How to Study for the LSAT While Working Full Time: Build the Week Around What Cannot Move

Test format · Build your week · Interactive planner · Interruptions · Worked examples · Try it yourself · Worksheets · Next step

By AccessPrep Editorial · Facts checked November 2, 2026

Keisha, Rafael and Mei, their jobs, schedules, study windows and practice logs are fictional illustrations, not real students or results. Every schedule here is illustrative, not a recommended number of hours, and nothing in this guide predicts a score or how long preparation will take. The window kinds, job sizes and rules are AccessPrep teaching proposals, not validated findings. AccessPrep and Lovare share a founder, and Lovare offers paid LSAT instruction and admissions consulting.

Plan the week from what cannot move, not from a target number of hours. List your fixed commitments, find the study windows that are actually left, and mark each one by two things: whether you are alert in it and how likely you are to be interrupted. Then put the five jobs of one review loop into those windows: a timed attempt, independent review before the key, resolving the misses and choosing one action, a transfer check on fresh questions, and a short evaluation. Place the timed attempt first, because it needs the rarest kind of window. Shorten jobs before you drop them, decide in advance what an interruption does, and judge each week by whether the loop closed, not by the hours you logged.

Who this is for and what success looks like

You work full time, or close to it, and your week is set by a job, a shift pattern or the people you look after. You may already have tried “two hours every weeknight” and watched the plan last until the first late meeting, the first night shift or the first sick child.

Success here is not a number of hours. It is a week in which one complete loop closes, from a timed attempt to a checked action, and a log that shows, week after week, whether the actions you practice appear on questions you have not seen before. If your week cannot hold even the smallest loop, success is learning that in one week rather than six, and knowing what to change.

This guide does not tell you how many hours or months to study. That depends on where you start and where you want to finish, and no schedule can promise a result. It gives you a way to find out how much usable study time your week really has, and to use it in an order that produces evidence.

What this adds to the other AccessPrep guides

The weekly study plan guide turns an error log into one practice action and lays out a focused two-hour week and a four-hour week as five blocks: collect evidence, review independently, resolve and choose, test transfer, and evaluate. This guide takes those five jobs as given and adds what that guide covers only in passing: deciding which of your real, uneven windows each job goes into, and what happens when a window disappears. Two of its calls become fixed rules here, because on a working schedule you make them tired: its answer to a missed session, “keep the sequence and reduce the workload,” becomes the move and shrink rules in Step 5, and its evaluation, which decides “whether to continue, simplify, or revise the action,” becomes the counts in Step 6.

Two other guides supply the methods inside the jobs. The blind-review guide explains how to review a timed attempt before you see the key. The plateau guide‘s diagnostic table answers “Results vary with interruptions, fatigue, or inconsistent timing” with a check of practice conditions: “make comparisons more consistent before attributing the change to a study method.” The window labels below are a way to do that on a working schedule. When the schedule is not the problem, the tutoring vs. course guide helps you decide what kind of help fits.

First, know which LSAT your blocks must fit

Choose the format for your intended test date before you size a study window. LSAC’s current specifications describe four 35-minute sections: two Logical Reasoning (LR), one Reading Comprehension (RC), and one unscored section. Its August 2027 announcement sets out a different arrangement for later administrations.

On small screens, tables scroll sideways.

The announced change embeds unscored questions within the three sections. LSAC says the question types, scoring scale and per-question time remain unchanged. Recheck its announcement before planning a later administration. We use the listed section lengths, rather than the announcement’s inconsistent total-duration comparison.

For a short LR attempt, this guide uses 10 questions in 14 minutes. That is a practice-set convention, not a rule to spend the same time on every question. Passage and question difficulty vary. Preserve the complete passage when selecting an RC set.

LSAC’s official practice page currently lists four free four-section PrepTests in LawHub and a one-year LawHub Advantage subscription at $124 (checked November 2, 2026; recheck before buying). Use official or otherwise licensed material; this guide does not reproduce LSAT questions.

Accommodations: the planner’s automatic section lengths are standard-time defaults. If you have approved timing or break accommodations, use your decision letter and the applicable LSAC instructions. The paper worksheet can hold your actual block lengths; do not treat the automatic planner’s verdict as validation of an accommodated schedule. A full practice test also needs a separate protected window for all sections, applicable breaks and setup; the short weekly loop is not a substitute for full-test practice.

Step 1: Inventory what cannot move

Start with a week you expect to repeat, not your best week. For each day, write the hours that are already spoken for, in four groups:

1. Sleep, at the amount you actually need to function at work.

2. Work, including commute, plus any regular overtime you cannot refuse.

3. Care and household: children, relatives, meals, errands, the chores that happen whether or not you study.

4. Other fixed commitments: classes, religious or community obligations, a standing family dinner.

Two rules keep the inventory honest. Count hours on the calendar day they actually occupy. Split overnight work, commuting and sleep at midnight: a 7 p.m.–7 a.m. shift occupies five hours on the first day and seven on the next, before adding commute time. Count each hour only once, including when care and work overlap. And never leave a group blank to mean zero: write 0. A blank means you have not checked, and the planner below shows it as missing rather than guessing.

What is left is uncommitted time, not study time. Meals, rest, exercise and everything unplanned come out of it too. The daily total is only a rough capacity check; the tool does not know the clock times of your commitments. Check each proposed window against your calendar yourself. For example, four uncommitted hours do not make a 2 p.m. window possible if you are working then. As a basic check: if your study windows on Tuesday add up to more than Tuesday’s uncommitted time, one of the two lists is wrong.

Step 2: Find your windows and mark their kind

A window is a specific start time and length on a normal week: “Monday 6:15 a.m., 45 minutes”, not “mornings”. For each window, mark two things.

• Alertness. Alert, or tired. Tired means you would hesitate to start a timed section right now: after a long shift, after the children are finally asleep, at the end of a twelve-hour day.

• Interruption risk. Low: nobody is likely to need you and you can silence your phone. Medium: an interruption is possible, but it is usually short or can wait until you finish the question in front of you. High: an interruption is likely, or would end the session on the spot.

Then give each window its kind. The kinds, like every size, threshold and rule below, are AccessPrep’s untested proposals. Apply these in order and take the first that fits:

The names say why the kind matters. A timed attempt that is interrupted stops measuring timed performance, so it needs a timer-safe window. Review can stop after any question and resume later, so a stop-safe window is enough. Length is checked separately, job by job, in the next step.

Step 3: Size the five jobs, full and shortened

Every week runs one loop of five jobs, in this order. Each has a full size and a shortened size, so that a smaller window shrinks the job instead of deleting it.

A full loop takes 160 minutes in the current format (170 with a new-format LR section, 177 with a new-format RC section). With every job shortened it takes 80 minutes, and it needs one timer-safe window of at least 20 minutes. Call that the minimum week. If your week cannot hold it, no amount of rearranging will close a loop; you need a different week, and Mei’s case below starts there.

These sizes are untested AccessPrep proposals, not measured requirements. If your windows are longer, the review guide’s two-hour and four-hour weeks show other ways to divide the same five jobs.

Step 4: Place the loop, hardest job first

Number your windows in time order, starting from Day 1. Day 1 does not have to be Monday: start your week on the day of your longest timer-safe window, because everything else in the loop follows the timed attempt (if your schedule rotates, see Troubleshooting). Then place the jobs with these steps, in order:

1. Try every job at full size.

2. Put the timed attempt, review, resolve and transfer jobs, in that order, each into the earliest window that comes at or after the previous job’s window, is a kind that can hold it, and still has enough minutes. A long window can hold more than one job.

3. Put the evaluation in the latest window of the week that comes at or after the transfer check and can hold it.

4. If any job will not place, shorten the timed attempt and start again from step 2. If it still fails, shorten the timed attempt and the review; then the first three jobs; and so on, one more job each time.

5. If the loop will not place even with all five jobs shortened, the week cannot hold a loop. Do not fill it with practice and hope; change the week (see Mei).

Three reasons sit behind the order. The timed attempt goes first because it is the only job that needs a timer-safe window, which is usually the scarcest kind. Jobs go as early as they fit so that the end of the week stays free, and free windows late in the week are what let you recover from an interruption. Shortening starts with the timed attempt because it is the largest job and the one that most often blocks the rest. These steps find a workable placement quickly; they do not search every possible arrangement. If you can see one that keeps more jobs at full size and still follows the window kinds, use it. The planner below checks any placement you type in.

Everything outside the loop is optional time: extra practice on the action you are testing, delayed checks of earlier actions, a second shortened timed set, or reading. Optional work is useful, and it is also the first thing an interruption takes.

Download the planner, blank and worked CSV worksheets, interruption rules and learning log (ZIP).

The planner on this page does the arithmetic for Steps 1 to 4: it shows each day’s uncommitted hours, gives each window its kind, places the loop with the steps above (or checks your own placement) and flags blank fields as missing. Nothing you type leaves the page. A blank copy of the same planner, for paper or a spreadsheet, is further down.

Step 5: Decide now what an interruption does

Interruptions are not a sign the plan failed; on a working schedule they are part of the plan. What fails is deciding what to do about them while you are tired and annoyed. Write these rules (an untested AccessPrep proposal) down before the week starts. When something goes wrong, apply them in order to each loop job the window held, earliest job first, and take the first rule that fits; if the window held only optional work, rule 1 is the whole answer.

Three definitions make the table usable. Re-placing means running Step 4 again on the windows that have not started yet, for the unfinished job and every loop job after it. A job cut short counts as not finished: keep what you did, labelled, and re-place the job. The one exception is a timed attempt you finish untimed later the same day, which counts as finished under worse conditions (rule 2).

One more rule keeps the week honest: you may add a window mid-week only if the time was freed from another commitment, such as a cancelled meeting. Time taken from sleep or from care is not a new window; it is a loan the next week repays.

For any one job, these five rules cover every case: writing the inputs as optional or loop work, finished or not, fits at planned size or not, and fits shortened or not gives 16 combinations, and each reaches exactly one rule. One combination, fitting at full size but not when shortened, cannot occur, because a shortened job needs fewer minutes and no stricter window.

Step 6: Close out the week in 20 minutes

The evaluation is the fifth job. It answers two questions, about the plan and about the action, and the stopping rule in the next section can override either answer.

The plan. Your loop closed this week if you reached the evaluation with the timed attempt, review, resolve and transfer jobs done, at any size, in order. A loop that started last week and closes this week counts for this week. For your first week, treat last week as closed. A rebuilt plan always gets a second week: its first failure falls to P2, and a second failure in a row meets stopping condition 1. “Your plan” means the week as placed in Step 4; a shrink caused by an interruption does not change it. Apply in order:

• P1 — Loop did not close, did not close last week either, and this was not the first week of a rebuilt plan — Rebuild: redo Steps 1 to 4 from the inventory, and change at least one window’s length, kind or existence. Repeating the same allocation does not address the constraint that caused the failures.

• P2 — Loop did not close — Keep the plan. Next week starts with the first unfinished job, and there is no new timed attempt until the loop closes.

• P3 — Loop closed this week and last, and your plan has a shortened job — Restore the last shortened job in loop order to full size if the loop still places without shortening anything else; otherwise keep the plan.

• P4 — Loop closed this week and last, and your plan has every job at full size — You may add one thing (one optional block, or a second shortened timed set), or keep the plan.

• P5 — Loop closed this week, not last week — Keep the plan.

The action. If the loop did not close, the action you were testing carries over unchanged. If it closed, count three things from the transfer check: N, the fresh questions where the action should apply (3 to 10); U, the questions where the action is written on your page and was written before you chose an answer; and C, the questions in U you answered correctly. A step you did in your head does not count toward U, because you cannot check it later. If the transfer check itself was labelled under rule 2, its counts are not comparable: decide A3, repeat, which is the review guide’s advice to “collect better evidence” when conditions changed. Otherwise apply in order:

• A1 — U is less than half of N — Rewrite the trigger. The action is not being called up when it should be, so rewrite the “when” part of it, then repeat.

• A2 — Two or more misses among the questions in U (U minus C is 2 or more) — Revise the action. You are using it and the error survives, so take those misses back through blind review and rewrite what the action does.

• A3 — U is less than N — Repeat the same action on new fresh questions.

• A4 — U equals N, with at most one miss — Advance. Book a delayed check (3 fresh questions, two weeks later, as optional work) and take your next action from the next timed attempt.

A delayed check passes if you used the action on all three questions with at most one miss; the action is then retired. If it fails, it becomes your current action again the next week, and any action it displaces waits its turn.

Counts this small are noisy, and the thresholds (half of N, two misses, a three-question delayed check) are untested proposals. On eight questions, one or two either way can be chance, which is why these rules only ever say repeat, revise or check again later, never “mastered”, and why every advanced action gets a delayed check. Usage is the count to trust most, because you can see it on the page.

When to stop adjusting the plan

Check these at every close-out. Condition 1 comes before the plan rules. Condition 2 applies as soon as the action rules give a rewrite or revise decision; weeks in which the loop did not close do not count toward it, either way.

1. The obstacle is outside the plan. Your plan has been rebuilt at least once, and since the latest rebuild the loop has failed to close two weeks in a row. Stop rearranging windows. Something outside the plan has to change: a protected window arranged with another person, a commitment dropped or moved, or a later test date. Then rebuild once more from the new inventory.

2. The obstacle is diagnosis, not schedule. The same error has produced three rewrite or revise decisions (A1 or A2) in a row. Stop changing the action on your own. Take your saved attempts through the blind-review and plateau guides, or get feedback from someone who can watch your reasoning; the tutoring vs. course guide helps you choose.

Neither condition is a verdict on your ability. Both say the same thing: more of the same planning will not fix this.

Case 1: Keisha, predictable hours (fictional)

Keisha is a fictional benefits analyst who works 8:30 to 5:00 on weekdays with a 45-minute train ride each way. Her hours rarely change. That is why her first plan failed in an instructive way: it ignored the week she actually had.

The first plan. She wrote “two hours every weeknight, 7:30 to 9:30, and four hours each weekend day”: 18 hours. Over two fictional weeks she logged 9.5 and then 7 hours. She took four timed LR sections, read the explanations before reviewing any of them, and never tested an action on fresh questions. No loop closed in either week, so by rule P1 the second close-out called for a rebuild. The inventory showed why. Her weeknight evenings follow ten hours away from home, so every one of them is low-load; Tuesday evening is choir; and Sunday afternoon is a standing family day she had planned over.

The rebuild. Her inventory, in hours (sleep + work and commute + care and household + other fixed):

• Day 1, Saturday — 8 + 0 + 4 + 0 = 12 — 12

• Day 2, Sunday — 8 + 0 + 3 + 4 = 15 — 9

• Days 3, 5, 6 and 7, Monday, Wednesday, Thursday, Friday — 7.5 + 9.5 + 2 + 0 = 19 — 5 each

• Day 4, Tuesday — 7.5 + 9.5 + 2 + 2 = 21 — 3

Work and commute is 9.5 hours, not 10, because her 30-minute lunch is not work time and holds two of her windows. That is 44 uncommitted hours, and she chose eight windows inside them. Her week starts on Saturday because Saturday morning is her longest timer-safe window. With the current format and LR timed attempts, Step 4 places every job at full size on the first try:

Her windows total 365 minutes: 160 for the loop and 205 optional. That is about a third of her first plan, and it is the first plan that produced a closed loop. The point is not that six hours is the right amount. It is that her week had about six usable hours, in a particular shape, and the 18-hour plan hid both facts.

An interruption, handled by the rules. In her second rebuilt week, an early meeting takes the Monday 6:15 window, where resolve and choose was planned. Rule 3 applies if the unfinished job and everything after it re-place at full size, and they do: resolve moves to the Monday lunch window (stop-safe, 30 minutes), the transfer check moves to Tuesday 6:15, displacing optional work, and the evaluation stays on Friday. The loop still closes at full size. Without the rule, the likely outcome was skipping the transfer check, the job that tells her whether anything she learned carries over.

What the format change does to her week. Keisha expects to test in August 2027 or later. When she switches to new-format practice in February 2027, her full LR timed attempt grows from 40 to 50 minutes. Her normal week still places at full size, because Saturday’s 100 minutes hold the timed attempt and review (90). The difference shows when Saturday is lost. In the current format, Step 4 moves everything at full size: the timed attempt fits Monday’s 45-minute window, with review on Tuesday, resolve on Thursday, the transfer check at Friday lunch and the evaluation on Friday evening. In the new format, no weekday window holds 50 minutes, so the same interruption becomes a shrink: a 10-question timed set on Monday morning, with every other job still at full size. Same week, same interruption, different rule. Running both formats through the planner shows this before it happens.

Growth, one change at a time. After two closed weeks at full size, rule P4 let her add one thing. She added a shortened RC timed set (one passage) on Thursday morning as optional work, and added nothing the following week. Her transfer-check results for six weeks are the worked log further down.

Case 2: Rafael, rotating shifts (fictional)

Rafael is a fictional maintenance technician at a distribution center. He works four 12-hour shifts a week, with a 30-minute drive each way, and his schedule rotates between day weeks (7 a.m. to 7 p.m.) and night weeks (7 p.m. to 7 a.m.). His days off move with the rotation. He is working on Reading Comprehension and testing in the current format.

Two failures. His first plan was fixed by clock time: “Wednesday and Thursday, 9 to 11 a.m.” It worked in day weeks and failed in every night week, when 9 a.m. is bedtime. To keep pace, he started taking timed RC sections straight after night shifts. His fictional record: three sections on days off at 19, 21 and 20 of 27, and two after night shifts at 15 and 17 of 27. Five sections cannot show how much the night shifts cost him; passages vary in difficulty, and a gap of about four questions across two sections could come from the passages alone. What the record does show is narrower and certain: the two groups were taken under different conditions, so averaging all five described neither. Under rule 2, a timed section done exhausted is kept and labelled; he still reviews it, but its result stays out of his timed comparisons.

The correction: one template per rotation. Rafael builds two weeks in the planner. His rotation changes on Saturday, so both templates start on Saturday; if they started on different days, one week’s loop would run into the next. In day weeks, his longest timer-safe window is on Wednesday, a day off: 120 minutes hold the timed attempt, review, and resolve and choose; Thursday morning holds the transfer check; the evaluation goes in a 20-minute slot on Friday evening; and the windows from Saturday to Monday, before the timed attempt, hold optional work. The night week is harder:

His alertness labels follow his sleep, not the clock. Four p.m. on a night-week Tuesday is a few hours after waking, so he is alert; the household is home, so the risk is medium. By Thursday the run of nights has worn him down, so that window is low-load whatever the time.

Overtime takes the best window. In his second night week, Rafael is called in for an extra shift on Saturday, the window that held the timed attempt and review. Rule 3 tries to re-place all five jobs at full size in the windows left: the timed attempt goes to Monday’s timer-safe window, review to Tuesday, resolve to Wednesday, and then the transfer check finds only low-load windows. Shortening the timed attempt alone still leaves the transfer check stranded. Shortening the timed attempt and the review works: both fit Monday’s 45 minutes (20 each), resolve goes to Tuesday, the transfer check to Wednesday and the evaluation to Friday. That is rule 4, shrink. Sunday’s stop-safe window now comes before the timed attempt, so it holds optional work: untimed RC practice with his current action.

Under the August 2027 format his RC timed attempt grows to 57 minutes. His Saturday window still holds it at full size; review moves to Sunday and the other jobs each move one window later, and the evaluation stays on Friday.

When the schedule is not the problem. Rafael’s action was “before answering, mark the sentence where the author’s own view first appears.” In week 1 he used it on 2 of 6 fresh questions (A1, rewrite the trigger). In week 2, after the shrink, he used it on 5 of 7 but missed 3 of those 5 (A2, revise the action). In week 3 he used it on 6 of 7 and missed 3 of the 6: a third rewrite or revise decision in a row. Stopping condition 2 applies. His loop closes every week, so more planning will not help; the error needs diagnosis. He takes three annotated passages through the blind-review method and considers getting feedback from someone who can watch him read. On the plan side, week 2’s shrink was a one-off: both of his templates place every job at full size, so from week 2 his close-outs fall under P4, and he adds nothing while the action is in question.

Case 3: Mei, caregiving interruptions (fictional)

Mei is a fictional payroll specialist who works from home, 8:30 to 5:00 with a 30-minute lunch. She is the main caregiver for her two-year-old on weekday evenings and weekends, because her partner works evenings and Saturdays. Her days are not short of minutes so much as short of minutes nobody can claim.

The first plan. Her windows were the child’s weekend naps (60 minutes at 1:00 p.m., alert but medium risk, since naps end early), lunch on Monday and Wednesday (25 minutes, medium risk because of work messages) and two evenings after bedtime (45 minutes, tired and medium risk). In Step 2 terms that is four stop-safe windows, two low-load windows and no timer-safe window at all. Step 4 cannot place even a shortened timed attempt, so her week could not hold the minimum loop. She did not know that yet. In week 1 she took a timed section during a nap, the child woke at minute 22, and she finished an hour later with the timer paused (rule 2: kept and labelled). Her review was cut short, and with no rule for what came next it never resumed. In week 2 a timed set stopped at question 9 and was never finished, and there was no timer-safe window to re-place it in (rule 5, carry). Two weeks without a closed loop meant rule P1: rebuild.

The rebuild changes a window, not just the schedule. P1 requires changing at least one window, and Mei’s missing piece was a timer-safe window. She arranged one with her partner: on Sundays he takes the morning, which gives her 60 minutes from 7:00, alert and low risk. Her week now starts on Sunday:

At full size the loop does not fit. The 40-minute timed attempt leaves only 20 minutes on Sunday morning, so review goes to the nap, and resolve, which needs 30 stop-safe minutes, cannot fit until Saturday; the transfer check then fills Saturday and the evaluation has nowhere to go. Shortening the timed attempt to a 10-question set lets review fit on Sunday morning, and the loop closes. Her loop totals 140 minutes. She also cuts every stop-safe job into question-sized pieces and writes where she stopped, so a waking child costs one question, not a session.

Week 3: a nap ends early. Twenty minutes into the Sunday nap, the child wakes. She has resolved one of three questions; resolve is not finished and the transfer check has not started. At full size, nothing before Saturday can hold a 30-minute stop-safe job, and Saturday cannot then hold the transfer check and the evaluation as well, so rule 3 fails. Shortening resolve works: one question and one log entry in Monday’s lunch window, then the transfer check and the evaluation in Saturday’s nap. Rule 4, shrink, and the loop closes. Her action was “on must-be-true questions, before crossing out an answer, write the fact it would need that the stimulus does not give.” She used it on 7 of 8 fresh questions with one miss among them: A3, repeat.

Week 4: the protected window disappears. Her partner is ill, and Sunday morning is gone. No other window is timer-safe, so even a shortened timed attempt cannot be re-placed: rule 5, carry. Her stop-safe windows go to untimed practice with the same action, which she does not count as a transfer check. At close-out, rule P2 keeps the plan, and week 5 starts with the timed attempt. Stopping condition 1 does not apply: since the rebuild, the loop has failed once, not twice in a row. In week 5 the loop closes again, and she uses the action on 8 of 8 fresh questions with one miss: A4, advance.

When P3 next applies, it will ask whether the timed attempt can return to full size without shortening anything else. Not yet: that is the full-size placement above, which runs out of week before the evaluation. Restoring it would take a longer Sunday arrangement, which is a conversation with her partner, not a planning trick. If the Sunday arrangement had failed in weeks 4 and 5 both, stopping condition 1 would have applied, and the next change would have been outside the plan: a second arrangement, other help with care, or a later test date.

Three weeks side by side

The pattern across the three: the first plan was written in hours, and the working plan was written in windows. Keisha and Rafael reused existing time more deliberately. Mei needed a real change in care coverage to create protected time. All three also needed rules for the weeks when a window vanished; rearranging a calendar alone could not supply missing support.

Try it yourself before copying a schedule

Use this new fictional week to test the method. You have Monday 6:30 a.m. for 45 alert, low-risk minutes; Wednesday lunchtime for 30 alert, medium-risk minutes; Thursday 6:30 a.m. for 45 alert, low-risk minutes; and Friday evening for 20 tired minutes. Assume the calendar and daily inventory checks pass. With current-format LR practice, can you place all five jobs? Write your answer before reading the explanation. Then remove Thursday and decide what changes.

Worked answer: a full timed attempt uses 40 of Monday’s minutes, leaving too little for review. Wednesday cannot hold the full 40-minute review, so it moves to Thursday, leaving no suitable transfer window. Shorten the timed attempt alone and the full review still cannot fit Monday. Shorten both to 20 minutes: Monday now holds attempt and review, Wednesday holds resolve (30), Thursday holds transfer (30), and Friday holds evaluation (20). The loop totals 120 minutes. The spare five minutes on Monday are not another study job.

After Thursday disappears: attempt and review still fit Monday. Shortening resolve to 15 minutes leaves 15 minutes in Wednesday’s stop-safe window for a shortened transfer check. Friday still holds the full evaluation. That is a 90-minute loop, with four shortened jobs and a full evaluation. If Wednesday also disappears, Friday’s tired window cannot hold a transfer check, even if its duration grows. The problem is the kind of window, not simply the weekly total.

Now test your own plan the same way: delete the most fragile window and name the first job that can no longer fit. Write the change you would make and what evidence would tell you it worked.

Copy this weekly planner

Use this on paper or in a spreadsheet if you cannot use the interactive planner. Keisha’s case above is a complete worked copy.

Part A: fixed hours, one row per day. Write 0 rather than leaving a group blank.

• {day 1} — {sleep} + {work} + {care} + {other} = {fixed} — {uncommitted}

Part B: windows, in time order.

• {day and start time}, {minutes} min — {alert or tired}, {low, medium or high} — {timer-safe, stop-safe or low-load} — {job, full or shortened, or optional}

Part C: the loop check. Before the week starts, confirm each line:

1. Every job sits in a window of a kind that can hold it, at the size you chose (Step 3 table).

2. The jobs in each window add up to no more than its minutes.

3. The jobs run in order: timed attempt, review, resolve, transfer check, evaluation.

4. No day’s windows add up to more than that day’s uncommitted time.

5. Your five interruption rules are written somewhere you will see them midweek.

Copy this learning and transfer log

One row per week. The log records whether the loop closed and whether the action showed up on fresh questions; your timed results belong in your error log, labelled by condition.

• {week} — {closed or not; jobs shortened; interruption rules used} — {action}: used on {U} of {N}, {C} correct where used. Delayed check: {action}, used on {u} of 3, {c} correct — Plan: {P1 to P5, or stop 1}. Action: {A1 to A4, or stop 2}

Worked log: Keisha, weeks 1 to 6 (fictional). Her two actions were “on flaw questions, write the conclusion in my own words before reading the answers” (K1) and “on most-strongly-supported questions, circle the strongest word in each finalist and find the sentence that supports it” (K2).

• 1 — Not closed (18-hour plan, no loop) — No action chosen — P2: keep the plan

• 2 — Not closed — No action chosen — P1: rebuild

• 3 — Closed, all full size — K1: used on 6 of 8, 5 correct where used — P5: keep. A3: repeat

• 4 — Closed, all full size; Monday morning lost, rule 3 (move) — K1: used on 8 of 8, 7 correct — P4: adds one optional RC passage set. A4: advance; delayed check booked for week 6

• 5 — Closed, all full size — K2: used on 3 of 8, 3 correct — P4: adds nothing. A1: rewrite the trigger

• 6 — Closed, all full size — Delayed check, K1: used on 3 of 3, 3 correct (retired). K2: used on 6 of 8, 4 correct — P4: adds nothing. A2: revise the action; a third A1 or A2 on this error would meet stopping condition 2

Read the K2 rows with the caution from Step 6: 4 correct out of 6 is two misses, and on six questions that could be bad luck. The rule still says revise, because two misses while using the action is enough reason to look at what the action does. It does not say she cannot learn the skill.

Put test day into the plan early

Plan the test appointment as a fixed commitment, including travel, check-in and any work or care cover you need. Under LSAC’s 2026–2027 scheduling process, registration order determines when your scheduling window opens. Available appointments remain subject to test-center capacity; early registration does not guarantee a particular appointment. Check the dates for your own administration and request leave once you can confirm the appointment.

In-center testing is the default for most candidates. LSAC’s remote-exception policy includes specified accommodations, military and distance circumstances. The domestic distance route applies beyond 180 miles or three driving hours from a center with available capacity. A preferred day or time alone is not a qualifying reason. Follow the request window for your route; do not assume remote testing is available because your work schedule is difficult.

LSAT Argumentative Writing is a separate remote-proctored task. LSAC’s specifications provide 15 minutes of prewriting analysis and 35 minutes of essay writing under standard conditions. Reserve more than those 50 minutes to allow for setup and required checks. An approved writing sample must be on file before the score can be released; check your account and current instructions rather than assuming you must repeat an already valid sample.

Troubleshooting

I have no timer-safe window at all. Then no loop can close, and Mei’s case is yours. A shortened timed attempt needs only 20 alert, low-risk minutes. Look for one arrangement: someone else covering a fixed hour, a quiet room at work before your shift, a library on a day off. Protect the sleep you need; do not treat sleep loss as a sustainable source of study time.

My schedule changes every week. If it rotates, build one template per rotation type, as Rafael did, and start every template on the same day of the week, usually the day the rotation changes, so one week’s loop ends before the next begins; windows before the timed attempt hold optional work. If there is no pattern at all, plan each week on its first day from the published schedule. The steps and rules are the same; only the windows change.

I keep doing optional work instead of the loop. Write the loop jobs into the plan by name and label everything else “optional”. When a window holds a loop job, the loop job goes first.

My evenings feel wasted. Tired windows are low-load, not useless: give them the low-load work from Step 2. What they should not hold is timed work, because changed conditions make the results harder to compare with rested practice; this planner cannot quantify the effect of fatigue.

Full practice tests do not fit anywhere. They are not loop jobs. Schedule one when you can protect a single window that holds every section and break of your format plus setup, as you would a medical appointment, and keep the weekly loop running around it.

I am testing in August 2027 or later. When you start new-format practice (LSAC says it arrives in LawHub in February 2027), switch the planner’s format setting and see which windows still hold a full section. Keisha’s case shows the usual effect: the normal week survives, and interruptions get more expensive.

I missed a whole week. That is one week without a closed loop. Rule P2 keeps the plan and starts the next week where you stopped. There is nothing to make up; if it happens twice in a row, rule P1 rebuilds the plan, or, if the plan was just rebuilt, stopping condition 1 applies.

Turn your next week into a plan you can test.

Free first step: before next week starts, open the planner on this page or copy the blank version above, enter your fixed hours and your windows, and place the five jobs, timed attempt first. Keisha’s week is a complete example to copy: eight windows, the timed attempt and review on Saturday morning, resolve on Monday morning, the transfer check at Monday lunch and the evaluation on Friday evening. Keep its shape and replace her windows with yours, then write the five interruption rules where you will see them. If you would rather start from a ready-made week, the free LSAT Study-Plan Check asks five questions, including whether you can give about 2, 4 or 6 hours next week, and gives you a seven-day plan on screen, with no email required. It plans from hours rather than windows, so check its blocks against your windows and their kinds before you rely on them.

Optional paid path: a tutor cannot add windows to your week, so paid help fits stopping condition 2, where the same error survives your revisions, not condition 1. Lovare Institut, which shares a founder with AccessPrep, offers paid one-on-one LSAT programs (current Lovare programs and prices). Its smallest published private options are 10-hour packages, and its terms of service state a six-month default expiry for unused hours and no package refund after the first completed session, subject to a different signed Order (checked November 2, 2026). Before buying, check that your planner has windows for the sessions and for the work between them. You can also discuss your plan with AccessPrep.

Continue your preparation

• Turn your error log into a weekly study plan

• Diagnose your timed-score gap with blind review

• Find the bottleneck before you add hours to a score plateau

• Decide whether tutoring, a course or self-study fits your problem

• Compare Georgetown Evening and GW part-time if you plan to keep working through law school

• Try the free LSAT Study-Plan Check