Find a Median From a Relative Frequency Table
A worked explanation of find a median from a relative frequency table, including the key limitation and a checkable example.
Cumulative Proportions Locate The Middle Observation
A student stares at a table of income brackets with 10,000 rows but no individual salaries. The median is the 5,000th value, yet no single row holds it. Cumulative proportions solve this by turning the relative frequency table into a running total. The point where that running total crosses 0.50 (50 percent) locates the median class. Once you know the class, a standard interpolation formula gives the exact median. Cumulative proportions locate the middle observation when raw data are unavailable. The answer requires you to find a median from a relative frequency table, not from a list of numbers.
Find a Median From a Relative Frequency Table: The 50 Percent Crossing Rule
Every relative frequency table lists each class, its relative frequency (proportion of total observations), and the cumulative relative frequency. The last class always reaches 1.0. The median is the value where the cumulative relative frequency first reaches or exceeds 0.5. This is the 50 percent crossing. It works whether the total n is 50 or 50,000 because relative frequencies remove the need to know n. The crossing tells you which class contains the middle observation. That class is the median class.
A Worked Example
Use the example data from the fact sheet. The table has four classes: 0-10, 10-20, 20-30, and 30-40. Frequencies are 5, 15, 20, and 10, for a total n of 50. Relative frequencies are 0.10, 0.30, 0.40, and 0.20. Cumulative relative frequencies are 0.10, 0.40, 0.80, and 1.00. The 50 percent crossing happens at 0.80 in the 20-30 class. That is the median class. This is the first step in every calculation.
Calculation Steps
Once you identify the median class, apply the grouped data formula: Median = L + ((0.5n - F) / f) * w. L is the lower boundary of the median class (20). n is total frequency (50). F is the cumulative frequency before the median class (5 + 15 = 20). f is the frequency of the median class (20). w is the class width (10). The formula becomes 20 + ((25 - 20) / 20) * 10 = 20 + (5 / 20) * 10 = 20 + 2.5 = 22.5. The median is 22.5. This value falls within the 20-30 class, confirming the crossing point is correct.
Why Relative Frequencies Work
Relative frequencies make the calculation independent of the total count. A relative frequency calculator uses the same logic: it finds the 50 percent crossing from the cumulative relative frequency column, then applies the interpolation formula. The result is the same whether you use raw frequencies or relative frequencies, because the crossing point is identical.
Why the Method Works
The interpolation formula assumes observations are uniformly distributed within the median class. In the example, the 20-30 class contains 20 observations spread across 10 units. The median is the 5th observation into that class (since 25 total observations fall before the median, and 20 are already before the class). The formula places the median at 20 + (5/20)*10 = 22.5. This is an estimate, not an exact value. If the actual data within the class clusters at the lower end, the true median would be closer to 20.
When Raw Data Are Unavailable
Cumulative proportions work because they only require the relative frequency of each class, not individual values. Government surveys, for example, publish income brackets with percentages but not raw salaries. The 50 percent crossing directly identifies which bracket holds the median household. No individual data is needed.
| Class | Frequency | Relative Frequency | Cumulative Relative Frequency | Median Class? |
|---|---|---|---|---|
| 0-10 | 5 | 0.10 | 0.10 | No |
| 10-20 | 15 | 0.30 | 0.40 | No |
| 20-30 | 20 | 0.40 | 0.80 | Yes — crosses 0.50 |
| 30-40 | 10 | 0.20 | 1.00 | No |
Limits of the Failure Case: Zero Frequency in the Median Class
The interpolation formula fails when the median class has zero frequency. Imagine a table with a gap: class 0-10 has 30 observations, class 10-20 has zero, and class 20-30 has 20. The cumulative relative frequency jumps from 0.60 to 0.60 to 1.00, skipping 0.50 entirely. The 50 percent crossing does not occur within any class. The formula cannot be applied, and the true median is ambiguous. The median may fall in the gap, but the formula gives no result. This is the single thing that most often goes wrong: assuming a median always exists in a grouped table, when gaps can make it undefined without additional assumptions about the data distribution.
What to Do Instead
If the median class has zero frequency, you cannot compute a median from the table alone. Check whether the gap is real or a data entry error. If real, report that the median falls within the empty class but cannot be estimated. This is an honest caveat, not a failure of the method. The method works only when every class in the table has a non-zero frequency and the cumulative relative frequency crosses 0.50 within a class.
Common Questions
What happens if cumulative relative frequency exactly equals 0.50?
The median is the upper boundary of that class. No interpolation is needed because the middle observation falls exactly at the class boundary.
Can I use this method with open-ended classes?
No. If the first class has no lower boundary or the last no upper boundary, you cannot compute the median without assuming a boundary. The formula is undefined.
Does the method work for ordinal data?
Yes, because the median only requires order. The 50 percent crossing identifies the median category. You cannot interpolate, but you can report the category.
Why does my calculator give a different median for the same table?
The most common cause is a different quartile method. this analysis uses the standard grouped data formula from NCERT. Some software uses R-6 or R-7 for percentiles.