Frequency is a simple count of how many times something appears in a dataset. Relative frequency takes that count and divides it by the total number of observations, giving you a proportion instead of a raw number. That single distinction changes how you can use the data, especially when comparing groups of different sizes.
Frequency: The Raw Count
Frequency (sometimes called absolute frequency) is the number of times a particular value has been observed. If you survey 128 people about their favorite superhero and 52 of them say Batman, the frequency for Batman is 52. No math beyond counting is involved.
Frequency is useful when you care about volume. A store manager looking at how many units of each product sold last week needs raw counts to decide how much to reorder. But frequency has a limitation: it doesn’t tell you anything about proportion. Knowing that 52 people chose Batman means little until you also know the total was 128, not 1,000.
Relative Frequency: The Proportion
Relative frequency answers a different question: out of all the observations, what fraction fell into this category? The formula is straightforward:
Relative frequency = frequency of the value ÷ total number of observations
Using the superhero example, Batman’s relative frequency is 52 ÷ 128 = 0.41. You can express that as a decimal (0.41), a fraction (52/128), or a percentage by multiplying by 100 (41%). All three say the same thing: 41% of respondents picked Batman.
A few properties make relative frequency especially useful. Every individual relative frequency falls between 0 and 1 (or 0% and 100%). And when you add up the relative frequencies for all categories in a dataset, they always sum to 1 (or 100%). If they don’t, something went wrong in the calculation.
A Side-by-Side Example
Imagine you surveyed 128 people about which superhero would win a fight. Here’s what both types of frequency look like together:
- Batman: frequency = 52, relative frequency = 52/128 = 0.41 (41%)
- Iron Man: frequency = 35, relative frequency = 35/128 = 0.27 (27%)
- Captain America: frequency = 24, relative frequency = 24/128 = 0.19 (19%)
- Superman: frequency = 17, relative frequency = 17/128 = 0.13 (13%)
The frequencies tell you how many people chose each hero. The relative frequencies tell you what share of the group chose each hero. Both describe the same data, just from different angles.
Why Relative Frequency Matters for Comparisons
The real advantage of relative frequency shows up when you’re comparing two datasets of different sizes. Suppose School A surveys 200 students and finds 80 prefer pizza for lunch, while School B surveys 500 students and finds 150 prefer pizza. Looking at raw frequency alone, School B has more pizza fans (150 vs. 80). But relative frequency tells a different story: School A’s pizza preference is 80/200 = 0.40 (40%), while School B’s is 150/500 = 0.30 (30%). School A actually has a higher proportion of pizza lovers.
This is why relative frequency is the standard choice whenever you’re comparing groups of unequal sizes. It standardizes the counts so that differences and patterns reflect genuine proportional differences, not just differences in how many people were surveyed.
How They Look on Charts
On a frequency histogram, the y-axis shows raw counts. Each bar’s height represents how many observations fell into that range or category. On a relative frequency histogram, the bars look identical in shape, but the y-axis shows proportions (or percentages) instead of counts. The overall pattern is the same because you’re dividing every bar by the same total. What changes is the scale, which makes the chart interpretable without knowing the total sample size.
This distinction is especially helpful in presentations or reports where your audience needs to grasp proportions at a glance. A bar reaching 0.41 on a relative frequency chart immediately communicates “about 41% of the data,” while a bar reaching 52 on a frequency chart requires your audience to also know the total.
The Connection to Probability
Relative frequency has a direct link to probability. If you flip a coin 1,000 times and get heads 510 times, the relative frequency of heads is 510/1,000 = 0.51. That number serves as an empirical (experimental) estimate of the probability of getting heads. The more times you repeat the experiment, the closer the relative frequency tends to get to the true probability.
This approach to probability is called the relative frequency method, and it’s how real-world probabilities are often estimated. Insurance companies, weather forecasters, and medical researchers all rely on observed relative frequencies to approximate how likely something is to happen. When a weather report says there’s a 30% chance of rain, that figure typically comes from the relative frequency of rain on days with similar atmospheric conditions.
When to Use Each One
Use frequency when the actual count matters. Inventory tracking, vote tallies, and headcounts all call for raw numbers because the totals drive decisions directly. Use relative frequency when you need to understand proportions, compare groups, or communicate patterns to an audience that doesn’t need to know the exact sample size. In practice, most data tables include both, because they complement each other: the frequency grounds you in the real numbers, and the relative frequency puts those numbers in context.