Average Calculator
What Is an Average?
An average is a single number that represents the central or typical value of a set of numbers, calculated by dividing the sum of all values by the count. In everyday language, "average" almost always refers to the arithmetic mean. Three other types exist — weighted, geometric, and harmonic — and each is suited to a different kind of data.
Statisticians prefer the specific term "mean" over "average" to avoid ambiguity, since an "average" could technically refer to the mean, median, or mode. When precision matters, say "mean." For most purposes, the two words are interchangeable.
The arithmetic mean accounts for every value in the dataset. The median accounts only for position. When a dataset contains extreme values, the two diverge significantly — and that gap is usually a signal worth paying attention to.
Average Formula
x̄ = Σx ÷ n
| Symbol | Meaning |
|---|---|
x̄ | Arithmetic mean ("x-bar") |
Σx | Sum of all values (sigma x) |
n | Count of values in the dataset |
xᵢ | Each individual value |
To find the average: add all the values, then divide by how many there are. The calculator above does this automatically and shows the full working.
How to Calculate an Average — Step by Step
How many numbers are in the dataset? This is n.
Find the total (sum = Σx). Include every number, positives and negatives alike.
Average = Σx ÷ n. Keep full precision until the final step, then round as needed.
Worked Example 1 — Test Scores
Dataset: 72, 85, 66, 89, 91, 78 (six test scores)
| Step | Calculation | Result |
|---|---|---|
| 1. Count | n = 6 | 6 values |
| 2. Sum | 72 + 85 + 66 + 89 + 91 + 78 | Σx = 481 |
| 3. Divide | 481 ÷ 6 | x̄ = 80.17 |
The average test score is 80.17 — equivalent to a solid B.
Worked Example 2 — Finding a Missing Value
This is the most common exam problem for averages, and the one most calculators ignore. If you know the target average and all but one value, you can work backwards:
Setup: You took 5 tests. After 4 tests, your average is 76. What score do you need on test 5 to achieve an overall average of 80?
Worked Example 3 — Negative Numbers
Dataset: −4, −2, 0, 3, 7. The process is identical — just include the negatives in the sum.
Count: 5 • Sum: −4 + (−2) + 0 + 3 + 7 = 4 • Average: 4 ÷ 5 = 0.8
Types of Average — Which One Should You Use?
The word "average" covers four distinct calculations. Using the wrong one can produce results that are technically correct but practically misleading. Here is how to choose.
| Type | Formula | Use When | Avoid When | Example |
|---|---|---|---|---|
| Arithmetic Mean | Σx ÷ n |
Data is symmetric, additive, same units | Strong outliers or ratio data | Average test score, average temperature |
| Weighted Mean | Σ(xᵢwᵢ) ÷ Σwᵢ |
Values have different importance/weights | All values contribute equally | GPA, portfolio returns, survey weights |
| Geometric Mean | (x₁×…×xₙ)^(1/n) |
Data is multiplicative: growth rates, ratios | Any value is zero or negative | Annual investment return, population growth |
| Harmonic Mean | n ÷ Σ(1/xᵢ) |
Averaging rates, denominator is constant | Data is not a rate | Average speed, average price per unit |
Arithmetic Mean — The Standard Average
The arithmetic mean treats every value equally. Its main limitation is sensitivity to outliers: one extreme value can pull the mean far from what most of the data looks like.
Example: Test scores of 60, 65, 70, 75, 95 → arithmetic mean = 73. Remove the 95: mean = 67.5. One high score raised the apparent average by 5.5 points.
Use the arithmetic mean for symmetric data with no extreme outliers, where values are additive (counts, temperatures, scores). Avoid it for skewed data, ratios, and growth rates.
Weighted Mean — When Values Have Different Importance
The weighted mean assigns each value an importance weight, then scales accordingly. The Weighted Average Calculator covers this in full, including GPA and portfolio presets.
| Course | Grade | Credits | Grade × Credits |
|---|---|---|---|
| Maths | 82 | 4 | 328 |
| English | 76 | 3 | 228 |
| Science | 90 | 4 | 360 |
| Art | 72 | 1 | 72 |
| Total | 12 | 988 |
Weighted Mean = 988 ÷ 12 = 82.3. The simple arithmetic mean of the four grades would be 80.0 — the heavier courses pull the accurate GPA higher.
Geometric Mean — For Rates, Ratios, and Growth
The geometric mean is the nth root of the product of n values. It is the right choice whenever data grows or compounds multiplicatively.
Why not arithmetic mean for investment returns? An investment that rises 50% in year 1 and falls 50% in year 2 has an arithmetic mean return of 0% — implying no loss. But $1,000 × 1.5 × 0.5 = $750. You lost money. The geometric mean: √(1.5 × 0.5) ≈ 0.866, which correctly shows −13.4% per year.
Example: Annual growth rates of 10%, 20%, 5% as multipliers (1.10, 1.20, 1.05). Product = 1.386. Geometric mean = 1.386^(1/3) ≈ 1.1152 → 11.52% average annual growth. The arithmetic mean (11.67%) slightly overstates it.
Harmonic Mean — For Rates and "Per Unit" Averages
The harmonic mean is used when averaging rates where the denominator stays constant (such as distance). It's the least well-known of the four but the most frequently misapplied.
Classic speed example: You drive 100 km at 60 km/h, then 100 km at 40 km/h.
- Arithmetic mean: (60 + 40) ÷ 2 = 50 km/h — this is wrong
- Actual time: 100/60 + 100/40 = 4.167 hours for 200 km
- True average speed: 200 ÷ 4.167 ≈ 48 km/h
- Harmonic mean: 2 ÷ (1/60 + 1/40) = 2 ÷ 0.0417 ≈ 48 km/h ✓
Whenever the denominator is held constant (same distance, same number of units), the harmonic mean is the correct choice.
When to Use Mean vs. Median vs. Mode
The right measure depends on your data's distribution, the presence of outliers, and your data type. For detailed worked comparisons, see the Mean, Median & Mode Calculator.
| Situation | Best Measure | Why |
|---|---|---|
| Symmetric data, no extreme outliers | Mean | Uses every value; most sensitive to small changes |
| Skewed data or significant outliers | Median | Unaffected by extreme values; more "typical" |
| Categorical data (colours, grades A/B/C) | Mode | Only measure that works for non-numeric data |
| Income, house prices, wealth | Median | A few very high values inflate the mean far above typical |
| Normally distributed exam scores | Mean | Symmetric distribution makes mean and median nearly equal |
| Bimodal data (two clusters) | Mode or both | Mean falls between the peaks, representing neither |
How Outliers Affect the Average
A single extreme value can shift the arithmetic mean far from the typical value of a dataset. Here's a concrete demonstration.
Seven employees earn these monthly salaries (£): 2,200 • 2,400 • 2,600 • 2,800 • 2,900 • 3,100 • 28,000 (the last figure is the CEO).
| Measure | Value | What it means |
|---|---|---|
| Mean | £6,286 | Heavily inflated by the CEO's salary |
| Median | £2,800 | The 4th of 7 values — unaffected by the extreme |
The mean of £6,286 suggests most employees earn comfortably above the team's typical wage — but six of the seven earn between £2,200 and £3,100. The median (£2,800) is far more representative of a typical employee's experience.
Two Mistakes That Look Like Averaging
These are the most common errors in real-world data analysis. Neither is covered fully by other calculator pages, so this section gives both the wrong method and the correct one, with numbers.
Mistake 1 — Averaging Percentages Without Group Sizes
Campaign B: 40% conversion
"Average" = (60% + 40%) ÷ 2 = 50%
Campaign B: 400 conversions from 1,000 visitors
True average = 460 ÷ 1,100 = 41.8%
Campaign B had ten times as many visitors, so its conversion rate should dominate the combined figure. Direct percentage averaging is only valid when both groups are exactly the same size. Otherwise, use the Weighted Average Calculator with group sizes as the weights.
Mistake 2 — Averaging Averages
Factory B average wage: £40,000 (100 employees)
"Average" = (£32,000 + £40,000) ÷ 2 = £36,000
Factory B total: 100 × £40,000 = £4,000,000
True average = £4,320,000 ÷ 110 = £39,273
Factory B has ten times more employees, so its wages should carry ten times the weight. When you average an average, always weight each one by its group size. The wrong figure (£36,000) understates the company-wide average wage by over £3,000.
Common Mistakes When Calculating Averages
1. Including zeros unintentionally. If a student was absent for one test and the teacher enters 0, the mean is pulled down. Zero is a real data value and is included in the calculation. If the absence should be excluded, remove the 0 from the dataset before calculating.
2. Using mean for heavily skewed data. When a dataset contains extreme values (high executive salaries, rare very long patient stays), the arithmetic mean is unrepresentative. Use the median instead. A quick check: if mean and median differ by more than ~20%, skew is likely.
3. Averaging percentages from different-sized groups. See Mistake 1 above. Direct percentage averaging is correct only when group sizes are equal.
4. Using arithmetic mean for growth rates or returns. A 50% gain followed by a 50% loss does not average to 0%. Use the geometric mean for multiplicative data.
5. Rounding intermediate steps. Rounding the sum before dividing introduces error into the final average. Keep full precision through all calculation steps; only round the final reported result.
Real-World Applications of Average
Education and Grades
The arithmetic mean is the basis for class averages and simple GPA calculations. When courses carry different credit hours, the weighted mean gives the accurate GPA. A student who scores 95, 92, 88, 85, and 12 on five tests has a mean of 74.4 — pulled down sharply by that one failing score. The median (88) is more representative of their consistent performance, and a teacher might consider both when assessing the student's understanding.
Finance and Investing
Compound annual growth rate (CAGR) uses the geometric mean, not the arithmetic mean. Companies report median salary benchmarks, not mean, because a few executives inflate the arithmetic figure. The P/E ratio of a stock portfolio is correctly averaged using the harmonic mean, since P/E is a price-to-earnings rate rather than an additive measure. For expected single-period returns, use the arithmetic mean; for multi-period compounding, use the geometric mean.
Science and Research
Researchers report the arithmetic mean when data is approximately normally distributed. When data is right-skewed (income, disease duration, study completion times), the median is preferred. The geometric mean appears in ecology (species counts), chemistry (pH averaging), and biological growth studies.
How to Calculate Average in Excel
Put your numbers in a column or row — for example, A1:A10.
Select the cell where you want the result to appear.
Use one of the formulas below depending on what you need.
| Formula | What It Does | When to Use |
|---|---|---|
=AVERAGE(A1:A10) | Arithmetic mean of range | Standard average |
=AVERAGEIF(A1:A10,">0") | Average of values meeting one condition | Average only positive values |
=AVERAGEIFS(range, crit1, val1) | Average meeting multiple conditions | Filter by multiple criteria |
=MEDIAN(A1:A10) | Median of the range | Skewed data or outliers |
=GEOMEAN(A1:A10) | Geometric mean | Growth rates, ratios |
=HARMEAN(A1:A10) | Harmonic mean | Averaging rates (speed, price/unit) |
=AVERAGE() ignores blank cells and text — a blank cell is not counted as zero. If blanks should be treated as zero, replace them with 0 before calculating. For weighted averages in Excel, use =SUMPRODUCT(values, weights)/SUM(weights).How to Calculate Average in Python
# Method 1: Built-in statistics module
import statistics
data = [72, 85, 66, 89, 91, 78]
mean = statistics.mean(data)
print(f"Average: {mean}") # 80.166...
# Method 2: NumPy (most common in data science)
import numpy as np
data = [72, 85, 66, 89, 91, 78]
mean = np.mean(data)
print(f"Average: {mean:.2f}") # 80.17
# Method 3: Pure Python (no libraries)
data = [72, 85, 66, 89, 91, 78]
average = sum(data) / len(data)
print(f"Average: {average:.2f}") # 80.17
# Geometric mean (scipy)
from scipy.stats import gmean
growth_rates = [1.10, 1.20, 1.05] # 10%, 20%, 5% as multipliers
geo_mean = gmean(growth_rates)
print(f"Geo mean: {geo_mean:.4f}") # 1.1152 → 11.52% per year
# Harmonic mean (scipy)
from scipy.stats import hmean
speeds = [60, 40]
h_mean = hmean(speeds)
print(f"Harmonic mean: {h_mean:.1f} km/h") # 48.0 km/h
np.mean() and statistics.mean() both return the arithmetic mean. For median: np.median(data). For mode: statistics.mode(data).
How to Calculate Average in R
data <- c(72, 85, 66, 89, 91, 78)
# Arithmetic mean
mean(data) # 80.16667
# Median
median(data) # 83.5 (average of 78 and 85 for even-count dataset)
# Geometric mean (no built-in; use exp(mean(log(x))))
growth <- c(1.10, 1.20, 1.05)
geo_mean <- exp(mean(log(growth)))
geo_mean # 1.115196 → 11.52% per year
# Harmonic mean (no built-in)
speeds <- c(60, 40)
harmonic <- length(speeds) / sum(1/speeds)
harmonic # 48 km/h
# Weighted mean (built-in)
grades <- c(82, 76, 90, 72)
credits <- c(4, 3, 4, 1)
weighted.mean(grades, credits) # 82.33333
R's mean() always returns the arithmetic mean. For the geometric and harmonic means, use the formulas above or the DescTools package (Gmean(), Hmean()).
Frequently Asked Questions
In everyday language, "mean" and "average" are used interchangeably. Statistically, "average" is informal — it could refer to the arithmetic mean, median, or mode. "Mean" specifically refers to the arithmetic mean (sum ÷ count) unless another type is specified. When precision matters in a statistical context, use "mean" rather than "average."
Three steps: (1) Count how many values there are. (2) Add all the values together to get the sum. (3) Divide the sum by the count. For example: average of 5, 10, 15 → count = 3, sum = 30, average = 30 ÷ 3 = 10. That is the arithmetic mean.
The four types are: (1) Arithmetic mean — sum divided by count, the standard average. (2) Weighted mean — values multiplied by their weights, used for GPA and portfolios. (3) Geometric mean — the nth root of the product, used for growth rates and investment returns. (4) Harmonic mean — count divided by the sum of reciprocals, used for speed and rate averaging.
Only if the groups they come from are exactly the same size. If Campaign A (100 visitors, 60% conversion) and Campaign B (1,000 visitors, 40% conversion) have different sizes, you cannot average the percentages directly. The true average is 460 ÷ 1,100 = 41.8%, not 50%. Use the weighted average with group sizes as weights.
Use this formula: Missing Value = (Target Mean × n) − Sum of Known Values. Example: if your average after 4 scores is 76 and you want an overall average of 80 from 5 scores, the required final score = (80 × 5) − (76 × 4) = 400 − 304 = 96.
Yes — the arithmetic mean includes every value in the calculation. One extreme outlier can shift the mean significantly. When outliers are present, the median is often a more representative "average." A quick check: if mean and median differ by more than about 20%, consider reporting the median as the main measure.
Use =AVERAGE(range) for the arithmetic mean. For example, =AVERAGE(A1:A10) returns the mean of the numbers in those cells. Excel ignores blank cells and text — they are not counted as zeros. For conditional averages, use =AVERAGEIF() or =AVERAGEIFS(). For geometric mean: =GEOMEAN(range). For harmonic mean: =HARMEAN(range).
Use the median when your data is skewed or contains significant outliers. Income and house prices are classic examples — a small number of very high values inflate the mean above the typical value. If your mean and median differ by more than about 20%, the median is likely the better representative measure.
The geometric mean is used for data that changes multiplicatively — investment returns, population growth rates, and compound rates. If an investment returns +50% and then −50%, the arithmetic mean (0%) implies no change, but you actually lost money. The geometric mean (−13.4% per year) correctly reflects the actual loss.
The harmonic mean (n ÷ Σ(1/x⊂i)) is used when averaging rates where the denominator is constant. The classic case is average speed over equal distances: if you drive 100 km at 60 km/h and 100 km at 40 km/h, the harmonic mean gives the correct average speed of 48 km/h, not the arithmetic mean of 50 km/h.
Related Calculators and Guides
The average is a starting point. These tools and guides take you further into descriptive statistics and statistical analysis: