How to Study Math Effectively: Beyond Doing Problems

To study math effectively, alternate between studying worked examples and solving similar problems yourself, then mix problem types so you practice choosing the right method. Explain each step as you go, and keep a log of every mistake so you fix the cause, not just the answer.

That's different from what most students do, which is to finish the homework, check the answers, and move on. Homework sets usually group problems by section, so the heading tells you the method before you read the problem. You get fast at executing a method you were told to use. The exam asks you to pick the method yourself.

This guide shows how to study math in five steps, with a fading worked-example sequence, an error log template, and a weekly routine you can copy for any course from algebra to calculus.

Why doing homework isn't the same as studying math

Homework builds speed on one method at a time. That feels like learning, and in the short term it shows up as good homework scores. But it skips two skills the exam needs: recognizing which method a problem calls for, and recovering when you get stuck partway.

There's also a recognition trap. When you read a solution in the textbook or watch it on the board, every step makes sense. That sense of "yes, I follow" is not the same as being able to produce the next step on a blank page. The only way to find out which one you have is to close the solution and try.

The five steps below fix both problems. None of them needs much more time than you already spend; they change what you do with it.

Step 1: Learn from worked examples, then fade them out

When a topic is new, studying a worked example teaches you faster than struggling through problems blind. In a classic series of algebra experiments, 8th and 9th graders who alternated four worked examples with four similar problems took less time and did better on the post-test than students who simply solved all eight. The US Institute of Education Sciences turned this into a recommendation for math and science teachers: alternate worked examples with problems.

The catch: the benefit showed up mainly on problems with the same structure as the examples. That's why steps 2 and 3 add explanation and mixing.

You can do this yourself with any textbook:

1. Cover a worked example's solution and read only the problem.
2. Uncover one line at a time. Before each line, predict what comes next.
3. Solve the next exercise of the same type on your own, without looking back.
4. Check, then move to the next example.

Then fade the help. Research on fading has learners complete more and more of each example themselves, until they're solving whole problems. Combined with prompts to name the principle behind each step, fading improved both similar and more distant problems without adding study time. Here's what a fading sequence looks like for the chain rule:

Once you can solve a problem type without help, stop studying full examples for it. For experienced learners, extra worked examples lose their benefit and can even hurt, an effect researchers call the expertise reversal effect. Examples are scaffolding. Take them down when the wall stands on its own.

When the idea itself is abstract, such as fractions, functions or limits, start one level lower: a picture or physical case first, then strip it back to the symbols. Our guide to learning with concrete examples calls this concreteness fading and shows how to pair any definition with examples.

Step 2: Explain each step out loud

You learn far more from an example when you explain it to yourself as you read. For each line, ask:

• What rule makes this line true? ("Distributive property." "Chain rule: derivative of the outside, times derivative of the inside.")
• Why this method and not another? ("Product rule doesn't apply; this is a function inside a function.")
• What would change if the problem changed? ("If the exponent were negative, the same rule still works.")

A 2018 meta-analysis of 64 research reports found that prompting learners to self-explain produced a medium-sized benefit on average (g = 0.55), across subjects including math and science. The self-explanation guide covers the research and better prompts. To test a whole method, explain it start to finish as if teaching a friend who missed the class; that's the Feynman technique.

Where you hesitate, write a question mark in the margin. Those steps are your study list, and they often point to a gap from an earlier course, like fraction rules or logarithm laws.

Step 3: Practice with mixed problem sets

After you can do each type alone, stop practicing them in blocks. Mix them, so each problem might need any method you've learned. This is called interleaving, and most of the classroom evidence for it comes from math.

In a 2007 study, college students learned formulas for the volumes of four unusual solids. One group practiced each solid in a block; the other practiced them shuffled. During practice, the blocked group looked far better. A week later, the result flipped:

The same pattern held in a randomized trial across 54 seventh-grade classes: students who did mixed homework scored 61% on a surprise test a month later, against 38% for blocked homework.

To build a mixed set:

1. List the problem types you've learned this unit and last unit.
2. Pick two or three problems of each from old homework, the review section, or past exams.
3. Shuffle them so neighbors need different methods.
4. Before solving each one, write the method in the margin: "u-substitution," "elimination," "complete the square."
5. Mark every problem where you chose the wrong method. Those are worth more than any calculation slip.

Expect your scores to drop at first. That drop is the point: you're now practicing the decision the exam will test.

Step 4: Keep an error log

A wrong answer you check and forget will come back on the exam. An error log makes you name the cause and test the fix. Use a notebook page or a spreadsheet with five columns:

The Type column is where the value is. After two weeks, count them. If most of your errors are sign slips, you need a checking habit, not more practice problems. If most are method choice, you need more mixed sets. If most are setup, study more worked examples of how problems start. The research on why errors help learning is in learning from mistakes.

On re-test day, copy the original problem onto a blank page and solve it without looking at your fix. Only then compare.

Step 5: Test yourself on methods, not just answers

Math exams test two kinds of memory: how to carry out a method, and when to use it. Most students only practice the first. Add these:

• Method cards. On the front, a problem with no instructions. On the back, the method and why. Front: "∫ x eˣ dx." Back: "Integration by parts. One factor (x) gets simpler when differentiated, the other (eˣ) is easy to integrate. Let u = x, dv = eˣ dx. Answer: x eˣ − eˣ + C."
• Blank-page procedures. Write the steps of a method from memory: "Solving a linear system by elimination: 1) line up variables, 2) scale one equation, 3) add to eliminate…" Then check against your notes.
• Formula reconstruction. Instead of memorizing a formula as a string of symbols, practice deriving it from something simpler. The guide to memorizing formulas shows derivation chains for algebra, trig, and calculus.
• Spaced review. Return to each problem type after a few days, then a week, then a few weeks. The spaced repetition guide explains the timing.

Flashcards work for definitions, theorems, and method choice. They don't replace solving full problems. Most of your math study time should still be pencil-on-paper problems; the cards just keep the older material from fading.

How to study math each week: a routine to copy

This assumes one math course, two lectures a week, and about 45 to 60 minutes a day. Adjust the days to your timetable. If you're juggling several problem-heavy courses, study tips for engineering students shows how to run the same habits across all of them.

How to study for a math test in the last week

1. Seven days out: list every problem type the test can include, from the syllabus, homework headings, and review sheet.
2. Days 7 to 5: one timed mixed set per day drawn from all types, with the method labeled first.
3. Day 4: a full practice or past exam under exam conditions. No notes, no calculator if the exam bans it.
4. Days 3 and 2: go through your error log and the practice exam. Re-study only the types you missed, then re-test them cold.
5. Day 1: a short mixed set and a blank-page run through every method. Then stop and sleep.

For a full plan across several subjects, see how to prepare for exams. If one of them is a reading-heavy course like history, how to study history with timelines and causes uses the same recall habits.