Two-Part Analysis is one of the five question families inside the GMAT Focus Data Insights section, and it is the family that most often punishes candidates who come in with strong Verbal instincts but no plan for handling a stem that asks for two answers at once. The screen looks like a cross between a quant problem and a logic puzzle: a small scenario, a table with two columns of candidate options, and a single prompt that has to be satisfied by picking one entry from column A and one entry from column B. The trick is that the scoring is binary. You get the question right only if both selections are correct. A half-right selection counts as a wrong answer, the same as a guess. That single design choice reshapes the entire approach to pacing, to reading the stem, and to the kind of partial work you should be willing to commit to on scratch paper.
This guide is written for candidates preparing for the GMAT or the GMAT Focus Edition who want a sharper, more disciplined method for the Two-Part Analysis item family. The goal is to teach a method, not a slogan. By the end of the article you should be able to identify the family on sight, read the table in the right order, separate the two sub-questions inside the stem, and execute a four-move protocol that protects you from the most common traps, including the choice that satisfies one column but breaks the other, the distractor built from a true individual fact that is not jointly possible, and the time-sink that masquerades as careful reading. The framing throughout is exam-specific: pacing budgets, scoring consequences, and the difference between an item family that rewards speed and one that rewards structured thinking.
What a Two-Part Analysis item actually looks like
A Two-Part Analysis question on the GMAT Focus presents a short scenario, typically one paragraph, followed by a table with two columns of candidate options. Each column holds between four and six answer choices. The stem is a single sentence that contains two linked but separable requirements. The candidate has to choose one option from column A and one option from column B, and both choices are scored together as a single correct or incorrect unit. There is no partial credit. The screen finishes with two oval-style selectors, one beneath each column, and a confirm button.
The first thing to register, before any content work, is the structural signature of the family. The stem almost always contains a coordinating conjunction that joins two requirements, frequently 'and', sometimes 'or' when the second condition is exclusive, and very occasionally 'but' when the two conditions must be evaluated against a common constraint. The conjunction is not decorative. It is the seam in the prompt, and the entire solution depends on finding it. Many candidates read the stem as a single integrated requirement and end up with an answer pair that satisfies the first clause and quietly violates the second. The defensive move is mechanical: locate the conjunction, write a slash on your scratch paper, and treat the two halves as if they were two separate sub-questions.
The second structural feature is the table itself. The two columns usually mirror the structure of the two sub-questions. If the stem asks 'which option in column A and which option in column B together satisfy the constraint', column A is reserved for the first sub-question and column B for the second. A small number of items, particularly the harder ones, scramble this and put the second sub-question in column A. The only safe habit is to read the column headers before reading the answer choices. Reading the column headers takes five seconds and saves the two minutes you would otherwise spend picking a structurally correct pair that the test treats as a misread.
A third feature, often missed, is that Two-Part Analysis items on the GMAT Focus almost always present a single shared pool of background data. The first paragraph describes a situation with several moving parts: a project with a budget, a schedule with two deadlines, a hiring decision with a set of constraints. The table is then populated with candidates that interact with that shared pool. A common mistake is to read only the column you are about to pick from and ignore the other column entirely. The other column often contains a distractor that looks plausible in isolation but is jointly impossible with the correct answer in the first column. The fix is to read every entry in both columns at least once, even if you think you already know the answer to the first sub-question.
How the binary scoring reshapes your strategy
The binary nature of Two-Part Analysis scoring is the single most important fact about the family, and it changes how you should think about time, risk, and commitment. On a multiple-choice question, a candidate who can narrow the field to two choices and guess has a 50 percent chance of earning the point. On a Two-Part Analysis item, the same candidate who narrows each column to two options and guesses independently has only a 25 percent chance, because the two guesses multiply. A pair of 80 percent confident eliminations per column, by contrast, gives a 64 percent chance, and that is the threshold at which committing an answer starts to beat walking away. The takeaway is practical: do not commit until both columns are simultaneously strong. A 90 percent read on column A and a coin flip on column B is functionally a coin flip on the whole item.
This multiplicative risk is why the most reliable Two-Part Analysis method is the four-move protocol, not the speed-read approach that works on other item families. The four moves are: identify the conjunction, read the column headers, do a full pass on column A only, and only then cross to column B. The temptation is to scan both columns simultaneously and start eliminating on the fly. That works on easy items and on items where the two sub-questions are independent, but it fails on the medium and hard items, where the test deliberately couples the columns so that a choice that looks correct in isolation is jointly impossible with a different choice in the other column. The full pass on column A first gives you a stable reading of the first sub-question, and only after that reading is locked in do you go looking for its partner in column B.
The binary scoring also changes the value of walking away. On the GMAT Focus Data Insights section, candidates have 45 minutes for 20 questions, an average of 2 minutes 15 seconds per question. Two-Part Analysis items, as a family, run slightly longer on average than Table Analysis and slightly shorter than Multi-Source Reasoning. A reasonable budget is around 2 minutes 30 seconds for an easy item and up to 3 minutes 30 seconds for a hard one, with a hard cap of 4 minutes. Past 4 minutes, the expected value of further work is negative, because the time cost shows up on the next two items in the section. The walk-away decision is part of the method, not a failure of it. If both columns are still wide open after two minutes of structured work, the correct call is to make the most defensible pair you can, commit, and move on.
The four-move protocol, step by step
Move one is identifying the conjunction. Re-read the stem with the explicit question 'what two things is this prompt actually asking for?' and write the two halves on your scratch paper separated by a slash. A typical stem might read: 'In the schedule above, which Project and which Manager are assigned to it such that the project is completed on time and the manager has at least one prior project of the same type?' The two halves are 'which project is completed on time' and 'which manager has at least one prior project of the same type'. The slash on your paper becomes the visual reminder that the question has two parts. For most candidates reading this article, the single highest-leverage habit change is this slash. It costs almost no time and it forces the rest of the protocol to be coherent.
Move two is reading the column headers before reading the answer choices. The headers usually name the entities in each column, but the test is happy to put the first sub-question in column B and the second in column A, especially on harder items. The headers also tell you what kind of entity you are picking: a project, a manager, a price, a probability, a year. If the headers read 'Project' in column A and 'Manager' in column B, and your first sub-question is about a manager, you already know you are looking in column B for the manager. This five-second check prevents the most embarrassing kind of Two-Part Analysis error: a structurally correct pair entered into the wrong selectors.
Pass one: column A only, with the conjunction in mind
Move three is a full pass on column A. Read every entry in the column, not just the ones that look promising. For each entry, ask whether it can satisfy the first sub-question on its own. Strike out the entries that cannot, and circle or mark the entries that can. If only one entry survives, that entry is locked in, and your second pass is forced into a single decision. If two or three entries survive, mark all of them and move to column B without committing. The discipline here is to not allow the contents of column B to influence the read on column A. Cross-contamination is the most common source of two-column error on this family, and the cure is to physically cover column B with your scratch paper during pass one. For most candidates this is a small but decisive change, because it stops the brain from pattern-matching across the columns before either read is complete.
Pass two: column B against the survivors in column A
Move four is the cross-column pass. With a shortlist in column A, go to column B and read every entry. For each surviving entry in column A, ask which entry in column B can satisfy the second sub-question jointly with it. The first surviving pair you find is not necessarily the answer. The correct answer is the pair that is jointly consistent. If column A has two survivors and column B has three, you have up to six pairs to evaluate. The evaluation is usually quick, because most joint combinations fail on the shared background data. The pair that passes both sub-questions simultaneously is the answer. Commit, confirm, and move on.
Common pitfalls and how to avoid them
The first pitfall is the true-individually, false-jointly trap. A choice in column A is factually correct on its own. A choice in column B is factually correct on its own. Together they violate the shared background. The test exploits this trap on roughly one in three Two-Part Analysis items, and the only defence is the cross-column pass described above. If you commit on the strength of two independent reads, you will pick up a steady drip of these. The defensive habit is to make the cross-column pass a non-negotiable part of the protocol, even when both columns feel locked in after pass one.
The second pitfall is the conjunctive-versus-disjunctive confusion. The stem uses 'and' in most items, but a meaningful minority use 'or' or 'but'. An 'or' stem requires the pair to satisfy one of the two clauses, not both. A 'but' stem requires the pair to satisfy a shared constraint and to differ on a specific dimension. The grammar of the stem is therefore a piece of the problem, not a stylistic choice. The slash on your scratch paper should also note which conjunction joins the two halves.
The third pitfall is column-header blind reading. The candidate reads the column entries, not the headers, and picks structurally valid pairs that get entered into the wrong selectors. The five-second header check is the cure, and it should be done before the first pass on column A.
The fourth pitfall is overwork. Two-Part Analysis items reward structured thinking and punish grinding. A candidate who spends five minutes on a single item and gets it right has lost more points than they earned, because the time cost carries forward to the next two items. The hard cap of 4 minutes is not a suggestion. When the cap arrives and the answer is not yet committed, make the most defensible pair and move on. In my experience coaching candidates in the 78 to 84 Data Insights band, the four-minute cap is the single biggest source of recovered points once it is enforced.
Comparing Two-Part Analysis with the other Data Insights families
The Data Insights section on the GMAT Focus contains five item families. Two-Part Analysis is one of them, and it is the family with the most distinctive scoring logic. The table below shows the four features that decide how you should triage each family during a section.
| Feature | Two-Part Analysis | Table Analysis | Multi-Source Reasoning | Data Sufficiency | Graphics Interpretation |
|---|---|---|---|---|---|
| Scoring unit | Binary pair (both picks must be correct) | Single answer | Single answer per tab | Binary five-way statement | Two drop-downs, both must be correct |
| Average time budget | 2:30 to 3:30 | 2:00 to 2:30 | 3:00 to 4:00 per tab pair | 2:00 to 2:30 | 2:00 to 2:30 |
| Primary cognitive demand | Joint reasoning across two columns | Pattern detection in a table | Synthesis across two to three tabs | Statement-by-statement logical check | Reading a graph, not a paragraph |
| Common failure mode | True-individually, false-jointly | Off-by-one row read | Tab-confusion | Premature commitment to one statement | Misread axis units |
The comparison makes the family-specific pattern visible. Two-Part Analysis is the only family in which the scoring is a strict binary pair, and that is the reason the four-move protocol is worth the time. Multi-Source Reasoning also requires synthesis, but the tabs are usually one-at-a-time scored. Graphics Interpretation is also a two-drop-down format, but the two drops are usually independent and can be solved as two sub-questions without the joint-evaluation pass. Data Sufficiency has its own binary structure, but the binary is across five statement combinations, not across two answer columns. Knowing the family-specific shape before the section starts is, in practice, a 3 to 5 point swing on Data Insights for candidates in the mid-70s.
How to read the shared scenario without overreading
The opening paragraph of a Two-Part Analysis item is a piece of structured business writing. It introduces entities, attaches a property to each, and often ends with a single constraint. The overreading trap is to treat the paragraph as a Reading Comprehension passage and look for nuance, voice, and implication. The paragraph is functional. Your job is to extract the entities, the properties, and the constraint, and to stop reading once those three are on the page. A useful exercise is to underline, on your first read, every proper noun, every number, and every comparative word such as 'more than', 'at least', or 'no later than'. The underlined words are the only things the test will quiz you on. Everything else is scaffolding.
Once the entities and properties are on the page, you have a small fact pattern in front of you. The table then gives you a set of candidates. The candidates either add new information (such as a candidate manager's prior project count) or impose new constraints (such as a candidate project's deadline). The reading order that saves the most time is paragraph first, table headers second, column A third, column B fourth. If you read the table entries before the paragraph, you will end up reading the paragraph twice, once to make sense of the entries and once to extract the constraint. If you read the paragraph first, you read it once, and the table becomes a set of questions with a clear schema.
Pacing, triage, and the section-level view
Two-Part Analysis items belong at the middle of a section-level triage plan, not at the start and not at the end. The section opens with whatever item family lets you settle into a steady rhythm, which for most candidates is Table Analysis or Graphics Interpretation. The middle of the section is where the harder Two-Part Analysis items live, and the structured four-move protocol gives you a way to handle them without spending more than your budget. The end of the section is reserved for the items you have not yet seen, and you want to leave 3 to 4 minutes at the end for a clean review pass, not for cracking a hard Two-Part Analysis item that you have been carrying since minute 22.
The pacing arithmetic is simple. With 20 questions in 45 minutes, you have an average of 2:15 per question. Two-Part Analysis runs long, so plan to take 3 minutes on each one and recover the time on the easier families. Three Two-Part Analysis items at 3 minutes each is 9 minutes, and the other 17 items have to average about 2:06 to keep the section inside the clock. That is achievable if you enforce a hard cap on the harder families. In practice, the candidates who score in the mid-80s on Data Insights are not the ones who spend 4 minutes on a hard item; they are the ones who spend 3 minutes, commit the best defensible pair, and bank the recovered minute for the next item.
How Two-Part Analysis preparation fits into a wider GMAT Focus plan
Two-Part Analysis is best prepared in two passes. The first pass is family-specific: take a block of 15 to 20 Two-Part Analysis items in isolation, time each one, and grade them by sub-question rather than by item. If a candidate misses an item because column A was wrong, the sub-question is A. If the candidate misses an item because column B was right but the cross-column evaluation was sloppy, the sub-question is joint. This sub-question grading surfaces the failure pattern that the four-move protocol is designed to fix. For most candidates in the 70s on Data Insights, the dominant failure pattern is joint-evaluation, not single-column elimination, and the four-move protocol closes that gap in a measurable way over a 10-day window.
The second pass is section-integrated. Mix Two-Part Analysis items into full-length Data Insights sections and grade them inside the section context. The point of the second pass is to enforce the pacing budget and the triage decision. Candidates who ace Two-Part Analysis in isolation but lose time on it inside a full section have a pacing problem, not a method problem. Candidates who lose the method under time pressure have a fatigue problem, and the cure is more timed mixed-family work, not more Two-Part Analysis in isolation. The mix should also include Multi-Source Reasoning and Data Sufficiency, because the two-drop-down logic of Graphics Interpretation and the binary statement logic of Data Sufficiency both shape the way you read the columns in Two-Part Analysis.
Preparation strategy at the section level also has a score-band reading. For candidates targeting 78 or below, Two-Part Analysis is a points-recovery family, and the four-move protocol should be drilled until it is automatic. For candidates targeting 84 or above, Two-Part Analysis is a tie-breaker family, and the marginal return comes from cross-column evaluation under fatigue, not from re-learning the protocol. The first group needs reps; the second group needs harder items and tighter pacing.
What the scoring report tells you about Two-Part Analysis performance
The GMAT Focus enhanced score report breaks Data Insights performance down by item family, including Two-Part Analysis. The report shows the percentage of items in the family that you answered correctly and the average time per item. Two readings matter. A high accuracy with high time usually means the protocol is working but the section pacing is bleeding points elsewhere. A low accuracy with normal time usually means the four-move protocol is not being executed, and the diagnostic move is to drop back into the family-isolation pass for a week. A low accuracy with high time is the worst combination and usually points to overwork on hard items, which the four-minute cap is designed to prevent.
The score report also shows the difficulty mix of the Two-Part Analysis items you received. The adaptive algorithm on the Focus edition adjusts the difficulty mix based on your performance on the first few items in the section. A candidate who finds Two-Part Analysis items in the easier band should still apply the full four-move protocol, because the binary scoring makes even easy items risky when the columns are coupled. A candidate who finds Two-Part Analysis items in the harder band should expect the conjunction to be load-bearing, the column headers to be scrambled, and the joint-evaluation pass to be the deciding move.
Putting it all together for test day
On test day, the only thing that matters is execution. Read the conjunction. Read the column headers. Pass one on column A, with column B physically covered. Pass two across the columns, looking for joint consistency. Commit the first jointly consistent pair, confirm, and move on. If both columns are still open at the four-minute cap, make the most defensible pair and move on. The protocol is the same on every item, regardless of difficulty, and that uniformity is the reason it works under time pressure.
The preparation strategy that supports this execution has two parts: a family-isolation pass of 15 to 20 timed items graded by sub-question, and a section-integrated pass of full-length Data Insights sections with the four-minute cap enforced. The first pass builds the protocol. The second pass builds the discipline. Candidates who do both passes typically recover 4 to 6 points on Data Insights over a 4 to 6 week window, and the recovered points come disproportionately from the Two-Part Analysis family.
TestPrep İstanbul's diagnostic assessment is a natural starting point for candidates building a sharper preparation plan around the Two-Part Analysis item family on the GMAT Focus Data Insights section.