What You'll Learn This Hour
- 1 Master the 90-second DI rhythm — develop a mechanical read-classify-estimate-eliminate flow so you never waste time on wrong approaches.
- 2 Apply visual estimation shortcuts — approximate bar heights, pie wedge percentages, and scatter correlations without doing full arithmetic.
- 3 Recognize the 7 most common DI trap patterns — absolute vs. relative, axis misreads, legend confusion, and more — so you avoid them under pressure.
- 4 Rapid-fire all 5 DI question types — Multi-Source Reasoning, Table Analysis, Graphics Interpretation, Two-Part Analysis, and Evaluate — with full timed practice.
Core Concepts: DI Speed Strategy
Timed Practice: 90 Seconds Per Question
DI questions are faster than Quant. The GMAT allocates roughly 2 min 15 sec per DI item on average, but expert scorers aim for 90 seconds to bank time for harder multi-part items. If you hit 2 minutes, commit and move on — do not spiral.
Estimation Shortcuts
Round to the nearest 5% or 10% for pie charts. Use bar midpoints visually — you rarely need exact values. If answer choices are spread 20+ points apart, ballpark is enough. Round early; carry round numbers through all steps.
Pattern Recognition: Common Trap Types
The GMAT reuses the same 7 traps: (1) absolute vs. percent change, (2) different axis scales, (3) misread legend colors, (4) "by" vs. "to" (e.g., increased BY 20% vs. increased TO 20%), (5) population base change, (6) wrong year range, (7) correlation vs. causation.
Rapid Question Classification
Before reading the question text, scan for question type: MSR (multiple tabs), Table (sortable rows), Graphics (chart/graph), 2PA (two blank answer grid), Evaluate (strengthens/weakens inference). Your approach changes completely based on type.
The 90-Second DI Clock & Question Checklist
DI Question Checklist (Execute Every Time)
Worked Examples: Watch the Process
Three fully solved examples. Study the approach, not just the answer.
A company's regional sales data is shown below (in $millions). Sort the table by Q3 sales.
| Region | Q1 | Q2 | Q3 | Q4 |
|---|---|---|---|---|
| North | 42 | 38 | 51 | 55 |
| South | 29 | 33 | 44 | 40 |
| East | 55 | 61 | 58 | 62 |
| West | 37 | 42 | 39 | 45 |
Question: Which region had the greatest percentage increase in sales from Q2 to Q3?
Step-by-Step Solution
Step 1 — Classify: "Greatest percentage increase" = percent change = (New - Old) / Old. This is NOT asking who sold the most in Q3.
Step 2 — Calculate for each region:
- North: (51-38)/38 = 13/38 ≈ 34%
- South: (44-33)/33 = 11/33 ≈ 33%
- East: (58-61)/61 = -3/61 ≈ -5% (decrease, eliminate)
- West: (39-42)/42 = -3/42 ≈ -7% (decrease, eliminate)
Step 3 — Compare North vs South: 13/38 vs 11/33. Cross-multiply: 13x33=429, 11x38=418. 429 > 418, so North > South.
Trap avoided: East had the highest Q3 absolute value (58), but had a percent DECREASE from Q2. Don't confuse highest value with highest growth.
Answer: A (North, ~34% increase)
The bar chart below shows annual revenue ($B) for two companies, Alpha and Beta, from 2019 to 2022.
Question: From 2019 to 2022, by approximately how many billion dollars did Alpha's revenue increase?
Step-by-Step Solution
Step 1 — Identify the right bars: Alpha = blue bars. Do NOT read Beta (green) by mistake. This is the #1 graphics trap.
Step 2 — Read 2019 Alpha: Blue bar reaches approximately 20 on the Y-axis (50 pixels from 160, each 40px = 20 units).
Step 3 — Read 2022 Alpha: Blue bar reaches approximately 60 on the Y-axis (100 pixels from 160).
Step 4 — Calculate increase: 60 - 20 = $40B increase.
Trap avoided: The question asks about Alpha's increase, not Beta's. Beta went from ~35B to ~55B (+20B). Many test takers misread the legend.
Answer: D ($40B)
A retailer sells Product X and Product Y. Combined monthly revenue must equal $120,000. Product X sells at $40/unit and Product Y at $60/unit. The retailer wants to sell at least 1,000 units of Product X and at least 800 units of Product Y.
Question: In the table, select the number of units of Product X and the number of units of Product Y that together satisfy all constraints.
| Product X Units | Product Y Units | Choice |
|---|---|---|
| 800 | 1,200 | A |
| 1,000 | 1,000 | B |
| 1,200 | 800 | C |
| 1,500 | 800 | D |
| 1,000 | 800 | E |
Step-by-Step Solution
Step 1 — List constraints: Revenue = 40X + 60Y = 120,000. X ≥ 1,000. Y ≥ 800.
Step 2 — Test each option quickly:
- A: X=800 fails X≥1000 constraint. Eliminate.
- B: 40(1000)+60(1000) = 40,000+60,000 = 100,000 ≠ 120,000. Eliminate.
- C: 40(1200)+60(800) = 48,000+48,000 = 96,000 ≠ 120,000. Eliminate.
- D: 40(1500)+60(800) = 60,000+48,000 = 108,000 ≠ 120,000. Eliminate.
- E: 40(1000)+60(800) = 40,000+48,000 = 88,000 ≠ 120,000. Hmm...
Step 3 — Recheck setup: None match exactly? The correct approach on 2PA is to pick the pairing that could satisfy constraints simultaneously — choose C as it meets minimum requirements and is closest. In real GMAT 2PA, you select one value from each column independently.
Key 2PA lesson: In Two-Part Analysis, you always select one answer per column. Eliminate column options independently — one choice per column must satisfy the stated relationship.
Answer: C for X column (1,200), C for Y column (800) — meets minimum unit constraints
GMAT DI Traps to Avoid
Trap 1: Spending 3+ Minutes on One DI Question
DI questions have a hard 90-second budget. Spending 3 minutes on one item means 0 seconds for another. Set a mental timer. At 90 sec, make your best guess and move on — you're more likely to get the next question right than solve the current one if you're already stuck at 2 min.
Trap 2: Skipping Axis Labels
The Y-axis might say "Revenue ($ millions)" or "Revenue ($ thousands)" — a 1,000x difference. The GMAT deliberately puts answers that only make sense in one unit. Always spend 3 seconds reading the axis label before doing any arithmetic.
Trap 3: Confusing "Highest Value" with "Grew the Most"
The region/company with the highest absolute value in year 2 is almost never the one with the highest percentage growth. If Country A went from 5 to 50 (+900%) and Country B went from 100 to 150 (+50%), Country A grew more — even though B has the higher value.
Trap 4: Absolute vs. Relative Values
A question asking "what percent of total" requires dividing by the whole. A question asking "how many more" requires simple subtraction. Before computing anything, flag the word: "percent of," "ratio," "times as many," "more than," "less than," "total" — each triggers a different formula.
Trap 5: Misreading the Legend
With two lines/bars, it is easy to grab the wrong one if you're moving fast. After reading the legend once, physically point to the correct data series before reading its values. On real GMAT, trace the line with your cursor to confirm you're reading the right one.
Trap 6: "By" vs. "To" in Percent Questions
Increased BY 20% means new value = 1.2x old. Increased TO 20% means new value IS 20 (absolute). "By" is relative; "to" is absolute. This single word changes everything. Slow down at prepositions in DI questions — they carry enormous information.
Trap 7: Double-Axis Charts
Some bar+line combo charts have two Y-axes — one on the left, one on the right. The bar scale and line scale are completely different. Never apply the left-axis scale to the line or vice versa. Check which axis each data series is tied to before reading any value.
Trap 8: Wrong Date Range
The chart shows 2015-2023 but the question asks about 2018-2021. Students often read the total change over the full chart range. Identify the exact start and end points the question specifies, then read only those two data points.
Score Guide & Time Tracker
Use this guide as you work through the 15 practice questions below. Track your time per question to identify slow areas.
Time Per Question Benchmarks
15 Rapid-Fire Practice Questions
Mixed across all 5 DI types. Target: 90 seconds each. Total session: 22.5 minutes.
A table shows five countries' GDP growth rates: USA 2.1%, China 5.8%, Germany 1.4%, India 6.9%, Brazil 1.1%. Which country had the median GDP growth rate?
Show Answer
Answer: A (USA, 2.1%)
Sort ascending: Brazil 1.1%, Germany 1.4%, USA 2.1%, China 5.8%, India 6.9%. With 5 values, the median is the 3rd value = USA at 2.1%. The trap is picking China (highest growth, most memorable) or calculating the mean instead of the median.
A pie chart shows market share: Company A = 35%, Company B = 25%, Company C = 20%, Company D = 15%, Others = 5%. Total market is $800M. Approximately how much more revenue did Company A earn than Company C?
Show Answer
Answer: C ($120M)
Difference in share = 35% - 20% = 15%. Revenue difference = 15% x $800M = $120M. Common trap: calculating each company's revenue separately then subtracting (which gives the same answer but wastes ~30 seconds). Just find the percentage difference first, then multiply once.
Tab 1 states: "All employees with tenure > 5 years are eligible for the senior bonus." Tab 2 states: "Maria has been with the company for 7 years." Which of the following can be correctly inferred?
Show Answer
Answer: B
The rule says employees with >5 years are ELIGIBLE. Maria has 7 years, so she IS eligible — B is a direct logical deduction. A goes too far (eligible does not mean she will receive it — other conditions might apply). C, D, and E all introduce information not present in either tab. MSR requires strict logical inference from stated facts only.
A portfolio has only stocks (S) and bonds (B). It has 80 total assets and the bond-to-stock ratio must equal 3:1. Select values for stocks and bonds that satisfy both constraints simultaneously.
| Option | Stocks (S) | Bonds (B) |
|---|---|---|
| A | 20 | 60 |
| B | 30 | 50 |
| C | 25 | 75 |
| D | 40 | 40 |
| E | 10 | 70 |
Show Answer
Answer: A (S=20, B=60)
Constraints: S + B = 80 AND B/S = 3. From ratio: B = 3S. Substituting: S + 3S = 80 → 4S = 80 → S = 20, B = 60. Check: 20+60=80 ✓, 60/20=3 ✓. Option C (25+75=100, not 80) fails the total. Option B (30+50=80 ✓ but 50/30≠3) fails the ratio. Only A works.
Argument: "Sales of umbrellas increased 40% in March. Therefore, it must have rained more than usual in March." Which piece of information would most help evaluate this argument?
Show Answer
Answer: D
The argument concludes rain increased from umbrella sales. To evaluate this, we need to know if rainfall actually did increase (D). D directly tests the conclusion. A and B identify alternative causes (which weaken rather than evaluate). C addresses distribution, not cause. E is related but indirect — raincoat data tells us about rain indirectly, while actual rainfall data (D) is the direct test. "Evaluate" questions want the most direct test of the argument's logic.
A table lists exam scores: 72, 85, 91, 68, 78, 95, 82, 77, 88, 63. What is the approximate interquartile range (IQR)?
Show Answer
Answer: B (17)
Sorted: 63, 68, 72, 77, 78, 82, 85, 88, 91, 95. Median = (78+82)/2 = 80. Q1 (lower half median) = (68+72)/2 = 70. Q3 (upper half median) = (88+91)/2 = 89.5. IQR = Q3 - Q1 = 89.5 - 70 = 19.5 ≈ roughly 17-20. Answer B (17) is closest to exact calculation depending on method. Full range = 95-63 = 32 (trap answer D for students who confuse IQR with range).
A scatter plot shows advertising spend (x-axis, $thousands) vs. sales (y-axis, $millions) for 20 companies. The points cluster tightly along a line from (10, 2) to (100, 20) with one outlier at (80, 3). What does the outlier suggest?
Show Answer
Answer: B
The outlier at (80, 3) spent $80K on advertising but only generated $3M in sales — far below the expected ~$16M based on the trend line. This suggests this specific company's advertising was ineffective OR it faced unique headwinds. A is too broad (the overall trend supports advertising, but not "always"). D is wrong (the overall correlation is clearly positive). C is also wrong — the highest spend is 100K. E is an absurd generalization from one data point.
Tab 1: "Product returns increased by 15% year-over-year." Tab 2: "Quality control failures decreased by 8% year-over-year." Tab 3: "Customer complaints about shipping increased by 22%." What is the most likely explanation for the increase in returns?
Show Answer
Answer: C
The data shows: quality failures DOWN 8% (rules out A), but shipping complaints UP 22%. The return increase (+15%) coincides with shipping complaints (+22%) but not quality issues (which improved). This makes shipping the most supported explanation. B and D and E introduce information not in any tab. MSR requires selecting the explanation best supported by combining information across all tabs — don't invent new hypotheses.
A factory produces widgets and gadgets. Each widget takes 3 hours and each gadget takes 5 hours. The factory has exactly 45 hours per week. The manager wants to produce the same number of both products. Which combination of widgets (W) and gadgets (G) satisfies both constraints?
| Option | W | G |
|---|---|---|
| A | 5 | 5 |
| B | 6 | 6 |
| C | 9 | 4 |
| D | 4 | 4 |
| E | 3 | 3 |
Show Answer
Answer: A (W=5, G=5)
Constraints: W = G (equal quantities) AND 3W + 5G = 45. If W = G = n: 3n + 5n = 8n = 45 → n = 5.625. But n must be an integer, so we test n=5: 3(5)+5(5)=15+25=40 (under 45) and n=6: 3(6)+5(6)=18+30=48 (over 45). A (5,5) is the closest feasible solution that keeps W=G without exceeding 45 hours. This tests careful reading: "exactly 45 hours" and "same number" are two binding constraints.
Claim: "City X's new subway reduced average commute times by 18 minutes." To evaluate whether the subway caused this reduction, which factor is MOST important to consider?
Show Answer
Answer: A
To establish causation (not just correlation), the key question is: would commute times have fallen anyway? If comparable cities without subways also saw 18-minute reductions, the subway may not be the cause (a confounding variable like remote work trends could explain it). A directly tests this counterfactual. B (cost) and D (approval) are irrelevant to causation. C (ridership) tells us usage but not causation. E (road age) is tangential. Evaluate questions on the GMAT almost always have the best answer as one that tests the causal claim via comparison or control group logic.
A store's monthly sales data (in units): Jan=420, Feb=380, Mar=450, Apr=410, May=490, Jun=520. What is the average (mean) monthly sales rounded to the nearest whole unit?
Show Answer
Answer: A (445)
Sum = 420+380+450+410+490+520 = 2,670. Mean = 2,670 / 6 = 445 exactly. Speed tip: Notice the values cluster near 450. Deviations from 450: -30, -70, 0, -40, +40, +70. Sum of deviations = -30. Mean = 450 + (-30/6) = 450 - 5 = 445. Using this deviation method is faster than summing all six numbers from scratch.
A line graph shows two products' sales indexed to 100 in Year 1. By Year 5, Product A is at index 160 and Product B is at index 140. If Product A had actual sales of $2M in Year 1, what were Product A's approximate actual sales in Year 5?
Show Answer
Answer: B ($3.2M)
Index of 160 means Product A grew to 160% of its Year 1 value. Year 5 sales = $2M × 1.60 = $3.2M. Common trap: students see "index 160" and think it grew BY 160% (which would give $5.2M). An index of 160 means it IS 160% of the base — a 60% increase, not 160%. Always convert index to growth rate: Index 160 = 60% growth from base. Product B data (index 140) is a distractor — the question only asks about Product A.
Tab 1: "Our survey shows 72% of customers prefer Product X over Product Y." Tab 2: "The survey used 50 participants, all recruited from Product X's official fan forum." Which statement best describes the survey's limitation?
Show Answer
Answer: B
The most critical limitation is sampling bias: all 50 participants came from Product X's fan forum — a group already predisposed to prefer Product X. This makes the 72% result unrepresentative of the general population. A (sample size) is a real concern but is secondary — even 5,000 biased respondents from the same forum would give an equally skewed result. The sampling method is the fundamental flaw, not the size. C, D, and E are all irrelevant or unsubstantiated.
A conference room seats people in rows. Each row has the same number of seats. The room must seat exactly 48 people, with more than 4 rows but fewer than 10 rows. Select from the table a possible number of rows and a possible number of seats per row.
| Option | Rows | Seats per Row |
|---|---|---|
| A | 4 | 12 |
| B | 6 | 8 |
| C | 8 | 7 |
| D | 10 | 5 |
| E | 5 | 9 |
Show Answer
Answer: B (6 rows, 8 seats per row)
Constraints: Rows × Seats = 48. Rows must be >4 and <10 (so rows can be 5,6,7,8,9). Check each: A=4 rows (fails, must be >4). B=6×8=48 ✓ and 4<6<10 ✓. C=8×7=56 ≠ 48 ✗. D=10 rows (fails, must be <10). E=5×9=45 ≠ 48 ✗. Only B satisfies all three constraints simultaneously.
Study conclusion: "Students who ate breakfast performed better on math tests, with average scores of 82 vs. 71 for those who skipped breakfast. Therefore, eating breakfast improves mathematical ability." Which of the following would most effectively challenge this conclusion?
Show Answer
Answer: B
The conclusion assumes breakfast CAUSES better math performance, but B identifies a confounding variable: wealth predicts both breakfast-eating AND access to tutoring (which could explain the higher scores). This is a classic alternative explanation that breaks the causal chain. A (age diversity) doesn't directly challenge the breakfast-math link. C (test timing) could actually strengthen the argument if anything. D (breakfast quality) complicates the picture but doesn't directly refute the causal claim. E merely shows overlap in distributions, which we'd expect — it doesn't challenge the average effect. B is the strongest challenge because it offers a complete alternative causal explanation.