Frequency is a weight
A frequency table compresses repeated observations by listing each value xᵢ and how many times fᵢ it occurs. The arithmetic mean of the expanded data equals the frequency-weighted mean: x̄=Σ(fᵢxᵢ)/Σfᵢ. Multiply every value by its frequency, add the products, and divide by the total frequency. Frequencies must describe counts for the same population and cannot be silently treated as percentages.
Inputs: values 2, 5, and 9 with frequencies 3, 4, and 1. Formula: x̄=(2×3+5×4+9×1)/(3+4+1). Calculation: numerator=6+20+9=35; denominator=8. Result: 35/8=4.375. Interpretation: the expanded data contain eight observations. Verification: listing 2,2,2,5,5,5,5,9 produces the same sum 35 and mean 4.375.
Build a reliable calculation table
Use columns for value, frequency, and product. Validate that values are numeric, frequencies are nonnegative integers for ordinary count data, and the frequency total is positive. A frequency of zero contributes nothing and may be omitted. Negative frequencies do not represent observation counts. Keep the product sum and frequency sum separate; dividing by the number of rows is the unweighted mean of distinct labels, not the mean of all observations.
Inputs: scores 60, 70, 80, 90 with frequencies 1, 2, 4, 3. Formula: Σfx/Σf. Calculation: products are 60, 140, 320, 270; totals are 790 and 10. Result: 79. Interpretation: 80 has the greatest influence because it occurs four times. Verification: 79 lies between the minimum 60 and maximum 90.
Relative frequency and percentage tables
If weights are relative frequencies or percentages, use Σ(wx)/Σw. When valid percentages sum to 100, dividing the weighted product sum by 100 gives the mean. When displayed percentages sum to 99.9 or 100.1 because of rounding, normalizing by the actual displayed total avoids assuming they are exact. Do not mix count frequencies and percentages in one weight column.
Inputs: values 10, 20, 40 with displayed shares 25%, 35%, 40%. Formula: mean=Σ(value×share decimal). Calculation: 10×0.25+20×0.35+40×0.40=2.5+7+16. Result: 25.5. Interpretation: the 40-value group has the largest stated share. Verification: shares sum to 1 and the result lies inside 10–40.
Grouped intervals use representative values
A grouped frequency table may list intervals rather than exact observations. A common estimate uses each class midpoint mᵢ and computes Σ(fᵢmᵢ)/Σfᵢ. This is an approximation unless every observation equals its midpoint or the original values are otherwise known. Open-ended classes such as “80 and above” lack a defined midpoint and require an explicit assumption or raw data.
Inputs: intervals 0–9, 10–19, 20–29 with frequencies 2, 5, 3. Formula: midpoint mean using 4.5, 14.5, 24.5. Calculation: products 9, 72.5, 73.5 sum to 155; frequency total 10. Result: estimated mean 15.5. Interpretation: grouping has discarded within-class positions. Verification: the estimate lies within the overall 0–29 range.
Find a missing frequency or value
Reverse problems turn the mean formula into an equation. Write every known weighted product, represent the missing frequency or value with a variable, and solve while preserving constraints. A frequency solution should normally be a nonnegative integer. If algebra gives 3.6 observations, either the inputs are rounded, the model is not a count table, or no exact count solution exists.
Inputs: values 4 and 10 have frequencies 6 and n, and the mean is 7. Formula: (4×6+10n)/(6+n)=7. Calculation: 24+10n=42+7n, so 3n=18. Result: n=6. Interpretation: equal frequencies of values equally spaced around 7 balance the mean. Verification: (24+60)/12=7.
Spreadsheet setup and precision
In a spreadsheet, place values in one range and frequencies in another, then use SUMPRODUCT(values,frequencies)/SUM(frequencies). Confirm ranges have equal length and exclude headers. Blank cells, text, hidden rows, and filters can change behavior between applications; inspect the included range rather than trusting a copied formula. Preserve full precision until the final display.
If values are currency or measurements, units follow the weighted mean while frequencies remain counts. Round according to the reporting purpose, not the number of rows. A mean of 79.46 rounded to one decimal is 79.5; rounding each value first can change the answer. Retain raw table values so the calculation can be reproduced.
Interpretation and audit limits
The weighted mean summarizes center but not spread, skewness, sample size uncertainty, or subgroup fairness. Different distributions can share the same mean. Frequencies from a biased sample do not become representative merely because the arithmetic is exact. For grades, finance, health, or official statistics, definitions and inclusion rules matter as much as calculation.
Audit by confirming the table population, weight type, total frequency, each fx product, numerator, denominator, and plausible range. Re-expand a small table or independently use a spreadsheet. Report midpoint substitution for grouped data and normalization for rounded percentages. NumUtility can calculate weighted values, but it cannot determine whether the frequency table was collected correctly or whether a mean is the appropriate summary.
Compare weighted and unweighted summaries
A frequency-weighted mean differs from the unweighted mean of the listed values whenever frequencies are unequal. Inputs: values 10, 20, and 30 with frequencies 1, 1, and 8. Formula: weighted mean=(10+20+240)/10. Calculation: 270/10=27. Result: 27, while the unweighted row mean is 20. Interpretation: most observations equal 30, so ignoring frequency understates the center.
Verification: expanding the table creates ten observations whose sum is 270. The median of this expanded set is also 30, showing that different location statistics answer different questions. Report the weighted mean as 27, not as a typical guaranteed observation. If the frequency column represents survey weights rather than literal counts, additional sampling methodology may be required and the effective sample size cannot be read directly from Σf.
A table can also contain cumulative frequency. Convert it to ordinary class frequencies before applying the formula by subtracting consecutive cumulative totals. Using cumulative figures directly double-counts earlier observations. Confirm that cumulative values never decrease and that the final cumulative value equals the total sample count; otherwise the table needs correction before any mean is trustworthy.