How to study for a test in one week: the day-by-day plan (and what changes for finals, three days, or one)

To study for a test in one week, work backwards from the date: on day 7 list every idea the test covers and test yourself once on each to find the gaps; days 6 to 3 are short daily sessions that attempt problems or recall before rereading, spaced so each idea comes back at least twice; day 2 is a full timed practice run; day 1 is a light review and an early night. The plan below gives each day its tasks, then shows what to cut with three days or one, and how to run several finals in the same week.

The plan rests on the four moves in the hub of this section: answer first, space it, mix it, sleep. Here they are turned into dates.

First, the list: what is actually on the test

Before any studying, write down what the test covers, as questions. Not "chapter 4", and not the pile of notes, but the ideas, one line each, from the course outline or the textbook's section headings. The plan that follows is per idea, not per page, and you cannot plan per idea until the ideas are named.

For a Grade 12 calculus (MCV4U) test on derivatives, the list is five lines:

1. Can I find a derivative from first principles, using the limit definition?
2. Can I apply the power rule, including to a linear term and a constant?
3. Can I use the product rule and the quotient rule, and tell which one a question needs?
4. Can I use the chain rule on a composite function?
5. Can I differentiate implicitly when y is not isolated?

Five questions. The whole week is spent getting an honest yes to each. A list of ideas written as questions is the difference between studying for a test and studying near one.

Day 7: the diagnostic

Take the list and attempt each idea once with the notes closed: one problem, or a blank page with the idea's name at the top and everything you can recall under it. Then open the notes and mark yourself into three piles.

• Can do. You got it right without hesitating. This pile gets one spaced check later in the week and no more.
• Shaky. You got there, but slowly, or with one slip you caught. This pile gets a session each.
• Cannot. You did not know where to start, or the answer was wrong and you could not see why. This pile gets the most hours and the first days.

Mark honestly. A problem you "would have got" is in the shaky pile. The rules for marking yourself without flattering yourself are in test yourself honestly. The piles decide the hours: the "cannot" pile gets the most, the "can do" pile gets the least, and nothing gets rereading.

Days 6 to 3: the spaced sessions

Four days, one session each of 40 to 60 minutes. Each session has three parts, in this order.

Day	New idea to work (from the "cannot" pile first)	Idea to revisit	Last 10 minutes
Day 6	The hardest "cannot" idea	Yesterday's diagnostic misses, cold	A mixed set: one problem from each pile
Day 5	The next "cannot" idea	Day 6's new idea, one problem	A mixed set
Day 4	The first "shaky" idea	Day 5's new idea	A mixed set
Day 3	The remaining "shaky" idea	Day 4's new idea, and one "can do" check	A mixed set across every idea on the list

The rule behind the middle column: every idea is attempted on at least two different days. That is the spacing effect, and it is one of the best-supported findings in the research on learning (Cepeda et al. 2006): the same minutes spread over two days produce better recall than the same minutes on one day, because the second attempt comes after you have started to forget. Why two separate days beat one long one explains the schedule in full.

Inside the "new idea" part, a session looks like this: read one worked example slowly, close it, redo the same problem from the question alone, then do a sibling problem with different numbers. Mark against the solutions and write every wrong line in an error log, with the right line beside it. Then redo the missed ones cold before you move on.

The last ten minutes are always a mixed set: problems from different ideas, shuffled, so that you practise choosing which rule applies. Blocked practice (ten chain-rule problems in a row) teaches the recipe; the shuffled set teaches the choice, and the choice is what the test asks for (Rohrer & Taylor 2007). It will feel worse than blocked practice. That is the point.

Day 2: the practice run

Build a practice test: the same number of questions as the real one, covering every idea on the list, with a time limit. Sit it under test conditions. No notes, no phone, no pausing the clock. Then mark it.

A practice test does two things. It is the strongest single study technique the research has (Roediger & Karpicke 2006: tested students remembered more a week later than students who reread), and it rehearses the conditions. You want to have already met the feeling of a blank moment with a clock running, so that on test day it is familiar rather than frightening. Keep the claim modest: practising under the test's conditions helps you perform under them; it does not make the test easy.

Spend the rest of the session only on what went wrong. Every wrong line goes in the error log. If the test is a standardized one rather than a school test, the exams section of this blog has the practice-test method for those.

Day 1: light review, early night

Recall the error log once: read each wrong line, say the right line before you uncover it. Redo two problems, one from the "cannot" pile and one mixed. Stop by early evening.

Then sleep. What you studied this week is consolidated while you sleep (Rasch & Born 2013), and a short night costs you recall and attention in the room tomorrow; what one night without sleep costs has the measured figures. If it is the nerves rather than the material that keep you up, if the nerves are the problem is the guide for that, and it is worth reading on day 2 rather than day 1.

What not to do on day 1: start new material, pull an all-nighter, or reread everything one last time. New material will not consolidate in time. The all-nighter costs more than it buys. The last reread produces a feeling of familiarity that the first question will remove.

Worked example: the whole week for one calculus test

The MCV4U list above, a test next Monday, one student.

Day	Date	What happens
7	Mon	Five questions written. Diagnostic: one problem each, notes closed. Piles: can do (first principles, power rule), shaky (product and quotient rules), cannot (chain rule, implicit differentiation). One slip noted in the power rule, below.
6	Tue	New: chain rule. One worked example, redone cold, then a sibling. Revisit: the power-rule slip from Monday, two problems. Mixed set, 10 minutes.
5	Wed	New: implicit differentiation, the same way. Revisit: chain rule, one problem. Mixed set.
4	Thu	New: product and quotient rules, with three problems that make you choose between them. Revisit: implicit differentiation. Mixed set.
3	Fri	Revisit: product and quotient (Thu), chain rule once more, and one first-principles problem as the "can do" check. Mixed set across all five ideas.
2	Sat	Practice run: ten questions, 45 minutes, timed, no notes. Mark. The rest of the hour on the wrong lines only.
1	Sun	Error log once, two problems, stop by 7 pm. Sleep.
Test	Mon	Breakfast. Error log once on the way in. Nothing new.

Here is Monday's power-rule problem and its marking, because this is where a slip hides. Differentiate f(x) = 3x⁴ − 2x. Power rule on the first term: 3 × 4 = 12, and the exponent drops by one, so 12x³. The second term is −2x, a linear term: its derivative is the coefficient, −2. So f′(x) = 12x³ − 2.

The common wrong line is f′(x) = 12x³ − 2x: the linear term left as it was, as if differentiating it did nothing. The student above wrote exactly that. The power rule itself was fine, so it stays in the "can do" pile, but the slip goes into the error log as "power rule on a linear term: the derivative of −2x is −2", and that line is what Tuesday's revisit column works on. Two problems on Tuesday, one on Friday, and the slip does not appear on Monday.

Three days left? One night?

Some of the week is already gone. Here is what the plan becomes.

Three days:

1. Day 3: write the list and run the diagnostic. Sort into the three piles. Spend the rest of the session on the hardest "cannot" idea: worked example, redo cold, sibling.
2. Day 2: the remaining "cannot" and "shaky" ideas, the same way, then a mixed set of ten problems across the whole list, timed.
3. Day 1: the error log once, two problems, an early night.

One night:

1. The list and the diagnostic, one hour.
2. The "cannot" pile by worked example, redo cold and sibling; then the "shaky" pile by recall.
3. The error log once, then stop in time for a full night.

One night buys recall of what you half-know and the procedures you have done before. It does not buy a new idea learned deeply, and most of what it does buy fades within a week. What still works the night before has the evening hour by hour, with the honest limit.

Several finals in the same week

Finals change two things: the list is longer (four or five units instead of one), and there are several of them. Both answers are the same: spread the courses across the days rather than giving each course a day of its own. Monday is calculus and chemistry, Tuesday is biology and calculus, Wednesday is chemistry and biology. Each course comes back on more days, which is the spacing rule applied across courses, and no course sits untouched for four days before its exam.

The earliest exam gets the first diagnostic. Each session is one course only; switching mid-session costs the restart. A weekly grid with a column per day and a row per course makes the whole thing visible on one page, and that grid is the weekly grid template in the study-plan guide. Start two to three weeks out for a final, at the same daily dose as the week above: the limit is the number of separate days each idea comes back, not the hours.

A math final has one more feature: every question is a problem, so the diagnostic and every session are problems, and the mixed set matters more than anywhere else. How to study math specifically is the version of this plan for a course where recall alone is never enough.

The week on one screen

1. Day 7. The list as questions. Diagnostic, notes closed. Three piles.
2. Days 6 to 3. One session a day: a new idea (worked example, redo cold, sibling), a revisit from two days ago, a mixed set to finish.
3. Day 2. A timed practice run under test conditions. Mark it. Work the wrong lines only.
4. Day 1. Error log once, two problems, early night.
5. Test day. Breakfast, error log once, nothing new.

This is chapter 9 of the free playbook, expanded; the playbook has the seven days as a one-page checklist you can print.