What is GMAT Data Insights?
Data Insights (DI) is the newest scored section of the GMAT Focus Edition: 20 questions in 45 minutes, spanning five distinct question formats — Data Sufficiency, Multi-Source Reasoning, Table Analysis, Graphics Interpretation, and Two-Part Analysis. It is also the only section where an on-screen calculator is permitted. The section measures data literacy: reading dashboards, weighing incomplete evidence, and deciding when information is sufficient to act on — the exact skills MBA classrooms and modern management demand.
DI is where the old exam's Integrated Reasoning content and Quant's Data Sufficiency format merged into one full-weight section. Because it pulls from the widest variety of formats, most candidates arrive underprepared for it — which makes it, paradoxically, the section with the greatest score upside: the population average here is the lowest of the three sections, so focused preparation moves both your section score and your total disproportionately.
Format at a glance
| Feature | Detail |
|---|---|
Questions | 20 questions across five formats (several are multi-part) |
Time | 45 minutes (average ≈ 2 minutes 15 seconds per question) |
Question types | Data Sufficiency · Multi-Source Reasoning · Table Analysis · Graphics Interpretation · Two-Part Analysis |
Calculator | On-screen calculator available throughout this section only |
Adaptive? | Yes — difficulty adjusts to your performance as you go |
Review & edit | Bookmark anything; change up to 3 answers per section if time remains |
Section score | 60–90 scale, weighted equally toward your 205–805 total |
Data Sufficiency — deep dive
Data Sufficiency (DS) moved into DI from the old Quant section, and it remains the format most responsible for section-score swings. Each question presents a problem followed by two statements. You never need to produce the final answer — you must judge whether the information given is sufficient to determine a unique answer.
The five answer options are identical on every DS question ever written:
- Statement (1) ALONE is sufficient, but statement (2) alone is not
- Statement (2) ALONE is sufficient, but statement (1) alone is not
- BOTH statements TOGETHER are sufficient, but neither alone is
- EACH statement ALONE is sufficient
- Statements (1) and (2) TOGETHER are NOT sufficient
The AD/BCE method
Why it matters: a fixed decision tree eliminates option-shopping forever
- Evaluate statement (1) alone. If it's sufficient, your answer is (A) or (D) — eliminate B, C, E.
- If (1) fails, your answer is (B), (C), or (E) — eliminate A and D.
- Then evaluate (2) alone. If it failed too, test them only together for C vs E.
- Never evaluate the statements together before testing (2) alone — that error produces exactly the wrong answer half the time.
Sufficiency means unique determination
Why it matters: the conceptual heart of every DS question
- “Sufficient” means the statements pin down exactly one value or a definite yes/no — not that you can compute it easily.
- A single equation with two unknowns is usually insufficient — but not always (x + y = 5 with x, y positive integers has exactly two solutions; check constraints).
- For yes/no questions, “always yes” and “always no” are both sufficient. “Sometimes yes, sometimes no” is not.
Content sources inside DS
Why it matters: DS borrows from every quant topic — your foundation carries over
- Algebra and equations: most common source.
- Number properties: parity, signs, divisibility constraints often decide sufficiency without any solving.
- Rates, percents, statistics: word-problem wrappers around the same sufficiency logic.
Multi-Source Reasoning — deep dive
MSR presents two to three tabbed pages — email threads, memos, tables, charts — and asks questions that require combining them. Questions come in sets sharing the same source material, so organization invested up front pays off three or four times.
Skim every tab first
Why it matters: answering from one tab while blind to another is the classic MSR failure
- Spend 30–40 seconds cataloguing what each tab contains before touching a question.
- Note cross-references: dates, names, and figures that appear on multiple tabs are usually load-bearing.
Expect binary decisions
Why it matters: most MSR items ask yes/no or true/false on multiple statements
- You may be judging three independent mini-statements; each is evaluated separately.
- Watch for contradictions between tabs — tension between sources is usually exactly what the question targets.
Table Analysis — deep dive
Table Analysis gives you a sortable spreadsheet — dozens to hundreds of rows — plus statements to judge true/false based on the data. The skill being tested is sorting strategy, not calculation.
Sort to split, then verify visually
Why it matters: each sort should do real analytical work
- Pick the column that divides the data closest to the threshold mentioned in the statement — then count rows above/below it directly.
- Averages and ratios: compute once from column totals instead of scanning row by row.
- Each sub-statement stands alone — one wrong judgment never cascades.
Graphics Interpretation — deep dive
You interpret a chart, scatter plot, or graph and complete two drop-down sentences. Formats mirror real analytics dashboards: clustered bars, time series, distributions, scatter plots with trend lines.
Read axes and units first
Why it matters: most GI errors live entirely in misread scaling
- Check axis labels, units (thousands? percent?), and whether scales start at zero before interpreting anything.
- Answer choices in GI are engineered far apart — estimate positions rather than demanding precision.
Know the vocabulary of trends
Why it matters: drop-downs frequently test statistical language
- Correlation direction and strength (positive/negative, strong/weak), clustering, outliers, and gaps between groups.
- “Approximately what fraction” questions: benchmark regions against halves and quarters of the plotted area.
Two-Part Analysis — deep dive
Two-Part Analysis poses one problem requiring you to pick a value in each of two columns so that both conditions hold simultaneously. Content may be quantitative or verbal/logical.
Treat the parts as one system
Why it matters: solving columns independently doubles the work
- Define what links the two answers (sum, difference, shared constraint) before computing anything.
- Back-solve strategically: test candidate pairs against both conditions rather than deriving algebraically from scratch.
- The two chosen values must work together — a choice correct for column one means nothing until the pair validates.
Using the calculator wisely
The on-screen calculator feels like a gift and quietly costs points. It handles arithmetic well but adds clicks and screen-focus to every operation. Strong performers use it surgically:
| Use the calculator for… | Skip it for… |
|---|---|
Messy multi-digit multiplication and division | Benchmark fractions and percents (know 1/8 = 12.5%) |
Verifying a computed ratio or percentage | Single-step arithmetic you can do mentally |
Compound growth computations | Estimation when choices are far apart |
Pacing strategy
| Checkpoint | Elapsed time | Budget left | If you're behind |
|---|---|---|---|
Question 7 | ~16 min | 29 min for 13 | Cut second passes on table sorts; trust first reads |
Question 14 | ~31 min | 14 min for 6 | Cap MSR sets: answer and move, don't perfect |
Question 18 | ~40 min | 5 min for 2 | Drop-down estimates over precise computation |
Question 20 | ≤45 min | — | Never leave blanks — guess on anything remaining |
DI's formats differ wildly in time cost: single DS items can take 60 seconds while an MSR set legitimately consumes 6–8 minutes across its questions. Budget per set, not just per question, and know your exit point if a set turns hostile.
How Data Insights scoring works
DI reports on the same 60–90 scale, weighted equally into the 205–805 total. Its historical population average is the lowest of the three sections — meaning the typical pool performs worst here. Two strategic consequences follow:
- DI is your cheapest percentile gain. The same absolute improvement yields more percentile movement in DI than elsewhere.
- Admissions committees notice balanced profiles. A strong total dragged down by weak DI raises data-literacy questions; solid DI alongside good Quant/Verbal signals classroom readiness.
A 12-week Data Insights study plan
| Phase | Weeks | Daily focus | Exit criteria |
|---|---|---|---|
DS foundation | 1–3 | Learn the five options cold; drill pure DS untimed using AD/BCE method; keep quant fundamentals warm | 80%+ untimed DS accuracy with clean method notes |
Format tour | 4–6 | One new format per week (Table → Graphics → Two-Part → Multi-Source); learn each format's specific protocol | No format feels unfamiliar; 70%+ accuracy everywhere |
Mixed integration | 7–9 | Timed 20-question mixed blocks at full pace; weekly error-pattern audit by format | Consistent ~2:15/question pacing with 70%+ accuracy |
Simulation | 10–12 | Full mocks weekly; targeted repair of weakest format between mocks | Stable target scores across consecutive mocks |
7 costly mistakes to avoid
| Mistake | Why it happens | The fix |
|---|---|---|
Treating DI as “extra math practice” | Assuming quant skills transfer automatically | Drill each of the five formats explicitly — they have distinct protocols |
Evaluating DS statements together too early | Natural instinct to combine information | Strict AD/BCE sequence until it's reflexive |
Calculator dependency | It's available, so it feels efficient | Reserve it for genuinely messy arithmetic only |
Reading MSR tabs mid-question | Rushing to the first question | Catalogue all tabs first — 40 seconds that saves 3 minutes |
Misreading axis units in Graphics Interpretation | Dashboards feel familiar, inviting skimming | Labels and units first, interpretation second — always |
Perfectionism on hostile MSR sets | Sunk-cost attachment to invested reading time | Set a per-set time ceiling; bank partial credit and move |
Leaving blanks when clock runs low | Hoping remaining time stretches further than it does | Hard rule: last minute = answer everything remaining |
Frequently Asked Questions
Q1Is Data Insights harder than Quant?
It's broader, not deeper. Quant tests one format deeply; DI tests five formats moderately. Most test-takers find DI unfamiliar rather than difficult — and unfamiliarity is exactly what structured preparation removes.
Q2Why did Data Sufficiency move out of Quant?
The Focus Edition redesign reframed DS as a data-literacy task — judging evidence sufficiency — which fits DI's mission better than computational math. Quant became pure Problem Solving.
Q3Can I use a calculator on all sections?
No. The on-screen calculator exists only during Data Insights. Quant and Verbal provide none, and physical calculators are banned everywhere.
Q4What percentage of DI questions are Data Sufficiency?
DS typically forms the largest single share of the section, but exact splits vary by exam form. Prepare as if any mix of the five formats can appear — that's the honest assumption.
Q5Do I need mental math even though I get a calculator?
Yes. Benchmark conversions, estimation, and sanity-checking remain faster and safer than calculator dependence, and the reasoning step — deciding what to compute — is always manual.
Q6How long is a GMAT DI score valid?
All GMAT scores, including Data Insights, remain valid for five years from your test date.
Q7What's a good Data Insights score?
Because the population skews lower on DI, matching your other section scores here already signals strength. Aim to keep DI within a couple points of your Quant and Verbal results for a balanced profile.
Q8Should I practice DI with third-party materials?
Prioritize official GMAC questions for format fidelity — DI's interfaces (sortable tables, tabbed sources, drop-downs) are part of what you're learning. Third-party content helps for volume after official material is exhausted.
Complete the picture
Explore the Quantitative Reasoning and Verbal Reasoning guides to cover every corner of the GMAT Focus Edition.