teaching and Learning Toolkit: feedback, one to one tuition, small group tuition, homework and reducing class size.
- Nickow, A., Oreopoulos, P. and Quan, V. (2020). The Impressive Effects of Tutoring on PreK-12 Learning. NBER Working Paper 27476.
- von Hippel, P.T. Two-Sigma Tutoring: Separating Science Fiction from Science Fact. Education Next.
- Evidence Based Education. Great Teaching Toolkit: Evidence Review.
- Rosenshine, B. (2012). Principles of Instruction: Research-Based Strategies That All Teachers Should Know. American Educator.
- Cambridge International Education. Cambridge Teacher Standards.
- Sadler, P.M. and others (2013). The Influence of Teachers' Knowledge on Student Learning in Middle School Physical Science Classrooms. American Educational Research Journal, 50(5).
- Bill and Melinda Gates Foundation, MET project. Ensuring Fair and Reliable Measures of Effective Teaching (2013) and Asking Students about Teaching (2012).
- ASER Pakistan. Annual Status of Education Report 2025, national (provisional).
- Gortazar, L., Hupkau, C. and Roldán-Monés, A. (2024). Online tutoring works: Experimental evidence from a program with vulnerable children. Journal of Public Economics, 232.
- Carlana, M. and La Ferrara, E. (2024). Apart but Connected: Online Tutoring, Cognitive Outcomes, and Soft Skills. NBER Working Paper 32272.
Comments