Every question format on the NextGen UBE, and what each one rewards.

Here is what each question type is and what it rewards. A separate page shows you what to actually do, step by step, inside each one: how to work each format.

How the exam runs

The NextGen UBE runs nine hours over a day and a half: three three-hour sessions on your own laptop, through NCBE’s secure platform. Scores come back on a 500 to 750 scale, and each jurisdiction sets its own passing line. Of the time, standalone multiple choice takes roughly forty percent, in two formats, and integrated question sets take just over a quarter. Performance tasks are the other measure: they carry roughly a third of the score, in two kinds.

Underneath are the seven skills the examiners name out loud: legal research, legal writing, issue spotting and analysis, investigation and evaluation, counseling, negotiation and dispute resolution, and client relationship management. The new question formats exist to test them. Each format asks you to do one piece of lawyering, and doing that piece is what earns the point.

One change cuts across all of them. The exam regularly hands you a statute, a rule, or an edited opinion and expects you to apply that text, even where it reaches further than the rule you learned. It can also test a subject you never memorized, by giving you all the law you need. NCBE tests family law and trusts and estates this way on every administration, and from July 2028 family law becomes a subject you also have to carry in your head. The old MBE reflex, recognizing the doctrine and reaching for the majority rule, can now cost you points.

That is as far as this page goes on structure. The scoring scale, the eight tested subjects, what NCBE’s published outline leaves out, the rollout jurisdiction by jurisdiction, and what the first administration showed are all on the exam page. What follows here is the formats themselves.

Select-one: the familiar format, doing unfamiliar work

One fact pattern, four options, exactly one of them credited. The early administrations deliberately look like MBE questions, and the variety grows from there. Select-one isn’t obsolete; NCBE just asks it to do more kinds of work than it used to.

What it still tests

The same three jobs, on a clock. Sort the facts with legal weight from the ones that are only there to sound dramatic. Name the doctrine before the options name it for you. Then apply the rule closely enough that the almost-right option doesn’t catch you.

Here is a well-built one, from a question of ours that is already public as a sample. A man means only to scare a rival, so he throws a rock inches past the rival’s head, and it breaks the nose of a bystander he never saw. Liable for battery? Yes. Transferred intent works twice here: the intent to assault carries over to a battery, and it carries from one person to another. Two of the wrong options say no, for reasons that feel completely fair, which is exactly why they’re there. What separates the right answer from the best wrong one is usually a single element, a single party, or a single word in the call.

Where it departs from the MBE playbook

Three things have changed. The call now asks what a lawyer would do, not what the law concludes: what to research next, what the court actually decided, which fact most hurts the defense. The law they give you controls as far as it reaches, and when a fact pattern quotes a statute, it quotes it because some small piece of that wording is the answer. And you’re given a role, so the real question is what that lawyer should do.

One habit is worth drilling until it’s automatic: read the call, the sentence that asks the question, before you read the facts. The same fact pattern can carry a doctrine question, a research question, or a counseling question, and only the call tells you which one you’re answering.

Select-two: six options, exactly two right, real partial credit

This is the one format with no old question bank behind it, and the one where candidates leave the most points on the table. Six lettered options, A through F, and exactly two of them are right. You pick exactly two, and the partial credit is real: getting one of them right earns points, and leaving it blank earns nothing.

The scoring, plainly

On the bench, our name for the practice app that comes with membership, each select-two is worth two points, scored 2 / 1 / 0. NCBE has said the partial credit is real. It hasn’t published the exact split, so the split here is our own convention: both right earns two points, one right earns one, and two wrong or a blank earns zero. Notice what that means. A half-right pair still pays on this format and a blank never does, so you always pick two.

What a blank actually costs

The math is small enough to check by hand. Six options, taken two at a time, make fifteen possible pairs. Only one of those fifteen is the fully correct pair. Eight of them are half right: one correct answer paired with one wrong one, and there are eight because each of the two correct answers can sit beside any of the four wrong ones. The remaining six pairs are two wrong answers together and score zero. So of the fifteen pairs, one earns full credit, eight earn half, and six earn nothing.

Read the first two rows of the table together. A blank is worth nothing here, the same as anywhere on the exam, but this is the one format where a coin-flip pair still returns about a third of the question’s credit on average. You are never so lost that a blank is the better play.

Here is the less obvious part. A blind guess on a four-option select-one gets you a quarter of that question’s credit. A blind guess on a six-option select-two gets a third, even though the field is bigger, because this format pays for being half right and the four-option one never does. So a select-two isn’t a harder select-one. It’s a format that pays out in proportion to what you actually know.

How the examiners build the field of six

Six options don’t feel generous in the room, because the four wrong ones aren’t filler. Each one is a specific way of being wrong, and once you can name the ways you start seeing them fast. Five kinds keep coming up in our own bank, and a question draws its four wrong options from among them.

• D1
  The right doctrine, one element short.A real rule from the right corner of the law that fails on an element the facts never supply. In a hearsay question about a startled shout, the dying-declaration exception is the classic: right area of law, wrong on every element.
• D2
  The rule for a different problem.A rule that governs something else entirely, offered because the subject heading matches. The best evidence rule governs writings, not shouts. Merchant status fills gaps in a sale of goods, and says nothing about whether a loan was fair.
• D3
  True, and beside the point.An accurate statement of law that has nothing to do with the call. The most dangerous kind, because confirming it is true feels exactly like finding the answer. The test is never “is this true,” it is “does this answer what the call asked.”
• D4
  Authority in the wrong order.Common in the research questions: a treatise ranked above a statute, an intermediate appellate decision above the supreme court, persuasive authority called binding. Half the work in the research format is throwing these out on sight.
• D5
  The answer that just feels fair.An option that follows ordinary fairness instead of the legal test. Buyer’s remorse is never the test for unconscionability, and an option saying the party “subjectively regrets the bargain” is there to catch anyone reasoning from sympathy.

Because two options earn credit, the call follows patterns worth knowing, and reading it is the fastest way in. Many select-twos are one test with two halves, where the two right answers are the two prongs of a single two-part rule. Unconscionability is the cleanest example. It needs both oppressive terms and an oppressive process, so a well-built question gives you one correct option for each prong and surrounds them with complaints about price and regret. Other questions want two independent reasons for one position, and there the second right answer should look nothing like the first.

What partial credit changes about how you work the item

Because the question pays for being half right, the goal isn’t to find the answer any more. It’s to turn whatever you already know into as many points as it will buy, and then move on.

Integrated question sets: one matter, told in pieces

This one is genuinely new. One client’s matter and one role, built on realistic documents: interview notes, a lease, a police report, a deposition excerpt, sometimes part of a statute or an edited opinion. A set mixes multiple choice, short answers, and medium-length written analysis, and new facts keep arriving as the matter moves. Some sets are pure doctrine. Others turn on drafting or editing a document, or on counseling a client whose stated goals and limits you’re expected to actually use.

What one of ours looks like, start to finish

Here is one of ours, already public as a sample, walked through from the start. A software founder pays a twelve thousand dollar deposit on a thirty thousand dollar venue contract, cancels seventy-five days out, and runs into a forfeiture clause. The statute you’re given makes keeping more than fifteen percent presumptively a penalty, unless the venue can prove actual damages. The set opens with a select-one on applying that statute, then a short answer asking for two specific facts that make keeping the deposit look like a penalty. Then the record grows: a demand letter arrives with another invoice, and a select-two asks which facts support the statute’s separate cause of action. Mediation produces a settlement proposal, and the next question asks what is wrong with it for this client, given what she has told you. Then a voicemail surfaces from the venue’s manager, the set turns from contracts into evidence, and it closes with a paragraph of written analysis worth four scoring points.

Six questions in all, which the exam calls components, on one continuous matter. That is the whole design: the facts never restart, they pile up.

The integrated setFacts unfold, nothing closes

“Given the facts now known”

“Given the facts now known,” or words like it, is the instruction you will find in every set. You can answer every question on the facts you have at that moment, and NCBE puts it in the exam instructions: no information provided later in a set is needed to answer an earlier component. That cuts against the instinct the format creates. An unfolding matter feels like it is holding something back, so candidates stall, or answer thinly and promise themselves a second pass. Neither is necessary. Read the question again instead of reading ahead. If you do leave something blank, the navigation pane is right there, because the graders score the answer you leave.

What the structure rewards

• Answering the record in front of you, not the case you think is coming. The temptation is to hedge toward the twist, and that spends points now on a fact you don’t have yet.
• Switching subjects without warning. A matter that opens in contracts can turn on an evidence question, and nobody tells you it happened. Studying one subject at a time never trains you for that.
• Counseling this client, not a generic one. These components are only answerable if you treat what the client told you as facts that matter. Say she has told you cash is tight, that she will never use this venue again, and that she publishes candid vendor reviews. Then a settlement demanding a future booking and mutual non-disparagement has specific problems you can name, and a generally sensible objection earns nothing. Weighing a settlement against what a client actually told you she wants has never been on a bar exam before. It rewards reading the interview notes as closely as the statute.
• The same partial credit, again. NCBE scores a select-two inside a set the same way as a standalone one: two points for both right, one for one, zero for none. So the same rule holds, never leave one blank.

Performance tasks: two kinds, one clock

Roughly a third of the score, in two formats that ask for very different things.

The standard task

It grew out of the MPT: a file, a library of authorities they give you, and one long writing assignment, on a sixty-minute clock. No outside law is wanted. What surprises people is that the graders score the instructions themselves. The memo assigning the work tells you what to write, what form to write it in, and what to leave alone, and every one of those is worth points rather than just politeness.

The legal-research task

This one is new, and NCBE puts it together differently: four multiple-choice questions, a short answer, and a medium-length writing assignment, worked over a small library. It tests your research judgment directly. Which authority is primary and which is secondary, which one binds the court and which only persuades it. It also asks what a case actually held, as opposed to what it said in passing, and whether the library you were handed even answers the client’s question.

That last one surprises people. One of our research tasks gives you four sources. The credited answer to its final multiple-choice question is that the library doesn’t answer what the client is asking, because the controlling case expressly leaves his exact situation open. Being able to say “my authorities don’t reach this” earns points here, and it’s the last thing any test has ever asked most candidates to say.

The two-pane workspace

You work both kinds in a document workspace rather than on a question screen: the file on one side, the library on the other, your answer in its own pane, and a clock running. Reading turns into looking things up. You’re moving between documents rather than holding one fact pattern in your head, and every minute spent re-finding a paragraph you already read is a minute you’re not writing. Map the library on your first pass, noting which source says what and where the sentence that matters sits, and you buy back more time than reading faster ever will.

The performance taskTwo panes, one clock

What the graders reward

Scoring is point by point, against a list. The list belongs to the grader while you write; on the bench you see it with your grade, when the grader shows its work. One of our standard tasks runs twenty points: two for following the instructions and the format, then six, eight and three across the three issues, and one for tying it together. Notice what that first number means. On that task you settle a tenth of the credit before anyone grades a word of your analysis, and across our standard tasks it runs one or two points in twenty. It goes purely on whether you wrote the section the task asked for, organized it the way the task said, and left out what the task said to skip.

Constructed response: written answers, graded by element

Written work comes in two sizes. The short one is the roughly one-sentence answer field inside integrated sets and research tasks. The long one is the analysis that closes a set or a task, anywhere from a paragraph to a memo. The graders score both against a list of specific points rather than an impression, and that gap is where most candidates lose written points.

The one-sentence field

Deceptively hard. The instruction is usually about one sentence, one answer per field, scored against a list of answers the graders credit. Each field is one slot, so packing two facts into one doesn’t earn two points and risks earning one. And specific is the operative word. “The venue behaved unreasonably” is not an answer. “The retained deposit equals forty percent of the contract price, far above the statute’s fifteen percent presumption threshold” is, because a grader can match it to something on the list.

What that grading actually looks like

A real grade from the grader members use shows it better than a description could. The setup is the deposit matter from above: a client canceled her event seventy-five days out, and the venue kept her twelve thousand dollar deposit. The task was to identify two specific facts supporting that keeping the deposit is presumptively an unenforceable penalty. The first answer cited the forty percent figure against the statutory fifteen percent threshold and earned the point. The grader quoted the writer’s own words back to show why. The second argued that the venue made no effort to rebook the date, so it couldn’t claim it had tried to limit its losses. It reads well. It scored nothing. The answers the graders credit are the seventy-five day notice, the clause that forfeits every deposit for any cancellation, or the forty percent figure. This one described what the venue did after the cancellation, not the facts of the cancellation itself. One of two.

That is what grading against a rubric means. The grader isn’t asking whether your answer sounds right, it’s checking whether your answer hits the specific points on the list. It’s also why a grader that flatters you does you no favors: a kind score tells you nothing about which sentence failed.

Writing to the rubric

Length isn’t what scores. A tight paragraph hitting four of the graded points beats a page hitting two. State the governing standard in the words the provided text uses. Quote the record, because those points come from specific facts. Leave out exactly what the task told you to leave out. And reach a conclusion, even a hedged one, because a supported prediction usually earns a point of its own, and an answer that just stops earns nothing.

Which format matters most

We’re not going to rank them. The share figures are the honest guide. Multiple choice takes the most time, performance tasks carry the most score, and the integrated sets are where the exam’s genuinely new skills live. Whichever exam date you take, all three are on it, and you get good at the thing you actually practiced. That’s the whole argument for working each format as itself, instead of translating from an exam that no longer exists. The companion page, how to work each format, takes each question type in turn and walks through what to do, step by step. Shorter questions about any of this, answered one at a time, are in the FAQ.

Where this page comes from

Every claim above about the exam’s structure comes from NCBE’s published materials and official statements, and where a figure describes our own materials rather than the exam, it says so. The scoring table is arithmetic on the 2 / 1 / 0 scale, and we present it as arithmetic rather than as a prediction of scaled scores. The samples come from our own original questions, set in the made-up state of Meridian. No real exam question appears here, and none ever will.

What you can check

• Independent of NCBE
• Professor and attorney authored and overseen
• Seventeen years with bar students
• Every claim labeled by its source
• No pass guarantees, no predicted questions
• Seven checks before a question reaches you

The .ai is for convenience. The vetting is not.

We built the bench for exactly what this page describes. It has 1,194 original questions across both multiple-choice formats, 750 select-one and 444 select-two, scored with real partial credit. It has all eighteen integrated sets, with the record growing as new facts arrive and nothing closing behind you, the way the exam does it. And it has both performance-task formats on the real clock, with written answers graded against a rubric you can see. Three real questions are free to try, the grader will grade an answer you write, and we publish the editorial standard so you can hold us to it. If that is the practice you want, full access is $399, one payment.