Free Mathematics Class 10 CBSE notes · practise this chapter with an AI quiz

← All study notes

How a Straight Line Between Two Known Values Estimates the One in Between

Estimate an unmeasured value with linear interpolation, see why the median of grouped data is interpolation on the ogive, find quartiles and percentiles the same way, and know when a straight-line estimate breaks down.

How can you estimate a value that nobody measured?

Draw a straight line between the two known values on either side and read the answer off that line. That is linear interpolation, and you have already used it without the name.

A thermometer read 24∘C24^\circ\text{C} at 10:0010{:}00 and 28∘C28^\circ\text{C} at 11:0011{:}00. Nobody looked at 10:3010{:}30, yet almost everyone would say it was about 26∘C26^\circ\text{C} — halfway in time, so halfway in temperature. That guess assumes the temperature rose steadily between the two readings, which is exactly what a straight line means.

The same idea sits inside a formula you are asked to use in Class 10: the median of grouped data is linear interpolation inside the median class. Once you see that, the formula stops being something to memorise and becomes a sentence you can rebuild.

This page covers the interpolation formula, the median, quartiles and percentiles as interpolation on the ogive, when a straight-line estimate goes wrong, and the difference between interpolating and extrapolating.
Formula

What is the linear interpolation formula?

For two known points (x1,y1)(x_1, y_1) and (x2,y2)(x_2, y_2), the value at any xx between them is estimated by

y=y1+x−x1x2−x1 (y2−y1)\boxed{y = y_1 + \frac{x - x_1}{x_2 - x_1}\,(y_2 - y_1)}

Read it as a fraction of the way across. x−x1x2−x1\dfrac{x - x_1}{x_2 - x_1} is how far along the gap you are, from 00 at x1x_1 to 11 at x2x_2. Multiply the rise y2−y1y_2 - y_1 by that fraction and add it to the starting value y1y_1.

Where it comes from. It is the equation of the straight line through the two points: the slope is y2−y1x2−x1\dfrac{y_2 - y_1}{x_2 - x_1}, and moving x−x1x - x_1 along the line raises yy by slope times that distance.

Worked example 1 — the thermometer. With (10,24)(10, 24) and (11,28)(11, 28), at x=10.5x = 10.5:

y=24+10.5−1011−10(28−24)=24+0.5×4=26∘Cy = 24 + \frac{10.5 - 10}{11 - 10}(28 - 24) = 24 + 0.5 \times 4 = 26^\circ\text{C}

Worked example 2 — a city's population. A town had 4040 lakh people in 20112011 and 5252 lakh in 20212021. For 20152015:

y=40+2015−20112021−2011(52−40)=40+0.4×12=44.8 lakhy = 40 + \frac{2015 - 2011}{2021 - 2011}(52 - 40) = 40 + 0.4 \times 12 = 44.8 \text{ lakh}

Check the answer sits between the two known values. 44.844.8 lies between 4040 and 5252, and 20152015 is a little under halfway, so the estimate is a little under halfway up. An interpolated value outside the two known values means a slip in the fraction.

Why is the median formula for grouped data really interpolation?

Because it draws a straight line across the median class on the cumulative frequency graph and reads off where that line reaches n2\frac{n}{2}.

Take the marks of 5555 students:

| Marks | 00–1010 | 1010–2020 | 2020–3030 | 3030–4040 | 4040–5050 |
|---|---|---|---|---|---|
| Frequency | 55 | 88 | 2020 | 1515 | 77 |
| Cumulative | 55 | 1313 | 3333 | 4848 | 5555 |

Plot the cumulative frequency at each upper limit — (10,5)(10, 5), (20,13)(20, 13), (30,33)(30, 33), (40,48)(40, 48), (50,55)(50, 55) — and join the points. This is the ogive. The median is the mark of the 552=27.5\frac{55}{2} = 27.5th student, so we want the xx where the ogive reaches 27.527.5.

27.527.5 lies between 1313 and 3333, so interpolate between (20,13)(20, 13) and (30,33)(30, 33), with the roles swapped: we know the height and want the position.

x=20+27.5−1333−13×(30−20)=20+14.520×10=27.25x = 20 + \frac{27.5 - 13}{33 - 13} \times (30 - 20) = 20 + \frac{14.5}{20} \times 10 = 27.25

Now compare with the Class 10 formula, Median=l+(n2−cff)×h\text{Median} = l + \left(\dfrac{\frac{n}{2} - cf}{f}\right) \times h:

- l=20l = 20 is x1x_1, the start of the gap
- cf=13cf = 13 is y1y_1, the height at the start
- f=20f = 20 is y2−y1=33−13y_2 - y_1 = 33 - 13, the rise across the class
- h=10h = 10 is x2−x1x_2 - x_1, the width of the gap

It is the interpolation formula, term for term. The straight line is the assumption that the 2020 students in the 2020–3030 class are spread evenly across it — which is why the formula needs cfcf from the class before: that is where the line starts.

How do you find quartiles and percentiles the same way?

Replace n2\frac{n}{2} by the fraction of nn you want and interpolate across whichever class the ogive crosses that height in. The method does not change.

Lower quartile Q1Q_1
is the n4=13.75\frac{n}{4} = 13.75th value. The cumulative column first passes 13.7513.75 in the 2020–3030 class, between 1313 and 3333:

Q1=20+13.75−1320×10=20+0.375=20.375Q_1 = 20 + \frac{13.75 - 13}{20} \times 10 = 20 + 0.375 = 20.375

Upper quartile Q3Q_3 is the 3n4=41.25\frac{3n}{4} = 41.25th value. That height is first passed in the 3030–4040 class, between 3333 and 4848, where f=15f = 15:

Q3=30+41.25−3315×10=30+5.5=35.5Q_3 = 30 + \frac{41.25 - 33}{15} \times 10 = 30 + 5.5 = 35.5

The 9090th percentile is the 0.9×55=49.50.9 \times 55 = 49.5th value, in the 4040–5050 class between 4848 and 5555, where f=7f = 7:

P90=40+49.5−487×10=40+2.14=42.14P_{90} = 40 + \frac{49.5 - 48}{7} \times 10 = 40 + 2.14 = 42.14

So the middle half of the class scored between 20.37520.375 and 35.535.5, an interquartile range of 15.12515.125 marks, and nine students in ten scored below about 4242.

One check covers every case. Each answer must lie inside the class it was interpolated in — Q1Q_1 in 2020–3030, Q3Q_3 in 3030–4040, P90P_{90} in 4040–5050 — and Q1≤Median≤Q3Q_1 \le \text{Median} \le Q_3, which here reads 20.375≤27.25≤35.520.375 \le 27.25 \le 35.5.

When does interpolation go wrong, and what is extrapolation?

A straight line is only an estimate of what happens between two points, so it is wrong whenever the true relationship bends — and it becomes unreliable once you leave the two points behind.

Error from a bending curve. Estimate 20\sqrt{20} from 16=4\sqrt{16} = 4 and 25=5\sqrt{25} = 5:

20≈4+20−1625−16×1=4+49=4.444\sqrt{20} \approx 4 + \frac{20 - 16}{25 - 16} \times 1 = 4 + \frac{4}{9} = 4.444

The true value is 4.4724.472. The straight line runs below the square-root curve, which bends, so the estimate is too small by about 0.030.03. The closer the two known points, the smaller this error: from 19.36=4.4\sqrt{19.36} = 4.4 and 20.25=4.5\sqrt{20.25} = 4.5 the same method gives 4.4724.472, correct to three decimal places.

Extrapolation means using the same line outside the two known points. From the town's 4040 lakh in 20112011 and 5252 lakh in 20212021, a straight line predicts

40+2041−201110×12=76 lakh in 204140 + \frac{2041 - 2011}{10} \times 12 = 76 \text{ lakh in } 2041

That number assumes the town keeps adding exactly 1.21.2 lakh people a year for twenty more years — an assumption the data never tested. Inside the gap, the two readings bracket the answer; outside it, nothing does.

In grouped data the straight line is the assumption that values are spread evenly within each class. When a class is lopsided — most students near the top of 2020–3030, say — the true median differs from 27.2527.25, and only the raw marks could show by how much.
Exam tip

What earns full marks when you interpolate?

Name the two known points, write the fraction of the gap, then substitute — and check that the answer lies between the known values.

- State the two bracketing points explicitly: "27.527.5 lies between the cumulative frequencies 1313 and 3333, at marks 2020 and 3030"
- Write the fraction x−x1x2−x1\dfrac{x - x_1}{x_2 - x_1} (or n2−cff\dfrac{\frac{n}{2} - cf}{f}) before multiplying
- Take cfcf from the class before the one you interpolate in; it is the starting height of the line
-
Use the plain frequency ff as the rise, never a cumulative one
-
Keep the units on the answer — marks, °C, lakh
-
Check the answer lies inside the gap
: a median outside its median class means cfcf was misread
- For quartiles and percentiles, locate the class from the cumulative column first; it is often not the median class
- Say "estimate": interpolation assumes a straight line, and the answer is only as good as that assumption

The misconception to name. Averaging the two neighbours, y1+y22\dfrac{y_1 + y_2}{2}, is interpolation only at the exact midpoint. At 20152015 it would give 4646 lakh instead of 44.844.8, because 20152015 is 0.40.4 of the way across, not 0.50.5.
Did you know

How did people read logarithm tables before calculators?

By interpolating — and the "mean difference" columns printed in every log table are linear interpolation done in advance.

A four-figure log table lists log⁡2.0=0.3010\log 2.0 = 0.3010 and log⁡2.1=0.3222\log 2.1 = 0.3222. To find log⁡2.04\log 2.04, which is 0.40.4 of the way from 2.02.0 to 2.12.1:

log⁡2.04≈0.3010+0.4×(0.3222−0.3010)=0.3010+0.4×0.0212=0.3095\log 2.04 \approx 0.3010 + 0.4 \times (0.3222 - 0.3010) = 0.3010 + 0.4 \times 0.0212 = 0.3095

The true value is 0.30960.3096. The mean-difference column simply lists 0.4×0.02120.4 \times 0.0212 and its cousins, so that the reader adds a printed correction instead of multiplying.

It works because the table's steps are small.
Over a gap from 2.02.0 to 2.12.1 the logarithm curve is very nearly straight, so the line through two neighbouring entries barely leaves it — the same reason 20\sqrt{20} became accurate once the known points were close together.

Generations of engineers, navigators and astronomers did their arithmetic this way, and the tables at the back of some school textbooks still carry those columns.
Exam relevance

How is interpolation used in board exams, JEE and NEET?

In Class 10 interpolation is examined through the median formula and the ogive, and the same idea returns in every later year under different names.

Board papers. Expect the median of grouped data by formula, reading the median from an ogive (draw a horizontal line at n2\frac{n}{2} and drop down to the marks axis — that is interpolation done with a ruler), and finding a missing frequency when the median is given. The ogive method and the formula must agree, and examiners sometimes ask for both.

Class 11 and JEE Main. Quartiles, deciles and percentiles of grouped data use the same interpolation with n4\frac{n}{4}, kn10\frac{kn}{10} or kn100\frac{kn}{100}. Mean deviation about the median starts from the interpolated median, and straight-line equations in coordinate geometry are the interpolation formula written for every xx at once.

Physics practicals and NEET. Reading a value between plotted points on a graph — a current between two measured voltages, a temperature between two times — is interpolation, and drawing a best-fit line through scattered readings and reading off it extends the same idea. Questions that ask for a value outside the measured range are testing whether you recognise extrapolation.

Where it leads. Calculus replaces the straight line between two points by the tangent at one point, and numerical methods use curves through three or more points when a line is too crude.

The trap that costs marks. Taking cfcf as the median class's own cumulative frequency. It must be the previous class's, and the slip shows up immediately: the median lands below its own class — 17.2517.25 for the class 2020–3030 in this data.
Key takeaways

What must you be able to do from this part?

One straight line, read in two directions.

- Linear interpolation: y=y1+x−x1x2−x1(y2−y1)y = y_1 + \dfrac{x - x_1}{x_2 - x_1}(y_2 - y_1) — the starting value plus the fraction of the gap times the rise
- Thermometer: 24∘C24^\circ\text{C} at 10:0010{:}00 and 28∘C28^\circ\text{C} at 11:0011{:}00 give 26∘C26^\circ\text{C} at 10:3010{:}30
- Town: 4040 lakh in 20112011 and 5252 lakh in 20212021 give 44.844.8 lakh in 20152015
- The ogive plots cumulative frequency at each upper limit; for the 5555 marks the points are (10,5)(10, 5), (20,13)(20, 13), (30,33)(30, 33), (40,48)(40, 48), (50,55)(50, 55)
- The median is interpolation at height n2=27.5\frac{n}{2} = 27.5: 20+14.520×10=27.2520 + \frac{14.5}{20} \times 10 = 27.25
- The median formula matches term for term: l=x1l = x_1, cf=y1cf = y_1, f=y2−y1f = y_2 - y_1, h=x2−x1h = x_2 - x_1
- Quartiles and percentiles use n4\frac{n}{4}, 3n4\frac{3n}{4} or kn100\frac{kn}{100}: Q1=20.375Q_1 = 20.375, Q3=35.5Q_3 = 35.5, P90=42.14P_{90} = 42.14
- Every answer lies inside its own class, and Q1≤Median≤Q3Q_1 \le \text{Median} \le Q_3
- Interpolation errs when the curve bends: 20≈4.444\sqrt{20} \approx 4.444 against a true 4.4724.472, and closer points shrink the error
- Extrapolation leaves the known points behind: 7676 lakh in 20412041 assumes growth never changes
- Averaging two neighbours is interpolation only at the exact midpoint

The self-test: from the 5555-mark table, find the 6060th percentile both from the formula and from an ogive you draw yourself, and explain in one sentence what assumption both methods share.

Ready to put this into practice?

Create a personalized quiz on this exact topic — free to start.

Create your own quiz on Statistics — Part 2Create a free account
← Back to all articles