The Two-Part Analysis item on the GMAT Data Insights section is a single question that asks for two coordinated answers, usually a pair of values, statements, or expressions, both of which must be selected before the item can be scored. It is the only Data Insights format where the test-taker's screen shows a small grid of choices rather than five lettered options, and that single visual detail is the source of most of the errors candidates make in the first ten practice items. Understanding the grid is the first task; solving the underlying relationship is the second; timing the whole thing as a single item, not a double, is the third.
What the item actually looks like on screen
On test day, Two-Part Analysis items are presented as a short stem, a small table or description of a real-world situation, and a grid of between five and seven answer choices. The candidate must select one cell from the first column and one cell from the second column, where the two columns typically correspond to the two requested values. A small banner above the grid indicates that exactly one choice in each column is required. There are no partial scores: the GMAT treats a Two-Part Analysis item as a single item worth roughly the same scaled points as one Quantitative Reasoning item or one Data Insights problem. The Data Insights section, which is the newer of the three sections on the current exam, includes roughly twenty items covering four item families, of which Two-Part Analysis is one of the four.
Candidates often arrive expecting a hybrid between Data Sufficiency and Quantitative Reasoning, and the resemblance is real only on the surface. The stem usually contains a piece of descriptive text, sometimes with a small table, and the candidate's job is to find two values that jointly satisfy a stated condition. There is no 'statement (1) and statement (2)' mechanic. There is no 'if and only if' sub-language. The work is to read the situation, translate it into equations or matching rules, and then map the two answers onto the grid.
The scoring grid is not just a different shape of multiple choice
Most test-takers treat the grid as decoration and click two cells based on the math they just did. That is the wrong order of operations. The grid is a constraint that the math must satisfy, and it changes which solving path is fastest. Three things to set up before reading the stem in any real sense:
- Count the columns. A grid with three columns and four rows is not the same item type as a grid with two columns and five rows, even when the stem looks identical. The number of columns tells you how many values you owe.
- Look for header units. The column headers almost always carry units, currency, or a yes/no meaning. Misreading a header turns a correct calculation into a wrong submission.
- Check for shared values. In a large share of items, one of the choices in column A is identical to one of the choices in column B, and the test-maker uses that overlap to tempt candidates into a self-contradicting pair.
For most candidates, the single highest-leverage habit to build is a ten-second pre-read of the grid before touching the stem. It costs nothing, it is allowed under the rules, and it is the cheapest way to avoid the classic 'I solved the right problem, I just clicked the wrong cell' error that shows up in roughly half of the missed Two-Part Analysis items in our error logs.
The two most common traps and how to avoid them
Two-Part Analysis items on the GMAT reward a particular kind of disciplined reading, and the same two traps show up in cohort after cohort. The first trap is treating the item as two independent sub-questions. It is not. The two requested values are constrained to come from one situation, so they share a budget, a total, or a structural relationship that a single equation can describe. The candidate who solves each column separately almost always finds an answer for column A that is inconsistent with the answer for column B, and only notices the contradiction after clicking submit. The fix is mechanical: write the joint equation first, then solve for the two values, then look for them in the grid.
The second trap is over-engineering the algebra. Two-Part Analysis items on the current exam tend to use numbers that collapse to a single arithmetic step once the relationship is identified. In my experience, candidates who reach for a system of two equations are usually responding to the visual shape of the grid rather than the structure of the stem. If a system of equations is genuinely the shortest path, the grid will usually have more columns or a yes/no structure. For two-value items with two columns, a single relation with one parameter is the typical move.
Rule of thumb: if you are writing more than two distinct algebraic lines before you touch the grid, you have probably misidentified the joint relationship.
A worked example, using a representative stem
Consider a stem that describes two investments, X and Y, with the total capital split between them in a 3 to 2 ratio. Investment X returns 4 per cent and investment Y returns 6 per cent, and the combined annual return is 1,200 dollars. The grid has two columns and five rows: column A lists dollar amounts for the return from X, and column B lists dollar amounts for the return from Y. Most candidates will set up two equations, solve for the split, and then walk into the grid. The cleaner path is to use the 3:2 ratio to parametrise the split. If X receives 3k and Y receives 2k, then 0.04 times 3k plus 0.06 times 2k equals 1,200, which simplifies to 0.12k plus 0.12k equals 0.24k, so k equals 5,000. The return from X is 0.04 times 15,000, which is 600; the return from Y is 0.06 times 10,000, which is 600. The pair to select is therefore 600 in column A and 600 in column B, a coincidence that the grid deliberately allows for.
Notice the structure. The arithmetic was light. The win came from reading the ratio as a single parameter, not as two unknowns. The grid was checked last, not first. That sequencing is what a senior tutor would drill before letting a candidate near a timer.
Pacing within a Data Insights section
The Data Insights section gives candidates roughly 45 minutes for about 20 items, which works out at just over two minutes per item on average, although the four item families are not paced uniformly. Two-Part Analysis items tend to take longer than a well-prepared candidate expects, because the grid adds a small fixed cost of fifteen to twenty seconds, and because the joint solving often forces a back-check. A realistic budget for a practiced test-taker is two and a half minutes per Two-Part Analysis item, against ninety seconds for a well-drilled Multi-Source Reasoning item. Plan for that gap, or the last two items of the section will be the ones you could have solved.
| Data Insights item family | Typical count per section | Suggested time budget |
|---|---|---|
| Data Sufficiency | approximately 5 | around 2 minutes |
| Table Analysis | approximately 3 | around 2 minutes |
| Multi-Source Reasoning | approximately 5 | around 90 seconds |
| Two-Part Analysis | approximately 5 | around 2.5 minutes |
These figures are not sourced to a single official table; they are the working budget that experienced tutors teach, derived from the section length, the section timing, and the structure of each family. Candidates using a stricter or looser personal budget should still keep the relative ranking intact.
How to drill Two-Part Analysis in a GMAT preparation programme
A focused block of ten to fifteen Two-Part Analysis items, solved in a single sitting with the timer running, will expose a candidate's habit patterns faster than any number of unmixed Quantitative Reasoning drills. Three tactical rules for that block:
- Pre-read the grid for ten seconds before reading the stem, and write down the units of each column header.
- Solve the joint relationship first, then look for the pair in the grid. Do not solve each column independently.
- After each item, write one sentence describing the trap you avoided or fell into. Review those sentences at the end of the block.
Candidates who follow that protocol for three sittings typically drop their per-item time on Two-Part Analysis by thirty to forty seconds and reduce their grid-misread error rate, in our internal logs, by a noticeable margin. The wins are mechanical and reproducible, which is exactly the point of a structured GMAT preparation course rather than open-ended self-study.
Common pitfalls and how to avoid them
- Reading the grid after the calculation. Reverse the order: grid first, then stem, then math, then back to the grid.
- Treating the two columns as independent sub-questions. They share a relationship. Solve the relationship before picking cells.
- Ignoring column units. A '6' in a per cent column and a '6' in a dollar column are not interchangeable.
- Over-reaching on the algebra. Most items collapse to a single arithmetic step. Two equations and two unknowns is usually a sign that you have misread the joint relationship.
- Skipping the back-check. The joint structure means that if column A contradicts column B, you have misidentified the relationship, not made a small slip.
Two-Part Analysis is the Data Insights item family that punishes sloppy reading the most heavily, and the one that rewards the cleanest prep routine the most generously. The candidates who improve fastest on it are the ones who stop trying to be clever and start reading the grid as a constraint, not a decoration.
Where this leaves a candidate
A 700-plus total on the current GMAT requires a steady hand on each of the four Data Insights item families, and Two-Part Analysis is the family where reading discipline matters as much as arithmetic. The next step is to take a block of ten to fifteen mixed items under timed conditions, apply the grid-first, joint-relationship-second protocol, and review the trap sentences at the end. That single routine is what converts a Data Insights score in the high teens into one in the mid-twenties, and it is the natural bridge between self-study and a structured programme.
GMAT Courses' one-to-one GMAT preparation programme works through each candidate's Two-Part Analysis error log line by line, and turns the grid-first protocol into a habit that survives timed conditions.