GMAT Data Insights lesson · Trends
GMAT Data Insights: trends and rates of change
Line-chart trends on GMAT Data Insights: counting rises, average change per period, crossing lines, projections and comparing growth.
The core idea
Compare each point with the one before it, divide total change by the number of intervals (not points), and project from the last value at the average rate.
Core ideas
Mark each segment of the line up or down. Most trend questions are then a count, a subtraction or one division.
A line chart with n points has n − 1 segments. Each segment is one period-to-period change. Keep the two words apart:
- Points are the values you read (one per year).
- Intervals are the changes between neighbouring points (one fewer than the points).
Here is a small line of yearly orders (in thousands), with each change written in:
| Interval | 2020→21 | 2021→22 | 2022→23 | 2023→24 |
|---|---|---|---|---|
| Change | +15 | −5 | +20 | +10 |
- Rises: 3 of the 4 intervals.
- Total change: 80 − 40 = 40.
- Average change per year: 40 ÷ 4 = 10.
The method:
- Read every point in the series the question names.
- Write the change for each interval (new − old), with its sign.
- Count rises, find the largest change, or take last − first.
- For an average, divide by the number of intervals.
Counting rises
Why it works. "Higher than the previous year" is a statement about one interval: the step from one point to the next. A chart with n points has n − 1 intervals, and each gives one yes or no. So you count intervals, never points.
How to spot it: "In how many years was X higher than in the previous year?"
Example. Museum visits, 2019 to 2025. In how many years were visits higher than in the year before?
- Seven points, so six changes: 2020, 2021, 2022, 2023, 2024, 2025.
- Compare each year with the year before: up, up, up, down (2,600 → 2,400), up, up.
- Rises in 5 of the 6 years.
Trap: comparing each year with 2019 instead of with the year before. Every year after 2019 is above 1,000, so that count is 6. "Higher than the previous year" always means the neighbour on the left.
Try it. On the museum chart, in how many years did visits rise by more than 300?
Show the answer
- Write each year's change: +800, +400, +400, −200, +600, +200.
- Tick the ones above +300: 2020 (+800), 2021 (+400), 2022 (+400) and 2024 (+600).
- The answer is 4 years. The +200 in 2025 is a rise, but not by more than 300.
Edge cases.
- A flat year is not a rise. If the value is unchanged, "higher than the previous year" leaves it out; "at least as high" keeps it.
- The first year never counts. It has no previous year to compare with.
- "In consecutive years" asks for a run of rises in a row. Find the longest run of up steps, not the total number of them.
GMAT-level example. Yearly riders on a city's buses and trams, in millions. In how many years did exactly one of the two rise?
| Year | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|
| Bus | +5 | −5 | +10 | −5 | +10 |
| Tram | +5 | +5 | −5 | +10 | +5 |
| Exactly one rose? | no (both) | ✓ | ✓ | ✓ | no (both) |
- Six points per line give five changes. Write each one with its sign, as in the table.
- Mark each year: both rose, exactly one rose, or neither rose.
- Exactly one rose in 2021, 2022 and 2023: 3 years.
- Check: 3 (exactly one) + 2 (both) + 0 (neither) = 5 intervals ✓.
The trap answers are 4 (Tram's rises alone), 2 (the years when both rose) and 7 (every rise on the chart added up).
Check yourself. A series reads 12, 15, 15, 14, 18 over five years. In how many years was it higher than in the year before?
Check your answer
Changes: +3, 0, −1, +4. Only 2 are rises; the 0 is not.
Average change
Why it works. Add up the yearly changes and every middle value cancels: (b − a) + (c − b) + (d − c) = d − a. So the total change depends only on the first and last points. The average spreads that total evenly over the number of intervals.
How to spot it: "the average annual increase", "on average, by how much per year".
Example. A town's bike lanes grew from 50 km in 2019 to 110 km in 2023. What was the average increase per year?
- Total change: 110 − 50 = 60 km.
- 2019 to 2023 has 5 points and 4 intervals.
- 60 ÷ 4 = 15 km per year.
The middle points don't matter: the yearly changes (+20, −5, +20, +25) add up to the same 60.
Trap: dividing by 5 (the number of points) gives 12, the fencepost trap. Quoting one year's change (+20 for the first year, +25 for the last) is the other wrong option.
Try it. On the bike-lane chart, what was the average yearly increase from 2020 to 2023?
Show the answer
- Total change: 110 − 70 = 40 km.
- 2020 to 2023 has 4 points and 3 intervals.
- 40 ÷ 3 ≈ 13.3 km per year.
- Dividing by 4 gives 10, the fencepost trap again.
Edge cases.
- Uneven gaps between points. If the points are years apart by different amounts, divide the total change by the number of years, not the number of intervals.
- A negative average is a fall. From 80 to 68 over 3 intervals is an average change of −4 per year.
- Average change is not average percent change. +10 a year on a base of 50 is 20% the first year but less every year after.
GMAT-level example. A food co-op's members, in thousands. The data points are not evenly spaced in time. What was the average yearly increase from 2010 to 2024, and in which period did members grow fastest per year?
| Period | 2010–14 | 2014–18 | 2018–20 | 2020–24 |
|---|---|---|---|---|
| Rise (thousand) | 12 | 8 | 8 | 14 |
| Years | 4 | 4 | 2 | 4 |
| Rise per year | 3 | 2 | 4 | 3.5 |
- Total change: 62 − 20 = 42 thousand.
- Years: 2024 − 2010 = 14. (There are only 4 intervals, but they cover 14 years.)
- Average: 42 ÷ 14 = 3 thousand members per year.
- For each period, divide its rise by its own number of years, as in the table.
- The fastest is 2018 to 2020: 8 ÷ 2 = 4 thousand per year.
- Check: 4 × 3 + 4 × 2 + 2 × 4 + 4 × 3.5 = 12 + 8 + 8 + 14 = 42 ✓.
The traps are 42 ÷ 4 = 10.5 (intervals instead of years), 42 ÷ 5 = 8.4 (points), and 2020 to 2024, which has the biggest rise (14) but takes twice as long.
Check yourself. A value is 80 in 2021, 74 in 2022, 71 in 2023 and 68 in 2024. What is the average yearly change?
Check your answer
(68 − 80) ÷ 3 = −12 ÷ 3 = −4 per year.
Crossing lines
Why it works. Two lines cross where the difference A − B changes sign. The chart only has values at its points, so "the first year A exceeded B" is the first point where A − B is positive. Writing the difference at every point turns the picture into a list you can scan.
How to spot it: "the first year in which A exceeded B", "the gap between A and B was greatest in".
Example. A retailer's sales by channel ($ million). In which year did Online sales first exceed In-store sales? In which year was the gap between them greatest?
| Year | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|---|
| In-store | 80 | 75 | 70 | 62 | 55 | 50 |
| Online | 30 | 45 | 60 | 68 | 80 | 95 |
| Gap | 50 | 30 | 10 | 6 | 25 | 45 |
- In 2021, Online (60) is still below In-store (70). In 2022, Online (68) is above In-store (62).
- The first year Online leads is 2022, the first point after the crossing.
- For the gap, subtract at every year: the largest is 2019 (50), when In-store led.
Trap: answering 2021, the last year before the crossing. For the gap, 2024 looks right because Online is at its highest there, but that gap is 45, smaller than 50. The widest gap can be on either side of a crossing.
Try it. On the retail chart, in which year were the two channels closest, and which channel led that year, by how much?
Show the answer
- Use the gap row: 50, 30, 10, 6, 25, 45.
- The smallest gap is 6, in 2022.
- In 2022, Online (68) is above In-store (62): Online led by $6 million.
Edge cases.
- A tie at a point. If A equals B in one year, "A first exceeded B" is a later year, because a tie is not "exceeded".
- Lines can cross more than once. "First" means the earliest crossing; "the last time B led" may be later.
- A crossing between points has no year. The retail lines cross somewhere between 2021 and 2022. You can name the first year Online led, not the year they crossed.
GMAT-level example. Monthly revenue at two cafés, in thousands of dollars. In which month did Market first lead Harbor? In how many months did Market lead, and in which month was the gap between them widest?
| Month | Jan | Feb | Mar | Apr | May | Jun | Jul |
|---|---|---|---|---|---|---|---|
| Harbor | 42 | 46 | 48 | 50 | 48 | 44 | 46 |
| Market | 36 | 42 | 50 | 52 | 46 | 46 | 50 |
| Market − Harbor | −6 | −4 | +2 | +2 | −2 | +2 | +4 |
- Write Market − Harbor for every month, with its sign, as in the table.
- The first positive difference is in March, so that's when Market first led.
- Count the positive differences: March, April, June and July, so 4 months.
- The widest gap is the largest difference ignoring the sign: 6, in January, when Harbor led.
- Check: 4 months with Market ahead + 3 with Harbor ahead = 7 months ✓.
Market's highest point is April (52), but the gap there is only 2. Counting the crossings gives 3, which is not the number of months Market led.
Check yourself. Over three years, A reads 10, 14, 18 and B reads 16, 15, 17. In which year did A first exceed B, and what was the gap in year 2?
Check your answer
A first exceeds B in year 3 (18 > 17). In year 2 the gap is 15 − 14 = 1, with B ahead.
Projections
Why it works. Growing at the average rate means adding the same step every period. So the projected values climb in equal steps from the last data point, and after n periods the value is last + n × average. You can solve for n instead of listing every year.
How to spot it: "If X continues to grow at its average annual rate, in which year will it first reach…"
Example. Solar panels installed (thousands): 120 in 2020, 150, 170, 200, then 240 in 2024. If the number keeps growing by the average yearly increase from 2020 to 2024, in which year will it first reach 380 thousand?
- Average increase: (240 − 120) ÷ 4 = 30 per year.
- Start from the last value, 240, and add 30 each year: 270, 300, 330, 360, 390.
- 360 is still below 380; 390 is the first value at or above it. That's the 5th year after 2024: 2029.
Trap: using the last change (200 → 240, +40) instead of the average. Steps of 40 give 280, 320, 360, 400, so 380 is reached in 2028, a year too early. Projecting from the first value, 120, is the other slip.
Try it. Suppose instead the number keeps growing by its average yearly increase from 2022 to 2024. In which year does it first reach 380 thousand?
Show the answer
- Average increase from 2022 to 2024: (240 − 170) ÷ 2 = 35 per year.
- From 240, add 35 each year: 275, 310, 345, 380.
- 380 is reached exactly in the 4th year after 2024: 2028. "Reach" includes landing on 380.
Edge cases.
- "Reach" vs "exceed". With steps of 30 from 240, the value is 390 in 2029. It reaches 390 in 2029 but first exceeds 390 in 2030.
- The base period matters. "The average from 2022 to 2024" uses only those points, not the whole chart.
- Percent growth multiplies. "At its average percent growth rate" multiplies by the same factor each year, so the steps get bigger.
GMAT-level example. Subscribers to two plans, in thousands.
| Year | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|
| Basic | 200 | 215 | 225 | 240 | 260 |
| Premium | 100 | 130 | 170 | 210 | 240 |
If each plan keeps growing by its own average yearly increase from 2020 to 2024, in which year will Premium first have more subscribers than Basic?
- Basic's average: (260 − 200) ÷ 4 = 15 per year.
- Premium's average: (240 − 100) ÷ 4 = 35 per year.
- In 2024 Basic leads by 260 − 240 = 20. Premium closes the gap by 35 − 15 = 20 a year.
- After 1 year (2025) the gap is 0: both are at 275. That's a tie, not a lead.
- After 2 years (2026): Basic 290, Premium 310. Premium leads for the first time in 2026.
- Check: 260 + 2 × 15 = 290 and 240 + 2 × 35 = 310 ✓.
The traps are 2025 (the tie) and 2027, which comes from using the last changes (+20 and +30) instead of the averages: those give a tie at 300 in 2026.
Check yourself. A value was 50 in 2021 and 74 in 2024. If it keeps growing at its average yearly increase, what will it be in 2027?
Check your answer
Average: (74 − 50) ÷ 3 = 8 per year. 2027 is 3 years after 2024: 74 + 3 × 8 = 98.
Comparing growth
Why it works. Percent growth divides each rise by its own starting value. A series that starts small can grow by a bigger percent with a smaller rise. And since percent growth = last ÷ first − 1, the series with the bigger last ÷ first ratio always grew by the bigger percent.
How to spot it: "which series grew by the greater percent", "grew faster".
Example. A river port's cargo (thousand containers), 2021 to 2024. Which series grew by the greater percent?
- Exports: 150 → 240. Rise 90, and 90 ÷ 150 = 60%.
- Transshipment: 90 → 170. Rise 80, and 80 ÷ 90 ≈ 88.9%.
- Transshipment grew by the greater percent. (Equivalently, 170 ÷ 90 ≈ 1.89 beats 240 ÷ 150 = 1.6.)
| Series | Start | End | Rise | Rise ÷ start |
|---|---|---|---|---|
| Exports | 150 | 240 | 90 | 60% |
| Transshipment | 90 | 170 | 80 | ≈ 88.9% |
Trap: Exports rose by more containers (90 vs 80), so it looks like the faster grower. Comparing growth means dividing each rise by its own starting value.
Try it. On the port chart, which series grew by the greater percent from 2023 to 2024?
Show the answer
- Exports: 200 → 240. Rise 40, and 40 ÷ 200 = 20%.
- Transshipment: 140 → 170. Rise 30, and 30 ÷ 140 ≈ 21.4%.
- Transshipment, just. Exports again rose by more containers (40 vs 30).
Edge cases.
- Same rise, different starts: the smaller start always gives the bigger percent.
- Compare the same period. Growth from 2021 to 2024 for one series and from 2022 to 2024 for the other is not a fair comparison.
- Cross-multiply to skip division. 240 ÷ 150 vs 170 ÷ 90: compare 240 × 90 = 21,600 with 170 × 150 = 25,500. The second is bigger, so Transshipment's ratio is bigger.
GMAT-level example. An outdoor store's revenue by product line, in thousands of dollars. Which line grew by the greatest percent from 2022 to 2025? By how many percentage points did its growth beat the growth of the line with the largest dollar rise?
| Line | 2022 | 2025 | Rise | Rise ÷ 2022 |
|---|---|---|---|---|
| Tents | 48 | 66 | 18 | 37.5% |
| Boots | 120 | 150 | 30 | 25% |
| Packs | 30 | 45 | 15 | 50% |
| Stoves | 64 | 92 | 28 | 43.75% |
- For each line, find the rise and divide it by the 2022 value, as in the table.
- The greatest percent is Packs, 50%, even though its rise is the smallest (15).
- The largest dollar rise is Boots (+30), which grew 25%.
- Difference: 50% − 25% = 25 percentage points.
- Check with ratios: 45 ÷ 30 = 1.5 and 150 ÷ 120 = 1.25, so 50% and 25% ✓.
Choosing Boots as the fastest grower is the trap, and so is answering "25%": the gap between two percents is in percentage points.
Check yourself. Series A rises from 20 to 30. Series B rises from 200 to 260. Which grew by the greater percent?
Check your answer
A: 10 ÷ 20 = 50%. B: 60 ÷ 200 = 30%. A grew by the greater percent, though B rose by more.
How hard questions disguise it
For harder questions (L4–L5)
At L4 and L5, a trend question rarely says "trend". "Grew fastest" can mean a percent change, not a rise, and "in how many years did both rise" is a counting question on the changes, not on the values. Hard charts space their points unevenly in time, put two lines close together so a crossing is easy to miss, or ask whether two series tend to move together, which is a correlation question. Each one starts the same way: read every point, write the change for each interval with its sign, and only then count, divide or project.
Practice
Practice trend questions on line charts in timed sets on the trends practice page.
Common traps
Points vs intervals
Five years of data have four year-to-year changes. Average change = (last − first) ÷ 4.
Comparing with the wrong point
'Higher than the previous year' compares neighbours, not each year with the first.
The wrong period
The first year A leads is right after the crossing; the steepest rise isn't always the step that ends at the peak.
Rises vs growth rates
Comparing growth means comparing last ÷ first for each line, not the rises.