Visual Tools
Calculators
Tables
Mathematical Keyboard
Converters
Other Tools

How to Find the Median and Mode of a Distribution

The median splits the probability in half: it is where the CDF reaches 12\frac{1}{2}. The mode is the most likely value, the highest point of the pmf or pdf. Both describe the center, alongside the mean.

Recognize the task

The question asks for the "middle" value, the 50th percentile, the most likely or most common value, or how the mean, median and mode compare (skew). Typical inputs: a pmf table, a density on an interval, an exponential waiting time. Test: for the median, can you get the CDF? For the mode, can you see or compute where the pmf or pdf is largest?
The two measures are defined in the median lesson and the mode lesson.

Steps

StepOn f(x)=2xf(x) = 2x on [0,1][0, 1]
1Build the CDF. Running total of the pmf (CDF from a PMF) or integral of the density (CDF from a PDF).F(x)=x2F(x) = x^2 on [0,1][0, 1]
2Find where the CDF reaches one half. Continuous: solve F(m)=12F(m) = \frac{1}{2}. Discrete: take the smallest value with F(x)≥12F(x) \geq \frac{1}{2}; if FF equals 12\frac{1}{2} exactly on a flat stretch, every point of that stretch is a median.m2=12m^2 = \frac{1}{2}, so m=12≈0.7071m = \frac{1}{\sqrt{2}} \approx 0.7071
3Find the peak of the pmf or pdf. Discrete: the value with the largest probability. Continuous: set f′(x)=0f'(x) = 0 and also check the endpoints of the support.f(x)=2xf(x) = 2x rises on [0,1][0, 1]: mode =1= 1 (endpoint)
4Compare with the mean. Mean below median below mode suggests a left skew; the reverse order suggests a right skew.mean 23<\frac{2}{3} < median 0.7071<0.7071 < mode 11: left-skewed

Worked examples

Example 1 · a pmf table

discrete
p=0.1,0.3,0.4,0.2p = 0.1, 0.3, 0.4, 0.2 on 0,1,2,30, 1, 2, 3
F(1)=0.4<0.5, F(2)=0.8≥0.5F(1) = 0.4 < 0.5,\ F(2) = 0.8 \geq 0.5
first value past one half
median =2= 2, mode =2= 2
p(2)=0.4p(2) = 0.4 is largest

Example 2 · an exponential waiting time

solve F = 1/2
F(x)=1−e−0.5xF(x) = 1 - e^{-0.5x}, x≥0x \geq 0
rate λ=0.5\lambda = 0.5
e−0.5m=12⇒m=2ln⁡2e^{-0.5m} = \frac{1}{2} \Rightarrow m = 2\ln 2
take logs
median ≈1.3863\approx 1.3863, mode =0= 0
density largest at x=0x = 0

Example 3 · a fair die

ties
F(3)=12F(3) = \frac{1}{2} exactly
CDF flat at one half on [3,4)[3, 4)
3≤m≤43 \leq m \leq 4: P(X≤m)≥12P(X \leq m) \geq \frac{1}{2} and P(X≥m)≥12P(X \geq m) \geq \frac{1}{2}
both median conditions hold
every m∈[3,4]m \in [3, 4] is a median; all six faces are modes
often reported as 3.53.5; no single mode
See where the CDF crosses one half in the CDF visualizer.

See it happen

The median is a balance of area, not of weight: half the probability lies on each side. The mode is simply the highest point of the graph. The mean is a balance of moments, so a long tail drags it out while the median barely moves. For f(x)=2xf(x) = 2x most of the area is near 11, so the median (0.70710.7071) sits right of the mean (23\frac{2}{3}) and the peak is at the right end.

Common mistakes

Looking for an exact half on a discrete CDF

✗ F(1)=0.4F(1) = 0.4, F(2)=0.8F(2) = 0.8, so the median is 1.51.5
✓ take the first value with F(x)≥12F(x) \geq \frac{1}{2}: median =2= 2

Taking the mode as the value of the pdf

✗ f(x)=2xf(x) = 2x peaks at height 22, so the mode is 22
✓ the mode is the location of the peak: x=1x = 1

Missing an endpoint maximum

✗ f′(x)=2≠0f'(x) = 2 \neq 0, so f(x)=2xf(x) = 2x has no mode
✓ check the ends of the support: the maximum is at x=1x = 1