Descriptive Statistics · Measures of Center
Mean, Median, Mode & Weighted Average: Understanding Measures of Centre
Learn how to calculate and interpret the most common measures of central tendency. Mean, median, and mode each describe the center of a dataset differently, while a weighted average allows some values to contribute more than others.
The right measure depends on the structure of the data, whether extreme values are present, and the question you are trying to answer. Understanding those differences is as important as knowing the calculation itself.
Four measures you will encounter
Add, then divide
Add all observations and divide by the number of observations. Every value contributes to the result.
Find the middle
Sort the observations and locate the middle value, or the two central values when the dataset contains an even number of observations.
Find what occurs most often
The mode is the most frequently occurring value. A dataset can have one mode, multiple modes, or no mode.
Let contributions differ
A weighted average is useful when observations do not contribute equally, such as course grades, portfolios, surveys, or index calculations.
What are you trying to describe?
Central Tendency Calculator
Enter dataset values to calculate mean, median, mode, or all central tendency measures. The related tool also supports weighted averages, frequency-table means, and missing-value calculations from a target mean.
Averages Fundamentals · Central Tendency
Mean, Median, Mode & Weighted Average Explained
Measures of central tendency summarize a dataset from different perspectives. The mean describes an arithmetic center, the median identifies the middle ranked position, and the mode identifies the most frequently occurring value. A weighted average extends the mean when observations should not contribute equally.
One dataset can have several meaningful measures of center
Choosing a measure begins with the dataset and the statistical question—not with the assumption that every type of average will produce the same answer.
A dataset may be summarized with the arithmetic mean, median, mode, or weighted average. Each measure uses a different property of the data, so the appropriate choice depends on what the user needs to describe.
The statistical language behind averages
Collection of observations
The values being analyzed. A dataset might contain grades, measurements, survey responses, business metrics, or other observations.
An individual value
Each individual value in the dataset is an observation. In notation such as x₁, x₂, …, xₙ, each x represents an observed value.
A way to describe the center
Central-tendency measures summarize where values in a dataset are centered, but different measures define that center differently.
A value far from the others
An unusually high or low observation can influence some summary measures substantially. This is especially important when interpreting the arithmetic mean.
Relative contribution
A weight determines how strongly a value contributes to a weighted average. Unlike an ordinary mean, observations do not have to contribute equally.
How often a value occurs
Frequency records the number of times a value occurs. Frequencies can identify modes and can also act as weights when calculating a mean from a frequency table.
The mean uses every observation
The arithmetic mean is calculated by adding the observations and dividing by the number of observations. Because every value contributes to the total, changing any observation can change the mean.
The detailed formula, variable definitions, and calculation procedure are covered in the formulas and methods section.
The median depends on ranked position
The median is the middle value after the observations have been sorted. It therefore depends on position rather than directly on the magnitude of every value.
With an odd number of observations there is one middle observation. With an even number, the median is found from the two central values. The calculation procedure is covered in formulas and methods.
The mode is determined by frequency
The mode is the most frequently occurring value. Unlike the mean or median, a dataset is not required to have exactly one mode.
This distinction matters when using a mode calculator : the result should not automatically be forced into a single-value answer.
A weighted average allows unequal contributions
An ordinary mean gives equal weight to each observation. A weighted average allows some observations to contribute more strongly than others.
Each observation contributes equally.
Each value contributes according to its assigned weight.
Weighted averages are particularly relevant to course grades, investment portfolios, survey weighting, index calculations, and combined averages. See the weighted-average method .
A frequency-table mean is a weighted mean
When a value occurs multiple times, its frequency can be treated as its weight. This means the mean of a frequency table follows the same underlying weighted-average structure.
Mean and median do not respond to the data in the same way
This difference becomes especially important when a dataset contains an extreme observation.
Uses the magnitude of every observation, so the large value contributes directly to the result.
Depends on ranked position and is therefore often less sensitive to an extreme value.
The choice between these measures is explored further in mean, median, mode, and weighted-average comparisons .
How the measures differ
| Measure | Based on | Core idea | Important distinction |
|---|---|---|---|
| Arithmetic mean | All observations | Sum the values and divide by their count. | Can be strongly influenced by extreme observations. |
| Median | Ranked position | Find the center after sorting the observations. | Often less sensitive to outliers than the mean. |
| Mode | Frequency | Identify the most frequently occurring value or values. | There may be one mode, several modes, or no mode. |
| Weighted average | Values and weights | Allow observations to contribute unequally. | Weights determine each value’s relative contribution. |
| Frequency-table mean | Values and frequencies | Use each frequency as the corresponding value’s weight. | It is mathematically a weighted mean. |
Match the statistical question to the relevant concept
Begin with the dataset and the statistical question. Select the measure whose definition matches the aspect of the data you need to summarize, then interpret the resulting value in that context.
Calculation Method · Central Tendency
How to Calculate Mean, Median, Mode & Weighted Average
Each measure of central tendency uses a different calculation rule. The arithmetic mean uses the sum and count of all observations, the median uses ranked position, the mode uses frequency, and the weighted average combines values according to their relative weights.
Identify the measure before applying a formula
Start with the dataset and decide which property is required. Use the mean for an arithmetic center, the median for the ranked center, the mode for the most frequent value, or a weighted average when values contribute unequally.
Variables used in the formulas
- xi
- An individual observation or value.
- n
- The number of observations.
- Σ
- Summation symbol: add all terms described by the expression that follows it.
- wi
- The weight assigned to observation xi.
- fi
- The frequency associated with value xi.
- x̄
- The arithmetic mean of the observations.
Add all values and divide by the number of values
For observations x1, x2, …, xn, the arithmetic mean is the sum of all observations divided by their count.
Equivalent form: x̄ = (x1 + x2 + … + xn) ÷ n
- Count the observations. Record the total as n.
- Add every observation. Calculate Σxi.
- Divide the sum by the count. The result is the arithmetic mean.
Sort first, then identify the middle position
Median calculation depends on whether the number of observations is odd or even. In both cases, the observations must first be arranged in numerical order.
Use the single middle value
For five observations, the median occupies position (5 + 1) ÷ 2 = 3.
Average the two middle values
Here the two middle values are 8 and 11, so the median is (8 + 11) ÷ 2 = 9.5.
Count occurrences and identify the highest frequency
The mode does not require averaging values. Instead, count how often each distinct value occurs and identify the value or values with the greatest frequency.
In this dataset, 4 occurs twice while every other value occurs once, so the mode is 4.
Multiply each value by its weight before combining the results
When observations have unequal importance or contribution, pair each value xi with a weight wi.
- Pair each value with its weight. Keep every value-weight pair aligned.
- Multiply each value by its weight. Calculate wixi.
- Add the weighted products. Calculate Σ(wixi).
- Add the weights. Calculate Σwi.
- Divide. Weighted products ÷ total weight.
If weights are proportions or percentages, use them consistently. The general formula divides by their total, so the weights do not have to be written in a particular scale as long as they are internally consistent.
When all observations receive equal weights, the weighted-average structure reduces to the ordinary arithmetic mean.
Treat each frequency as a weight
A frequency table compresses repeated observations. Instead of writing the same value many times, multiply each distinct value by the number of times it occurs.
| Value (x) | Frequency (f) | Product (fx) |
|---|---|---|
| 10 | 2 | 20 |
| 20 | 3 | 60 |
| 30 | 1 | 30 |
| Totals | 6 | 110 |
The denominator is the total frequency, which is the total number of represented observations—not merely the number of rows in the table.
Solve for a missing value when the required mean is known
The mean formula can also be rearranged. If the desired mean, total number of observations, and all but one observation are known, the missing value can be calculated.
Required sum = target mean × total number of observations
Known sum = x1 + x2 + … + known values
Missing value = required sum − known sum
This is an algebraic rearrangement of the arithmetic-mean relationship. It does not apply unchanged to median or mode problems.
Precision, validity & edge cases
Sort before finding the median
Median position is meaningful only after observations are ordered.
Do not assume one mode
Check the complete frequency pattern for ties, multiple modes, or the absence of a mode.
Keep weights aligned
Each weight must remain paired with the observation it describes.
Use total frequency
For a frequency-table mean, divide by the sum of frequencies, not by the number of distinct values.
Round at the end
Keep sufficient precision during intermediate arithmetic and round the final result to a level appropriate for the data.
Check the dataset
An empty dataset does not provide observations from which these numerical measures can be calculated.
Which calculation rule should you use?
| Task | Method | Key requirement |
|---|---|---|
| Arithmetic mean | x̄ = Σxi ÷ n | Every observation contributes equally. |
| Median, odd count | Use the observation at position (n + 1) ÷ 2. | Sort the observations first. |
| Median, even count | Average the two central observations. | Sort the observations first. |
| Mode | Find the value or values with greatest frequency. | Allow multiple modes or no mode. |
| Weighted average | x̄w = Σ(wixi) ÷ Σwi | Pair every value with the correct weight. |
| Frequency-table mean | x̄ = Σ(fixi) ÷ Σfi | Frequency acts as the weight. |
| Missing value for target mean | (target mean × total count) − known sum | The target mean and final observation count must be known. |
Worked Examples · Central Tendency
Mean, Median, Mode & Weighted Average Examples
Worked examples make the differences between averages easier to see. The examples below identify the appropriate method, substitute the values, show the calculation, and explain what the result means in context.
Find the mean of five quiz scores
Suppose five quiz scores are 72, 78, 81, 84, and 85. Each quiz contributes equally, so the arithmetic mean is appropriate.
- Add the values 72 + 78 + 81 + 84 + 85 = 400
- Count the values n = 5
- Divide the sum by the count 400 ÷ 5 = 80
The five scores have an arithmetic mean of 80. This summarizes their arithmetic center; it does not mean that 80 must itself appear among the original observations.
Find the median of an even-sized dataset
Consider the values 18, 12, 25, 16, 20, and 14. Because the median depends on position, the first step is to sort the observations.
The median is 17. Three observations lie below the midpoint between 16 and 18, and three lie above it. With an even number of observations, the median does not have to be an observed value.
Identify the most frequently occurring value
A small dataset contains 6, 8, 8, 9, 10, 10, 10, and 12. The mode is determined by counting how often each distinct value occurs.
| Value | Frequency |
|---|---|
| 6 | 1 |
| 8 | 2 |
| 9 | 1 |
| 10 | 3 |
| 12 | 1 |
Ten occurs more frequently than any other value, so this dataset is unimodal with a mode of 10.
See how an extreme value can separate the mean and median
Suppose five values are 20, 21, 22, 23, and 150. The final observation is much larger than the other four.
(20 + 21 + 22 + 23 + 150) ÷ 5
236 ÷ 5 = 47.2Middle value after sorting
22The extreme value of 150 contributes directly to the arithmetic total, pulling the mean upward to 47.2. The median remains 22 because it is determined by ranked position. This illustrates why the two measures can describe the same dataset differently.
The implications of this distinction are explored further in the mean, median, mode, and weighted-average comparison section .
Calculate a course grade with unequal category weights
Suppose a course grade is based on three categories. The scores are 80, 90, and 70, but the categories contribute 20%, 30%, and 50% respectively.
| Category | Score | Weight | Weighted contribution |
|---|---|---|---|
| Category A | 80 | 20% | 80 × 0.20 = 16 |
| Category B | 90 | 30% | 90 × 0.30 = 27 |
| Category C | 70 | 50% | 70 × 0.50 = 35 |
| Total | — | 100% | 78 |
The weighted average is 78. The score of 70 has the largest influence because its category carries the largest weight. Simply averaging 80, 90, and 70 would ignore the unequal contributions.
Calculate a mean without expanding repeated observations
Suppose the value 10 occurs twice, 20 occurs three times, and 30 occurs once. Frequency can be used as the weight instead of writing all six observations separately.
| Value (x) | Frequency (f) | f × x |
|---|---|---|
| 10 | 2 | 20 |
| 20 | 3 | 60 |
| 30 | 1 | 30 |
| Totals | 6 | 110 |
The six represented observations have a mean of approximately 18.33. The denominator is 6 because that is the total frequency—not 3, the number of distinct values shown in the table.
Find the missing value needed for a target average
Four values will have a target arithmetic mean of 75. Three known values are 68, 74, and 80. What must the fourth value be?
A fourth value of 78 produces the required arithmetic mean of 75. This method works by reversing the arithmetic-mean relationship; it is not a general missing-value rule for the median or mode.
Where these calculations are useful
The appropriate measure depends on what the data represents and what the calculation is intended to summarize.
Scores & performance
Arithmetic means can summarize equally contributing scores, while weighted averages can handle categories with different contributions.
See the weighted-grade exampleData with extreme values
Comparing mean and median helps reveal when unusually high or low observations substantially affect the arithmetic center.
See the outlier exampleRepeated observations
Frequency tables summarize repeated values efficiently and can be used directly to calculate a frequency-weighted mean.
See the frequency exampleMost common values
Mode is useful when the question concerns which observed value occurs most frequently rather than the arithmetic or ranked center.
See the mode exampleTarget calculations
Reverse mean calculations can determine the value required to reach a specified arithmetic average when the other values are known.
See the target-mean exampleGeneral data summaries
Mean, median, and mode can be viewed together to understand different aspects of the same dataset rather than treating “average” as a single universal statistic.
Compare the measuresMatch the question to the calculation
| Question | Measure or method | Calculation focus |
|---|---|---|
| What is the arithmetic center? | Mean | Sum ÷ number of observations |
| What is the ranked middle? | Median | Sort, then identify the central position(s) |
| What occurs most often? | Mode | Compare frequencies |
| Do values contribute unequally? | Weighted average | Value × weight, then combine |
| Are repeated values summarized by frequency? | Frequency-table mean | Frequency acts as weight |
| What value produces a required mean? | Target-mean calculation | Required total − known total |
Comparisons · Assumptions · Limitations
Mean vs Median vs Mode: Which Measure Describes the Data?
Mean, median, mode, and weighted average describe different properties of a dataset. They can produce different results without any of those results being mathematically incorrect. The important question is what each measure represents and whether its assumptions match the data.
The measures answer different statistical questions
Arithmetic center
Uses every numerical observation in the arithmetic calculation.
- Based on
- Magnitude of all values
- Requires sorting?
- No
- Can be unobserved?
- Yes
- Extreme values
- Can materially affect it
Ranked center
Uses the middle position after observations have been ordered.
- Based on
- Ordered position
- Requires sorting?
- Yes
- Can be unobserved?
- Yes, with an even count
- Extreme values
- Usually less sensitive to magnitude
Most frequent value
Identifies the observation or observations occurring most often.
- Based on
- Frequency
- Requires sorting?
- No, but counting is required
- Observed value?
- Yes, when a mode exists
- Possible result
- One, multiple, or no mode
Weighted arithmetic center
Allows specified observations to contribute by different amounts.
- Based on
- Values and assigned weights
- Requires weights?
- Yes
- Can be unobserved?
- Yes
- Key dependency
- Valid value-weight pairing
| Measure | Main question answered | Uses every value’s magnitude? | Uses rank? | Uses frequency? |
|---|---|---|---|---|
| Mean | What is the arithmetic center? | Yes | No | Only indirectly through repeated observations |
| Median | What is the middle of the ordered data? | No | Yes | No |
| Mode | What occurs most frequently? | No | No | Yes |
| Weighted average | What is the arithmetic center after weighting? | Yes | No | Only when frequency is used as weight |
Mean and median respond differently to extreme values
Consider the ordered dataset 20, 21, 22, 23, 150. Four observations are close together while one is much larger.
47.2
The value 150 enters directly into the sum, so its magnitude pulls the arithmetic mean upward.
22
The median is determined by the middle ranked position, so the magnitude of the highest observation does not move that middle position.
Arithmetic mean vs weighted average
Arithmetic mean
Each listed observation contributes equally to the calculation.
Weighted average
Contribution is controlled by the weight assigned to each observation.
The results differ because the questions differ. The ordinary mean treats the three scores equally; the weighted calculation intentionally gives the score of 70 half of the total contribution. See the weighted-average calculation .
Not every measure applies equally to every type of data
Arithmetic operations can be meaningful
Mean and weighted average require numerical values for which addition, multiplication, and division have a meaningful interpretation.
Median requires meaningful ranking
The observations must be capable of being placed in an order from lower to higher or according to another meaningful ranking.
Mode is frequency-based
Mode can identify the most frequently occurring observation or category because its calculation depends on frequency rather than arithmetic magnitude.
| Measure | Needs numerical arithmetic? | Needs meaningful order? | Needs repeated values? |
|---|---|---|---|
| Mean | Yes | No | No |
| Median | Not for ranking alone; arithmetic is used for an even-count numerical median | Yes | No |
| Mode | No | No | A repeated value is needed for a unique modal frequency above all others |
| Weighted average | Yes | No | No |
One dataset can legitimately have several “average” values
For the dataset 2, 3, 3, 4, 8, each measure focuses on a different feature.
The mean is 4 while the median and mode are both 3. Reporting simply “the average is 4” would therefore be ambiguous unless average has already been defined to mean arithmetic mean in that context.
Check what must be true before interpreting the result
The dataset is the intended dataset
An accurate calculation can still answer the wrong question if relevant observations are omitted or unintended observations are included.
Values are represented consistently
Numerical observations should use compatible units or scales before arithmetic calculations are combined.
The selected measure matches the question
Arithmetic center, ranked center, frequency, and weighted contribution are different concepts.
Weights have a defined meaning
Weighted calculations depend on weights that correctly represent the intended contribution of each observation.
Median values are ordered correctly
The median cannot be identified reliably from an unsorted list by selecting whichever entry happens to appear in the middle.
Frequency ties are preserved
If several observations share the greatest frequency, the result should preserve that multimodal structure.
What an average does not tell you
An average does not show the full distribution
Different datasets can share the same central value while having very different spreads, clusters, gaps, or extreme observations.
The mean does not show whether observations are close to it
A mean summarizes arithmetic center, not how dispersed the individual observations are.
The median does not use every value’s magnitude
Once the ranked center is known, changing an extreme value may leave the median unchanged even though the dataset itself has changed substantially.
The mode may be absent or non-unique
Some datasets have no uniquely most frequent observation, while others have several values tied at the highest frequency.
A weighted average depends on the weighting scheme
Changing the weights can change the result even when the underlying observations remain identical.
Avoid these common interpretation errors
| Shortcut | Why it fails | Use instead |
|---|---|---|
| “Average” always means mean | Average can refer to different measures depending on context. | Name the specific measure. |
| Mean = median = mode | The measures are defined by different mathematical properties. | Calculate each measure separately when comparison matters. |
| The middle listed entry is the median | Median depends on ranked position. | Sort the data first. |
| A dataset must have one mode | There may be multiple modes or no uniquely most frequent value. | Inspect the complete frequency pattern. |
| Weighted average = ordinary mean | Unequal weights change each observation’s contribution. | Use the specified weights and divide by their total. |
| A central value describes the whole dataset | Center alone does not describe spread or distribution shape. | Interpret the average alongside relevant data context. |
Results that require extra care
Empty dataset
With no observations, the usual numerical mean and median calculations do not produce a value.
Single observation
For a one-value numerical dataset, that observation is both the arithmetic mean and median.
Even observation count
For numerical data, the median is the arithmetic mean of the two central ordered observations.
Multiple modal values
Preserve every value tied for the greatest frequency rather than selecting one arbitrarily.
Zero total weight
The weighted-average formula cannot divide by a total weight of zero.
Rounded inputs
An average calculated from rounded observations reflects those rounded inputs and should not imply greater source precision.
Match the measure to the property you need
Related Tool · Mean · Median · Mode
Central Tendency Calculator
Once you know what each measure represents, use the calculator to evaluate your own dataset. It can handle the arithmetic needed for mean, median, mode, and related average calculations while you retain control over how the result should be interpreted.
Start with the question you want the data to answer
The appropriate calculation depends on the statistical property you need, not simply on which result is easiest to calculate.
Calculate the mean
Use the arithmetic mean when the calculation should combine the magnitude of all numerical observations with equal contribution.
Review the mean methodCalculate the median
Use the median when the question concerns the middle position of an ordered dataset.
See how median differs from meanIdentify the mode
Use the mode when you need to identify the value or values that occur most frequently.
See the mode exampleUse a weighted average
Use a weighted calculation when observations have defined weights and are not intended to contribute equally.
See the weighted-average examplePrepare the data before entering it
Clean, consistently represented observations make the resulting statistics easier to calculate and interpret correctly.
Dataset values
Enter the numerical observations that belong to the dataset you intend to summarize.
Value separation
Keep individual observations clearly separated so that each value is treated as its own data point.
Weights, when applicable
A weighted calculation additionally requires the intended weight associated with each value.
Consistent representation
Values being combined arithmetically should use compatible units and a consistent numerical scale.
How the calculation changes with the selected measure
Calculation sequence: enter the intended dataset, determine which measure is appropriate, apply that measure’s calculation rule, and interpret the resulting statistic in the context of the data.
Sum the numerical observations and divide by the number of observations.
Order the observations and identify the central value, or the two central values when the numerical dataset has an even count.
Determine how often each value occurs and identify the greatest frequency, preserving ties when applicable.
Multiply each value by its corresponding weight, combine those products, and divide by the total weight.
Need the complete manual procedure? Return to the formulas and calculation-method section .
Interpret each output according to its definition
Reflects the magnitude of every numerical observation included in the calculation.
Represents the middle of the observations after they have been placed in order.
Identifies the observation or observations associated with the greatest frequency.
Reflects both the numerical observations and their assigned contributions.
Match the calculation to the data question
| Question | Calculation | Primary requirement | Important check |
|---|---|---|---|
| What is the arithmetic center? | Mean | Numerical observations | Values contribute equally |
| What is the middle ranked value? | Median | Meaningful ordering | Sort before locating the center |
| What occurs most often? | Mode | Frequency information | Preserve tied modal values |
| Do observations contribute unequally? | Weighted average | Values plus valid weights | Total weight must not be zero |
| Are repeated values summarized in a table? | Frequency-weighted mean | Value-frequency pairs | Use total frequency as the denominator |
Calculator use vs manual calculation
Use the calculator for
- Checking arithmetic for a numerical dataset
- Comparing central-tendency measures
- Working with longer lists of observations
- Checking a manually calculated result
- Evaluating repeated or weighted numerical values where supported
Use the educational method for
- Understanding why a formula works
- Showing calculation steps in an assignment or report
- Deciding what a statistic means
- Checking assumptions about the dataset
- Recognizing limitations and unusual cases
A central-tendency result is not a complete statistical analysis
Spread
A central value alone does not quantify how far observations are dispersed around that value.
Distribution shape
The result alone does not show clusters, gaps, symmetry, skew, or other structural features of the complete dataset.
Data quality
Correct arithmetic cannot establish that the entered observations are accurate, complete, or representative.
Interpretive context
A numerical output does not determine whether that statistic is appropriate for the user’s substantive question.
Run a quick statistical sanity check
- Dataset: Did you enter the intended observations?
- Measure: Are you reading mean, median, mode, or a weighted result?
- Weights: If used, do they represent the intended contributions?
- Representation: Are numerical units and scales compatible?
- Interpretation: Does the selected statistic answer the question you are asking?
Common Mistakes · Questions · Advanced Considerations
Common Mistakes When Calculating Mean, Median, Mode & Weighted Average
Most errors with averages come from choosing the wrong measure, preparing the data incorrectly, or interpreting a correct calculation too broadly. These checks help separate arithmetic mistakes from statistical interpretation mistakes.
Errors to check before trusting the result
Mistake
Assuming “average” always means mean
Mean is one type of average, but median, mode, and weighted averages answer different questions.
Mistake
Dividing by the wrong number of observations
For an ordinary arithmetic mean, the denominator is the number of observations included in the sum.
Mistake
Finding the median before sorting
The median is based on ranked position. The visually central entry in an unsorted list is not necessarily the median.
Mistake
Using only one middle value when the count is even
An even number of numerical observations has two central ordered values rather than one central observation.
Mistake
Assuming every dataset has exactly one mode
Several values can tie for the greatest frequency, and a dataset with no repeated value has no uniquely most frequent value.
Mistake
Averaging averages without considering group sizes
Taking the ordinary mean of subgroup means gives each subgroup equal influence, even when the groups contain different numbers of observations.
Mistake
Pairing a value with the wrong weight
A weighted average depends on the relationship between each observation and its corresponding weight.
Mistake
Rounding too early
Repeated intermediate rounding can alter the final result, especially in weighted calculations or calculations with many observations.
A short troubleshooting sequence
-
1
Check the data
Confirm that all intended observations are present and that unrelated observations have not been included.
-
2
Check the measure
Verify whether the question requires mean, median, mode, or a weighted average.
-
3
Check the method
Recount observations, verify ordering for the median, check frequencies for the mode, and confirm value-weight pairs.
-
4
Check precision
Look for premature rounding or a final answer that implies more precision than the input data supports.
-
5
Check interpretation
Ask whether the resulting statistic actually describes the feature of the dataset you intended to summarize.
Need to reconstruct the arithmetic? Return to the manual calculation methods .
Questions that commonly arise when working with averages
Can the mean, median, and mode be different?
Yes. They are calculated from different properties of a dataset. The mean uses numerical magnitude, the median uses ranked position, and the mode uses frequency. Different results can therefore all be correct.
Compare mean, median, and mode in detailCan the mean be a number that does not appear in the dataset?
Yes. The arithmetic mean is the sum of the observations divided by their count. That quotient does not have to equal one of the original observations.
Can the median be a number that does not appear in the dataset?
Yes. With an even number of numerical observations, the usual median is calculated from the two central ordered values. Their arithmetic mean may not be one of the original observations.
What happens if two values occur equally often?
If both values share the greatest frequency, both are modal values. More generally, several observations can tie for the highest frequency.
What if every value occurs only once?
No observation occurs more frequently than the others. Under the usual convention used on this page, the dataset therefore has no mode rather than treating every observation as a useful modal value.
Why can an extreme value change the mean so much?
Every numerical observation contributes directly to the sum used for the mean. An unusually large or small observation can therefore change that sum substantially. The median depends on ranked position instead, so it generally responds differently to extreme magnitudes.
Review the extreme-value comparisonIs the median always better when there is an outlier?
No universal rule makes one measure automatically preferable. The median is less sensitive to the magnitude of extreme observations, but the appropriate statistic still depends on what the analysis is intended to describe.
Do percentage weights have to add to 100%?
When weights are explicitly defined as shares of a whole, expressing them as percentages that total 100% is a common representation. The general weighted-average formula instead divides by the sum of the supplied weights, so weights can also be represented on another consistent scale when that scale has a valid meaning.
Review the weighted-average formulaCan I just average two group averages?
Only when giving the two groups equal influence is appropriate. If the group sizes differ and the goal is to reconstruct the overall mean of all observations, the subgroup means must be weighted by their respective observation counts.
Does a weighted average have to fall between the values?
With nonnegative weights and a positive total weight, the weighted average lies between the smallest and largest values. If a different weighting convention permits negative weights, that familiar boundary does not necessarily hold.
Should I round before or after calculating the average?
When practical, keep sufficient precision during intermediate steps and round the reported result at the end. If the source observations are already rounded, the final answer should not imply precision that the original data does not support.
Does an average tell me how spread out the data is?
No. Measures of central tendency summarize center or frequency; they do not by themselves quantify variability. Two datasets can have the same mean or median while their observations are distributed very differently.
Review the limitations of central tendencyA few deeper points that improve interpretation
Overall means depend on group size
Suppose Group A has mean x̄A from nA observations and Group B has mean x̄B from nB observations. Their combined mean can be reconstructed as:
Frequency can act as a weight
When a table records each value once together with the number of times it occurs, its frequency can be used as that value’s weight.
Relative weights matter
Multiplying every weight by the same positive constant does not change the weighted average because both the numerator and denominator are multiplied by that constant.
Missing is not automatically zero
A missing observation and a measured value of zero represent different information. Replacing unknown data with zero changes the dataset and can change its calculated averages.
Useful checks for catching calculation errors
| Calculation | Check | If the check fails |
|---|---|---|
| Mean | For a nonempty numerical dataset, the mean should lie between the minimum and maximum observations. | Recheck the sum, observation count, signs, and division. |
| Median | The result should correspond to the center of the ordered observations. | Re-sort the values and recount their positions. |
| Mode | Every reported modal value should have the greatest observed frequency. | Recount all occurrences and check for ties. |
| Weighted average | With nonnegative weights and positive total weight, the result should lie between the smallest and largest weighted values. | Recheck value-weight pairs, products, and the total weight. |
| Frequency-weighted mean | The denominator should equal the total number of observations represented by the frequencies. | Recalculate the frequency total before dividing. |