GMAT Two-Part Analysis: two answers, one chance.

Two-Part Analysis is the most under-prepared question type on the GMAT Focus Edition. Students drill Data Sufficiency and Critical Reasoning for weeks, then meet Two-Part Analysis for the first time inside Data Insights and discover that it behaves like nothing else on the exam. It is the one format where every prompt is a small puzzle with its own rules, where the answer choices are shared across two questions, and where getting one column right buys you nothing if the other is wrong. The whole item is scored as a single unit.

Two-Part Analysis is graded all-or-nothing. You can do 90% of the work correctly, miss one column, and earn exactly the same credit as a student who guessed both. That scoring rule is the reason this format rewards precision over speed.

What Two-Part Analysis actually is

Two-Part Analysis sits inside the Data Insights section alongside Data Sufficiency, Multi-Source Reasoning, Table Analysis, and Graphics Interpretation. Each item gives you a short prompt — a paragraph of conditions, a scenario, a small set of relationships — followed by a table. The table has two answer columns with headers describing two related quantities or claims, and a list of candidate answers running down the side. Your job is to select exactly one option for the first column and exactly one for the second.

The candidates are shared. The same row of options serves both columns, and in most items a single option could legitimately be selected in either column — the prompt is what decides which one belongs where. That shared list is the source of most of the difficulty, because it lets the test-writer build a close decoy that is correct for the other column.

The two flavors

Two-Part Analysis items come in two broad flavors, and knowing which one you are looking at within the first ten seconds tells you how to attack it.

Quantitative two-part items

These give you a scenario governed by numbers — rates, ratios, a system of relationships, a budget, an allocation. The two columns ask for two values that together satisfy the scenario. Most of these reduce to solving a system: two unknowns, two constraints, and a list of candidate values to test or compute against.

Verbal and analytical two-part items

These give you an argument or a set of logical conditions. The columns might ask you to select a statement that, if true, most strengthens the conclusion and a separate statement that most weakens it; or a claim that must be true and one that cannot be true given the constraints. The math is gone, but the two-column trap is identical: the options are shared, and the wrong answer for one column is usually the right answer for the other.

The distinction that decides everything

Before you compute or evaluate anything, answer one question: are the two columns independent or dependent?

In an independent item, the answer to column one does not constrain the answer to column two. You solve each column on its own terms. A strengthen-and-weaken pair is usually independent: the statement that strengthens has nothing to do with the statement that weakens.

In a dependentitem, the two columns are tied together by the scenario, and the answers have to be consistent with each other. A system of two equations is the canonical case: the value you pick for the first unknown forces the value of the second. In dependent items you cannot evaluate a column in isolation — you have to find the pair that satisfies the whole scenario at once.

Misclassifying the relationship is the most expensive mistake in this format. Treat a dependent item as two independent choices and you will pick a value that is locally plausible but globally inconsistent — the exact decoy the writer built.

A worked quantitative example

Here is an invented item in the quantitative style. A courier service charges a fixed dispatch fee plus a constant rate per kilometer. A 10-kilometer delivery cost a customer 26 dollars. A 25-kilometer delivery cost 50 dollars. The table asks you to select the dispatch fee in the first column and the per-kilometer rate in the second, from a shared list of candidates: 1.20, 1.40, 1.60, 8, 10, 12.

This is a dependent item — the fee and the rate have to jointly produce both totals. Set up the system. Let f be the dispatch fee and r the per-kilometer rate. Then:

• 10-km trip: f + 10r = 26
• 25-km trip: f + 25r = 50

Subtract the first from the second: 15r = 24, so r = 1.60. Substitute back: f + 16 = 26, so f = 10. The dispatch fee is 10 dollars and the per-kilometer rate is 1.60 dollars.

Notice the decoys. The list contains 1.20 and 1.40, both plausible-looking rates that a student who divides a single total by its distance would land on (26 divided by 10 is near neither, but a careless 50 divided by 25 gives 2, and rounding errors steer panicked test-takers toward the nearby values). The 8 and 12 sit on either side of the correct fee of 10, ready to catch an arithmetic slip in the substitution step. Every wrong option was placed to reward a specific error.

The five recurring traps

Trap 1 — Reversing the columns

You do the math correctly, then enter the rate in the fee column and the fee in the rate column. The values are right; the placement is wrong; the item is marked wrong. The shared answer list makes this trivially easy to do under time pressure. Fix: read the column headers last, right before you select, and physically check that the number you are about to choose answers the header above it.

Trap 2 — Solving one column and guessing the other

You nail the first column, the clock is loud, and you eyeball the second. Because the item is all-or-nothing, a confident first column plus a guessed second column is worth zero. Fix: budget for the whole item, not the first half. If you cannot finish both columns, the rational move is often to guess both quickly and bank the time elsewhere.

Trap 3 — Treating a dependent item as independent

You evaluate each column on its own and pick the locally best option for each, ignoring that the scenario ties them together. The result is a pair that looks fine column by column but cannot both be true. Fix: classify the item as dependent or independent before you compute, and in dependent items always verify the pair against the full scenario.

Trap 4 — The shared-option lure

In verbal two-part items especially, the statement that most strengthens and the statement that most weakens are placed next to each other in the list precisely because they are easy to swap. A statement that weakens a conclusion can read as a strengthener if you misread the direction of the argument. Fix: restate the conclusion in your own words before you evaluate any option, so the direction is fixed in your head.

Trap 5 — Over-reading the prompt

The prompt paragraph is short, but anxious test-takers re-read it three times looking for hidden conditions that are not there. Two-Part Analysis prompts are spare by design; the difficulty lives in the columns, not in buried clauses. Fix: read the prompt once for structure, extract the constraints into a quick note, and move to the table.

Timing

Data Insights gives you 45 minutes for 20 questions — an average of 2 minutes 15 seconds each. But the section mixes five formats of very different weight. Multi-Source Reasoning sets and dense Table Analysis items run long; a clean Data Sufficiency question runs short. Two-Part Analysis sits in the middle, and a fair budget is 2:15 to 2:45 depending on flavor.

• Quantitative two-part: aim for 2:30. Most of the time goes into setting up and solving the system; the selection itself is fast once you have the values.
• Verbal two-part: aim for 2:15. The evaluation is quick if you fix the conclusion first; the risk is re-reading, not computation.
• Hard cap at 3:30. Past that point, on an all-or-nothing item with no partial credit, the expected value of staying is low. Pick the most defensible pair and move on.

On every other GMAT question, slow accuracy still earns the point. On Two-Part Analysis, slow accuracy on one column with a rushed error on the other earns nothing. The format punishes uneven effort more than any other on the exam.

How to drill Two-Part Analysis

1. Classify before you solve.For every practice item, write “dependent” or “independent” before doing any work. Build the habit until the classification is automatic.
2. Separate the two flavors. Do a set of only quantitative two-part items, then a set of only verbal ones. The skills barely overlap, and mixing them early hides which one is actually weak.
3. Audit your wrong column.When you miss an item, identify which column you got wrong and why. If you consistently miss the second column, it is a timing problem, not a content problem — you are running out of clock.
4. Verify the pair, every time. On dependent items, plug both selected values back into the scenario as a final step. This single habit kills the most common silent error.
5. Tag the miss in your error log. A reversed column is a Careless tag; a misclassified dependency is a Strategy tag; a guessed second column is a Time Pressure tag. The tag tells you which fix to drill next.

The short version

Two-Part Analysis is two questions scored as one, with a shared answer list built to make you swap columns or pick a locally plausible but globally wrong pair. Classify the item as dependent or independent before you compute. Solve the whole item, not the first half. Read the column headers last and confirm each value answers the header above it. On dependent items, plug the pair back into the scenario. Budget 2:15 to 2:45, hard-cap at 3:30, and remember that on an all-or-nothing item a guessed second column erases a perfect first one.

The platform

Zakarian GMAT's Data Insights chapters teach Two-Part Analysis as its own discipline, with the dependent-versus- independent decision built into the worked examples and the problem sets. The practice runner tracks per-question time so you can see exactly where the second column is eating your clock, and the error log's six-tag taxonomy separates a reversed column from a misclassified dependency from a time-pressure guess. The sample chapter is free if you want to see the teaching first.