How To Find The Median From A Frequency Table

How to Find the Median from a Frequency Table
Ever stared at a table of numbers and wondered where the middle of the data really sits? That’s the median, the value that splits the set in half. It’s a staple in statistics, but the trick is getting it right when you’re only given a frequency table. Let’s walk through the process step by step, clear out common pitfalls, and arm you with tricks that actually work.

What Is a Frequency Table?

A frequency table is a tidy way to show how often each value or group of values occurs in a data set. Instead of listing every single observation, you group them into classes* (like age ranges, income brackets, test scores) and record how many observations fall into each class. The table usually looks something like this:

Class	Frequency
10–19	4
20–29	7
30–39	3
…	…

The numbers in the Frequency* column are the counts. When you add them up, you get the total number of observations, n.

Why It Matters / Why People Care

Knowing the median is useful because it’s less sensitive to extreme values than the mean. If you’re looking at exam scores, for instance, a single outlier won’t pull the average up or down. The median tells you the “typical” performance. In business, median sales can reveal a realistic picture of a product’s success without being skewed by a handful of huge deals.

When you only have a frequency table, you can’t just eyeball the middle value. You need a systematic method to locate the class that contains the median and then calculate the exact value within that class. That’s where the cumulative frequency and a bit of interpolation come in.

How It Works (or How to Do It)

1. Compute the Total Frequency

Add up all the frequencies to get n, the total number of observations. If the table is small, you can do it mentally; for larger tables, a quick spreadsheet or calculator helps.

2. Find the Median Position

For an odd n, the median sits at position ((n+1)/2). On the flip side, for an even n, it’s the average of the two middle positions, so you’ll need the values at positions (n/2) and ((n/2)+1). In practice, you’ll locate the class that contains either of those positions (or both, if the median falls between classes).

3. Build the Cumulative Frequency Table

Create a new column that adds each class’s frequency to all previous frequencies. The cumulative frequency tells you how many observations fall below or at the upper bound of each class.

Class	Frequency	Cumulative Frequency
10–19	4	4
20–29	7	11
30–39	3	14
…	…	…

4. Locate the Median Class

Find the first class where the cumulative frequency is greater than or equal to the median position. Which means that class contains the median. For an even n, you’ll need to find the two classes that contain the two middle positions; if they’re the same class, you’ll use that one for both.

5. Interpolate Within the Class (if needed)

If the class interval is a single value (e.g.Also, , 30), the median is that value. If the class spans a range, you’ll use linear interpolation to estimate the exact median within that range.

[ \text{Median} = L + \left(\frac{M - F}{f}\right) \times w ]

Where:

(L) = lower boundary of the median class
(M) = median position (or average of the two positions for even n)
(F) = cumulative frequency of the class before the median class
(f) = frequency of the median class
(w) = class width (upper bound – lower bound)

Let’s walk through a concrete example.

Example: Finding the Median of Exam Scores

Suppose you have the following frequency table for a class of 30 students:

Score Range	Frequency
0–10	1
11–20	3
21–30	5
31–40	9
41–50	8
51–60	3
61–70	1

Step 1: Total Frequency
(n = 1+3+5+9+8+3+1 = 30)

Step 2: Median Position
(n) is even, so we need the 15th and 16th observations. The median will be the average of the values at those positions.

Step 3: Cumulative Frequencies

Score Range	Frequency	Cumulative
0–10	1	1
11–20	3	4
21–30	5	9
31–40	9	18
41–50	8	26
51–60	3	29
61–70	1	30

Step 4: Locate Median Class
The 15th observation falls in the 31–40 range (cumulative 18 ≥ 15). The 16th observation also falls in the same range. So the median class is 31–40.

Step 5: Interpolate
Because both middle positions are in the same class, we can treat the median as the midpoint of that class, or we can do a precise interpolation:

(L = 31)
(M = 15.5) (average of 15 and 16)
(F = 9) (cumulative before the class)
(f = 9) (frequency of the class)
(w = 10) (40–31)

[ \text{Median} = 31 + \left(\frac{15.5 - 9}{9}\right) \times 10 \approx 31 + 0.722 \times 10 \approx 38.

If you found this helpful, you might also enjoy is sodium a metal or nonmetal or the shape of the water molecule h2o is.

So the median score is about 38.Plus, 2. If you prefer a simpler approach, you could say the median is roughly 38, the middle of the 31–40 range.

Common Mistakes / What Most People Get Wrong

Skipping the cumulative column
Some folks jump straight to the class with the largest frequency, thinking it must contain the median. That’s a trap;

the median depends on position*, not frequency magnitude. A class with a huge frequency at the very top or bottom of the distribution won't contain the middle observation if the bulk of the data lies elsewhere. Always build the cumulative frequency column first.

Using class limits instead of class boundaries
If your classes are listed as 0–10, 11–20, 21–30, the lower boundary of the median class (31–40) is 30.5, not 31. The upper boundary is 40.5. The class width (w) remains 10 (40.5 – 30.5), but using (L = 30.5) in the formula shifts the result slightly. For discrete data (like integer scores), the "true" boundaries are halfway between the upper limit of one class and the lower limit of the next. Ignoring this introduces a systematic bias of half a unit.
Confusing the median position with the median value**
(M = (n+1)/2) (or the average of (n/2) and (n/2 + 1)) gives you the rank* of the observation you are hunting for. It is not the score itself. Plugging the raw frequency or the class midpoint into the formula in place of (M) is a category error.
Forgetting to handle open-ended classes
Tables often end with "60+" or "Less than 10." If the median falls inside an open-ended class, you cannot calculate a precise interpolated median because (w) (class width) is undefined. In this case, you can only report that the median lies within* that class (e.g., "Median > 60") or make an external assumption about the width based on context.
Applying the formula to nominal or ordinal data
The interpolation formula assumes the underlying variable is continuous (or at least interval/ratio) and that observations are evenly spread across the class interval. If your data is strictly ordinal (e.g., Likert scales: Strongly Disagree, Disagree, Neutral, Agree, Strongly Agree), calculating a median of "3.7" is meaningless. For ordinal grouped data, simply identify the median category*.

When to Use the Median vs. the Mean for Grouped Data

You’ve calculated the median (≈ 38.In real terms, you could also calculate the grouped mean (using class midpoints). 2). Which should you report?

Use the Median when the distribution is skewed or contains outliers. In our exam example, if the 1 student in the 0–10 range actually scored a 0, the mean would be dragged down significantly. The median remains stable at ~38.2 because it only cares about the position* of the middle student, not the magnitude* of the extremes.
Use the Mean when the distribution is roughly symmetric and you need a measure that incorporates every data point for further algebraic manipulation (e.g., calculating standard deviation, combining groups, or regression analysis).

A quick visual check: compare the median class (31–40) to the modal class (also 31–40, frequency 9). Since they align and the frequencies taper off somewhat evenly on both sides (1, 3, 5 → 9 → 8, 3, 1), this distribution is approximately symmetric. Here, the mean and median will be close. But if the left tail were longer (e. g., many low scores), the median would be > mean; if the right tail were longer, median < mean.

Summary Checklist

Next time you face a grouped frequency table, run through this mental checklist:

[ ] Verify (n) (sum of frequencies).
[ ] Determine median position(s) ((n/2) and (n/2+1) for even (n); ((n+1)/2) for odd).
[ ] Construct cumulative frequencies (running total).
[ ] Identify the median class (where cumulative frequency first equals or exceeds the median position).
[ ] Extract parameters: (L) (lower boundary*), (F) (cumulative before), (f) (frequency of median class), (w) (class width).
[ ] Plug into formula: (L + \frac{M - F}{f} \times w).
[ ] Sanity check: Does the result lie inside* the median class boundaries? (If not, re-check (L) and (w)).

Conclusion

Finding the median from grouped data is an exercise in controlled estimation. Even so, because binning discards individual observations, we surrender precision for manageability; linear interpolation buys back a reasonable approximation by assuming a uniform spread within the median class. While the formula (L + \frac{M - F}{f} \times w) looks mechanical, its power lies in what it ignores: it renders the median immune to the distorting pull of extreme values that would cripple the mean.

Mastering this technique means you no longer have to treat grouped data as a "lesser" version of raw data. You can confidently report a central tendency that respects the data’s structure while sidestepping its limitations. Whether you are analyzing test scores, income brackets, or sensor readings binned for storage, the interpolated median gives you a single, defensible number that represents the typical* observation—without letting the extremes write the narrative.