An LSAT question type error log is the single most effective study tool for turning practice mistakes into score gains, because it forces you to tag every miss by question type and find the patterns hiding underneath. Most students log mistakes; very few log them in a way that produces a visible weakness map. This guide gives you the exact column structure, the 17 Logical Reasoning question types to tag, an error-category framework, and a weekly review schedule that converts log data into a concrete study plan.
An LSAT error log is the workhorse of test prep — but most students build one wrong. They list the questions they missed, write a short note about why, and move on. The log fills up. Their score does not move. The problem is rarely effort; it is structure. Without a question-type tag on every entry, the log is unsortable, and unsortable data hides exactly the patterns it should reveal.
When your log is a chronological list of mistakes, the only pattern you can see is "I am still missing some questions." That tells you nothing actionable. Compare that to a log filtered by question type: suddenly you can see that 14 of your last 20 misses were Necessary Assumption, that your Flaw accuracy dropped only after you started timed practice, or that your Reading Comprehension Inference misses cluster around science passages.
Cambridge Coaching, Odyssey Test Prep, Magoosh, and Blueprint Prep all emphasize the same point in their methodologies: an error log only works when it is searchable. Tagging by question type is what turns the log from a journal of failures into a diagnostic instrument that points to your next study action.
LSAT errors cluster by question type far more than by date or by test number. Two students at the same overall score can have completely different weakness profiles — one missing mostly Parallel Reasoning, the other missing mostly Strengthen — and they need completely different study plans. Without the question-type column, you cannot distinguish a student who needs to drill Parallel Reasoning from one who needs to drill Strengthen, even if their raw score is identical.
Before you can tag accurately, you need a clean taxonomy. There is no point logging "LR mistake" — that's like noting "math problem" on a math error log. The whole point of the system is granular categories. Here is the full list every entry should map to.
The Logical Reasoning section uses 17 distinct question types according to 7Sage's classification, organized by what the question stem asks you to do. Flaw is the most common type — averaging 3.9 questions per LR section — which means a single LSAT contains nearly eight Flaw questions across its two scored LR sections. Most Strongly Supported (2.45 per section) and Necessary Assumption (2.25 per section) come next.
When you account for the top seven types — Flaw, Most Strongly Supported, Necessary Assumption, Weaken, Resolve, Strengthen, and Sufficient Assumption — you cover roughly three-quarters of the LR section. Mastering these seven categories has the largest score impact, so your log should track them most aggressively.
| Question Type | Avg. Per LR Section | Total Per Exam (2 LR Sections) | Priority |
|---|---|---|---|
| Flaw / Descriptive Weakening | 3.9 | ~7.8 | High |
| Most Strongly Supported | 2.45 | ~4.9 | High |
| Necessary Assumption | 2.25 | ~4.5 | High |
| Weaken | 2.25 | ~4.5 | High |
| Resolve / Reconcile / Explain | 1.95 | ~3.9 | High |
| Strengthen | 1.95 | ~3.9 | High |
| Sufficient Assumption | 1.35 | ~2.7 | Medium |
| PSA — Find the Rule | 1.3 | ~2.6 | Medium |
| Main Conclusion | 1.25 | ~2.5 | Medium |
| Argument Part | 1.1 | ~2.2 | Medium |
| Parallel Method of Reasoning | 1.0 | ~2.0 | Medium |
| Parallel Flawed Method | 1.0 | ~2.0 | Medium |
| Point at Issue — Disagree | 1.0 | ~2.0 | Medium |
| Must Be True | 0.95 | ~1.9 | Medium |
| Method of Reasoning | 0.75 | ~1.5 | Low |
| PSA — Find/Complete Application | 0.35 | ~0.7 | Low |
| Point at Issue — Agree / Evaluate / Must Be False | ~0.1 each | <1 combined | Low |
Reading Comprehension doesn't have a tidy 17-type taxonomy, but most students get good mileage out of six core tags. Main Idea questions ask for the passage's central argument. Inference questions reward staying close to the text and rejecting one-step-too-far answers — the same skill as LR's Most Strongly Supported. Function questions ask why an author included a sentence or paragraph. Detail questions test direct recall. Tone (or Author Attitude) questions assess the author's stance. Analogous Reasoning questions test parallel-structure recognition.
Tagging your RC misses by these six categories is what reveals whether your weakness is global comprehension (Main Idea, Function) or local precision (Detail, Inference). Each demands a different study approach, just like LR.
The Argumentative Writing sample is part of the LSAT but is not scored as part of the 120 to 180 scaled score. It is sent to law schools alongside your score, and admissions committees do read it — but since it does not generate "right" or "wrong" answers, it does not belong in your error log. Track your writing process separately if you want feedback on it, but keep the error log focused on the scored sections.
Every competitor article tells you to use a spreadsheet, but almost none specify the columns. That's the biggest content gap on this topic, and it's why most students abandon their log within two weeks: their format is either too sparse to find patterns or too elaborate to maintain. The structure below balances those two failure modes — ten columns that are powerful but realistic.
Start with these ten columns. They take less than five minutes to set up in Google Sheets, Excel, or Notion, and they give you everything you need to filter, sort, and spot patterns. Skip any of these columns and you lose a critical analytical lever.
| Column | What to Record | Example | Why It Matters |
|---|---|---|---|
| Date | When you reviewed the question | May 10, 2026 | Lets you see if certain mistakes recur on stressed or rushed days |
| PT Number | Preptest source | PT 89, S2 | Required to revisit the original question |
| Section | LR1, LR2, RC, or Experimental | LR2 | Reveals whether fatigue affects later sections |
| Q Number | Question number on that section | Q14 | Lets you re-attempt the question later |
| Question Type | Specific LSAT type (Flaw, Necessary Assumption, etc.) | Flaw | The most important column — makes patterns filterable |
| Difficulty | Star rating or 1–5 scale | 4 stars | Separates Q-type weaknesses from difficulty curves |
| Your Answer | The choice you picked | C | Reveals tempting wrong-answer patterns |
| Correct Answer | The right choice | D | Pair with Your Answer to spot trap-answer types |
| Error Category | Content / Misread / Process / Time / Careless | Misread | Tells you what to fix, not just what you missed |
| Takeaway | One actionable sentence | When the stimulus says 'most,' don't read it as 'all' | The single sentence that drives future improvement |
Worked Example: Logging a Real Miss
Setup: You take PrepTest 89, miss question 14 in LR section 2, and need to enter it in your log. The stimulus argued that since most successful CEOs jog in the morning, jogging caused their success. You picked C; the correct answer was D.
Once you've logged 20 to 30 questions consistently, you can add columns that give you finer-grained signal. Time Spent (in seconds) tells you whether your misses are pacing-driven or comprehension-driven. Stimulus Topic (science, law, ethics, business) reveals whether you're worse on certain content domains. Re-Do Date implements spaced repetition — set it two weeks out, then re-attempt the question without notes. Confidence Rating (1 to 5 before checking the answer) flags overconfidence patterns.
Resist adding all four at once. Add one new column per week; if you stop maintaining it within a session or two, drop it. The log only helps if you keep using it.
Question type tells you what you missed. The error category tells you why. These are two separate dimensions, and your log needs both. Two students can miss the same Flaw question for completely different reasons, and they need completely different fixes.
Synthesized from competitor methodologies and standard test-prep theory, every LSAT mistake falls into one of these five buckets. Pick one per entry; if multiple seem to apply, pick the upstream cause (the misread that triggered the process error, not the process error itself).
| Error Category | What Happened | Example | How to Fix |
|---|---|---|---|
| Content Gap | You didn't know the rule, vocabulary, or logical pattern | Confused 'sufficient' with 'necessary' assumption | Re-study the question type from a curriculum (7Sage, LSAT Demon, Powerscore), then drill 15+ untimed questions |
| Misread | You missed a scope word, qualifier, or contrast indicator | Read 'most' as 'all' in the stimulus | Slow down on the first pass; annotate key qualifier words (some, most, all, only, except) |
| Process | You chose a tempting wrong answer without eliminating distractors | Picked an out-of-scope answer because it 'sounded right' | Practice four-step elimination: prephrase, scan for matches, eliminate three, defend the last one |
| Time Pressure | You rushed because the clock pressured you | Skipped a stimulus re-read on the last 5 questions | Train pacing with sectional drills; build a 'guess and flag' policy for the last 60 seconds |
| Careless | You knew it but slipped — wrong bubble, wrong answer letter | Marked C instead of D on the answer grid | Add a final-step check before bubbling; use ScoreGap or similar to spot careless-cluster patterns |
The error category is what separates surface-level logging from diagnostic logging. To make this concrete, here's a worked example.
Worked Example: Same Miss, Different Fix
Setup: Two students both miss the same Necessary Assumption question. Student A picked the answer that strengthened the argument; Student B picked the answer that addressed an irrelevant scope.
A common failure mode is to build a beautiful log and never review it — or worse, to review it passively by re-reading takeaways. The 7Sage methodology and Cambridge Coaching's recommendations both point to a different approach: re-solve, don't re-read. Active review is the conversion mechanism between log data and score gains, and tracking LSAT errors by question type only pays off when this loop is closed.
The single most effective review technique is this: pull up a question you previously missed, ignore your notes, and solve it from scratch. Then compare your process to the takeaway you wrote earlier. Did you apply the lesson, or did you fall into the same trap? If the process is unchanged, the question still needs review. If the process is now correct and confident, the lesson has stuck — move to the next entry.
This protocol forces you to actively reconstruct the reasoning rather than passively recognizing the answer. Without it, you are training pattern matching on the explanation rather than on the question, which is exactly the muscle you can't use on test day.
Cadence matters. A log reviewed once a month is barely better than no log; a log reviewed daily becomes overwhelming and stops getting maintained. The sweet spot is twice a week plus pre-test reviews, with a deeper analysis every two weeks.
| Review Type | Frequency | Time | What You Do |
|---|---|---|---|
| Quick Read-Through | Twice per week | 15–20 min | Read every takeaway from the past week; reinforce lessons without re-solving |
| Active Re-Solve Session | Weekly | 30–45 min | Redo 5–10 missed questions from at least one week ago without checking notes; compare your process |
| Pattern Analysis | Every 2 weeks | 20–30 min | Filter log by question type and error category; identify any cluster above 30 percent |
| Pre-Test Read-Through | Before every full-length practice test | 20–30 min | Read all takeaways to activate learned patterns before timing pressure starts |
| Monthly Deep Audit | Monthly | 45–60 min | Compare error rates by question type to the prior month; adjust study allocation |
The day before any full-length practice test, sit down with your log and read every takeaway. Don't re-solve. Don't add new entries. Just read. The goal is to surface your most recent lessons to working memory so they trigger automatically during the timed test. Most students skip this step and then wonder why the same patterns keep tripping them up.
The questions below test the four habits that make an error log productive: tagging the right question type, tagging the right error category, choosing the right review cadence, and responding correctly to pattern data.
The end goal of an LSAT mistake log is not a tidy spreadsheet — it's a study plan that changes based on what the log shows. This is where most students get stuck: they build the log, they review it, but they don't translate the data into a different week of practice. The next three subsections give you concrete decision rules.
When you run a pattern analysis every two weeks, count the misses in your most-missed question type. If a single type accounts for more than 30 percent of all your misses across the last 14 days, that type is the bottleneck — and untimed drilling on that type is the highest-leverage thing you can do. Take a full week off timed practice, run 15 to 25 untimed questions of that type per session, and only re-introduce timing once your untimed accuracy is above 85 percent.
The reason this works: a 30-percent cluster means the question type is so confused that timed practice is making you worse, not better — you are re-rehearsing the wrong process under stress. Untimed drilling lets you rebuild the underlying skill without contaminating it with timing pressure.
The second decision the log forces is whether you need new content or better process. If your error categories cluster around "content gap," you need to re-study fundamentals — pull out the 7Sage course, Powerscore Bible, or your prep curriculum and re-learn the rules. But if your error categories cluster around "misread," "process," or "careless," more content study will not help. You need process drills: annotation routines, elimination protocols, and pre-bubble checks.
A useful diagnostic: if you can explain why the correct answer is right and why each wrong answer is wrong, but you still missed the question under timing, you have a process problem, not a content problem. Don't waste a week studying concepts you already understand.
Within four weeks of test day, the ratio flips. Early in prep, spend 60 to 70 percent of your study time on log review and the rest on new questions, because the log is where compounding happens. Within four weeks of the LSAT, switch to 70 percent timed practice and 30 percent log review — you need to stress-test the patterns you've internalized rather than continuing to add new ones. Add the full pre-test read-through before every timed full-length PT to keep your most recent lessons active.
Format is the silent killer of error logs. Students pick the wrong tool, hit friction, and stop using the log within two weeks. The right choice depends on how you study and what kind of friction you tolerate.
Spreadsheets — Google Sheets or Excel — are the default for one reason: filtering. The single most valuable thing you can do with a log is sort by question type and immediately see the cluster of misses, and that requires structured columns and a filter row. No other format does this as quickly. Google Sheets in particular wins for cloud access (you can log a question on your phone during a commute) and zero setup cost.
If you want to attach screenshots of the original question, link to explanation videos, or build a tagging system more elaborate than question type plus error category, Notion or Airtable becomes worth the higher setup cost. Airtable in particular handles both database-style filtering and rich media. The downside is the initial setup time — budget 30 to 45 minutes to build the template, longer if you've never used the tool. Don't switch to Notion if you're already two weeks into a Google Sheet that works.
Paper has one real advantage: it forces slow, deep engagement. Some students who study on paper genuinely retain more from writing out a takeaway by hand. But paper loses every pattern-detection advantage — you can't filter, sort, or count entries by question type without re-tabulating manually. Use paper only if your handwriting is legible enough to reread, your discipline is excellent, and you're willing to maintain a separate analysis sheet for pattern detection.
If you take most of your practice tests on LawHub, dedicated tools like ScoreGap can auto-import results and tag questions by type automatically, which removes the largest friction point in maintaining the log. The trade-off is dependency on the platform — if you ever switch back to PDF preptests, you lose the integration. Worth considering if you take more than 20 LawHub tests during prep.
Identify the question type you're missing most often, and get a focused next-step study action.
At minimum, log the date, preptest number, section, question number, question type, your answer, the correct answer, your error category (content, misread, time, process), and a one-sentence takeaway. Add optional columns for difficulty, time spent, and re-do date as you grow comfortable. The two non-negotiable columns are question type and error category, because they make patterns filterable.
Log every wrong answer, every correct answer that came from a lucky guess, and every correct answer that took more than 90 seconds in Logical Reasoning. Easy questions you missed are the most diagnostic, since they reveal process errors rather than content gaps. Limiting your log to hard questions hides the most fixable mistakes and slows your score improvement.
Review your log twice a week: one quick read-through to refresh takeaways, and one active re-solve session where you redo missed questions without checking your notes. Run a full pattern analysis every two weeks to look for question-type clusters, and read all takeaways the day before any timed practice test or the real LSAT.
Logical Reasoning has 17 question types: Flaw, Necessary Assumption, Sufficient Assumption, Strengthen, Weaken, Must Be True, Most Strongly Supported, Resolve the Paradox, Main Conclusion, Argument Part, Method of Reasoning, Parallel Reasoning, Parallel Flaw, Principle (PSA), Point at Issue, Evaluate, and Must Be False. Reading Comprehension uses about six categories: Main Idea, Inference, Function, Detail, Tone or Author Attitude, and Analogous Reasoning.
Spreadsheets win for most students because they sort and filter by question type, which is the entire point of the log. Use Google Sheets if you want cloud access, Notion or Airtable if you want to attach screenshots, and dedicated tools like ScoreGap if you take a lot of LawHub tests and want auto-tagging. Avoid paper unless your handwriting and discipline are exceptional.
You never stop reviewing the log, but the ratio shifts. Early in prep, spend 60 to 70 percent of time on log review and 30 to 40 percent on new questions. Within four weeks of test day, flip it: 70 percent timed practice, 30 percent log review, and read the full takeaway list before every full-length test to lock in patterns.