GMAT Data Insights: Data interpretation, information sufficiency and source reasoning

Data Insights

Independent data-reasoning practice develops table and graph reading, information sufficiency, multi-source reconciliation and linked constraints. Current official question counts, timing, calculator rules and interaction formats are not established by these notes.

The five question types

Data Sufficiency: a question plus two statements; decide whether the data are sufficient, not what the value is.
Multi-Source Reasoning: two or three tabs of text, tables, or charts; questions ask you to judge statements as true/false or to select an option.
Table Analysis: a sortable table; each statement is judged true or false against it.
Graphics Interpretation: a graph with drop-down, fill-in-the-blank statements.
Two-Part Analysis: a table where you pick one value in each of two columns; both must be correct.

A five-case information-classification model

A: (1) alone sufficient, (2) alone not. B: (2) alone sufficient, (1) alone not. C: together sufficient, neither alone. D: each alone sufficient. E: together still not sufficient.

Worked example

Is x > 0? (1) x² > 0. This only establishes x ≠ 0; x could be 3 or −3. Not sufficient. (2) x³ > 0. Cubing preserves sign, so x must be positive. Sufficient.

Within that five-case model, statement 2 alone is sufficient. Classify against the options actually given in a practice item.

The discipline is to test awkward values - negatives, zero, proper fractions, non-integers - before declaring sufficiency. Statement (1) looks sufficient only if you try positive numbers exclusively.

Traps that cost marks

Calling a statement sufficient after finding a single satisfying value.
Forgetting that a variable may be negative or fractional when the question imposes no restriction.
In Two-Part Analysis, choosing two values that satisfy one constraint but violate the other.
In Table Analysis, misreading a column or failing to re-sort before judging a statement.
Skipping a tab in Multi-Source Reasoning and missing the data that decides the question.
Reaching for the calculator before deciding what actually needs computing; it verifies arithmetic, not reasoning.

Evidence, sufficiency and quantitative comparison

Information is sufficient when it determines the answer to the question being asked, within the stated domain. A statement need not identify an exact value to answer a yes-or-no question consistently. Conversely, finding one example that gives a yes does not establish that every permitted example does. First rewrite the target question precisely; then test each supplied statement independently before combining them.

For example, if x is a real number and x²=49, its value could be 7 or-7. This is insufficient to determine x uniquely, but it is sufficient to answer whether x is nonzero. Adding x>0 determines x=7. The question matters as much as the numerical equation. Do not import a second statement when checking the first alone, and do not assume positivity, integrality or distinctness unless it is actually given.

Tables and charts introduce source discipline. Read the label, unit, date or period, denominator and any note before comparing values. A total in thousands is not directly comparable to a single-unit count without conversion. A percentage based on all responses can differ from the percentage based on respondents to one item. A share of a total, a percent increase from a baseline and a percentage-point difference answer different questions. State which one the problem requests.

In a multi-source model, assign each fact to its source and time. An original estimate and a revised estimate are not necessarily contradictory: they may answer different versions of the same planning question. If three people cost 120 per hour and two cost 90 per hour, the combined hourly rate is 540. At 200 hours the modeled labor cost is 108000; at 250 hours it is 135000. A budget note cannot silently replace the requested original duration. These are fictional calculations, not labor-market facts.

Finally, distinguish consistency from proof. A rule may fit every supplied term of a finite sequence without being the only mathematical continuation possible. Such exercises ordinarily ask for the intended simple pattern represented by the available choices. Explain that rule and test it on all supplied pairs. In an argument question, similarly identify the premise, proposed conclusion and any missing assumption. A plausible conclusion is not automatically entailed. A counterexample can disprove a universal claim, while several supporting examples alone do not establish one.