Secondary 3 Mathematics Learning Guide | Missing Values, Combined Means and Statistical Data Corrections

A mean is a total divided by a count. When a value is missing, a group is added or an entry is corrected, rebuild those two quantities before calculating the new mean. Many apparently different statistics questions become manageable once the learner keeps an accurate record of what changed and what did not.

This Secondary 3 Mathematics Learning Guide develops missing observations, combined means, unknown group sizes, missing frequencies, recording errors and corrected standard deviations. It concentrates on the algebra beneath statistical summaries rather than repeating how to read a histogram or box plot. Later sections also explain why standard deviations cannot generally be averaged and why a rounded mean may not determine an exact missing value.

Scope and source: the 2027 SEC G3 Mathematics syllabus includes means and standard deviations for grouped and ungrouped data. The questions here connect those ideas with equations and data interpretation. They are original consolidation tasks; school sequencing and assigned depth may differ.

Keep a Record of the Quantities Behind the Summary

Let N be the number of observations and S their sum. Then mean = S/N, so S = N × mean. When standard deviation is involved, also record Q, the sum of the squared observations. These are three different quantities: the count, the total and the total of squares.

Adding one observation changes N and S. Correcting one misrecorded observation changes S but leaves N unchanged. Removing an accidentally duplicated observation changes both. A missing frequency changes how many copies of a value belong in the data, so it changes the count and weighted total together.

This record is not an extra formula sheet. It is a way of preventing the denominator from being treated as background decoration. A mean can change because the observations change, because the number of observations changes, or because both change. The correct calculation depends on which of those events actually occurred.

A Diagnostic About Totals and Counts

Five values have mean 18. What is their total? A sixth value of 30 is added: what changes in the calculation? An original value of 30 is instead corrected to 20: does the number of observations change? Two groups have means 60 and 80: is their combined mean necessarily 70?

The first total is 90. Adding a sixth observation changes the total to 120 and the count to six, giving mean 20. Correcting an existing value leaves the count unchanged while replacing its contribution to the total. The combined mean is not necessarily 70 because the group sizes are needed.

A wrong answer of 24 after adding the sixth observation would suggest the new total was divided by the old count of five. An automatic answer of 70 for the two groups suggests the means were treated as two equally weighted observations. These are different denominator errors, and each becomes visible when the counts are written down.

Worked Example 1: Recover a Missing Observation

The five values 8, 10, 11, x and 16 have mean 12. Find x. The required total is 5 × 12 = 60. The four known values total 8 + 10 + 11 + 16 = 45. Therefore x = 15.

The same reasoning can be written as (45 + x)/5 = 12. Multiplying by five gives 45 + x = 60. The arithmetic-total route and the equation route express the same relationship.

Check by replacing the missing entry: 8 + 10 + 11 + 15 + 16 = 60. The median of the ordered data is 11, not 12. This illustrates why “average” should be identified precisely in a question. The mean imposes a sum condition; the median imposes a positional condition.

Worked Example 2: Add a New Observation

Eight observations have mean 15. A new observation of 24 is added. Find the new mean. The original total is 8 × 15 = 120. The new total is 144 and the new count is nine. The new mean is 144/9 = 16.

The added value is nine above the old mean. That extra nine is shared across the nine observations in the new collection, raising the mean by one. This second interpretation provides a useful check without repeating the same total calculation.

Adding a value above the old mean raises the mean, but the amount of increase depends on the enlarged count. It is not correct to average the old mean and the new observation as (15 + 24)/2. The old mean summarises eight observations; the new value represents only one.

Worked Example 3: Remove an Observation

Twelve observations have mean 18. The value 40 is removed. Find the mean of the remaining eleven. Original total = 216. Remaining total = 216 − 40 = 176. Therefore the new mean is 176/11 = 16.

Removing a value above the old mean lowers the mean, as the result shows. The denominator decreases because there is one fewer observation. Subtracting 40 from 18 would combine an individual data value with a summary that represents an entire collection; it would not calculate any meaningful new mean.

Do not confuse removing a value with correcting its recorded value to zero. Removing it reduces the count. Replacing it with zero preserves the count. The totals might change by the same amount, but the final means would be different because their denominators differ.

Worked Example 4: Combine Groups of Different Sizes

Group A contains eighteen students with mean score 64. Group B contains twelve students with mean score 79 on the same assessment. Find the combined mean. Recover each total: A contributes 18 × 64 = 1152 and B contributes 12 × 79 = 948.

The combined total is 2100 across thirty students, so the combined mean is 70. The simple average (64 + 79)/2 = 71.5 is not correct because the lower-mean group is larger.

The weighted-mean expression is (N₁m₁ + N₂m₂)/(N₁ + N₂). Each group mean receives the weight of its number of observations. This is not a special exception to the ordinary mean formula; it is the ordinary formula after the group totals have been reconstructed.

Check that 70 lies between 64 and 79, and closer to 64 because more students belong to Group A. These bounds do not prove the answer, but they can reject an impossible combined mean outside the two group means when both group sizes are positive.

When Is Averaging Two Means Valid?

The unweighted average of two group means gives the combined mean when the groups have the same size. It also gives the correct value when both means are already equal, regardless of positive group sizes. Outside such cases, reconstruct totals rather than assuming equal weights.

Before combining anything, check that the observations measure the same quantity on a compatible scale. Scores out of twenty and scores out of one hundred cannot simply be pooled as though their raw numbers had the same meaning. The question must specify the intended comparison or conversion. Correct weighting does not repair incompatible measurement definitions.

Worked Example 5: Recover an Unknown Group Mean

Thirty students have combined mean 70. Eighteen of them have mean 64. Find the mean of the other twelve. The complete total is 30 × 70 = 2100. The known group’s total is 18 × 64 = 1152.

The other group’s total is 2100 − 1152 = 948. Divide by its count, twelve, to obtain 79. This reverses the previous example. The difference 70 − 64 = 6 is not the unknown mean; it only compares two summary values.

The original combined mean gives one total condition. Knowing one group’s size and mean supplies a removable contribution. The unknown group is then what remains, both in count and in total. This same subtraction structure appears when a known class leaves a larger cohort or a subset of measurements is separated from a complete record.

Worked Example 6: Find an Unknown Group Size

One group has n observations with mean 60. A second group has twelve observations with mean 75. Their combined mean is 66. Find n. The combined total is 60n + 900 and the combined count is n + 12.

Set (60n + 900)/(n + 12) = 66. Then 60n + 900 = 66n + 792, so 108 = 6n and n = 18. Check: the combined total is 1080 + 900 = 1980 across thirty observations, giving 66.

A balance interpretation checks the structure. Each of the twelve observations in the higher-mean group contributes an average of 75 − 66 = 9 above the combined mean, for total excess 108. Each observation in the lower-mean group contributes an average shortfall of 66 − 60 = 6. Eighteen such shortfalls balance the excess.

This explanation does not claim every member of a group equals its mean. It compares group totals relative to a common baseline. That distinction matters: summaries can support calculations about totals without describing every individual observation.

Worked Example 7: Correct a Misrecorded Value

Forty measurements were reported with mean 18.6. One value was entered as 75 instead of the correct 57. Find the corrected mean. Treat the reported mean as exact for this exercise. The recorded total is 40 × 18.6 = 744.

Remove the incorrect contribution and insert the correct one: corrected total = 744 − 75 + 57 = 726. The number of measurements remains forty. The corrected mean is 726/40 = 18.15.

The mean falls by 18/40 = 0.45, which agrees with 18.6 − 18.15. The error changed one value by eighteen, but its effect on the mean is distributed across forty observations.

“Recorded incorrectly” is a different action from “a measurement was removed”. There are still forty genuine observations. Reducing the denominator to thirty-nine would create a second error while attempting to correct the first.

Worked Example 8: A Duplicate Entry Changes the Count

A computer record contains eight entries with mean 15. One entry of 20 is an accidental duplicate and must be deleted. Find the mean of the seven genuine entries. The recorded total is 120. Removing the duplicate leaves total 100 across seven entries, giving 100/7 ≈ 14.3.

Compare a different correction: one of the eight genuine entries was written as 20 but should have been 10. That would give total 110 across eight observations, so mean 13.75. The value twenty appears in both stories, but the action determines whether the count changes.

When reviewing a data-cleaning question, mark three words explicitly: add, remove or replace. Then write the new count before the new mean. This practical order prevents a common denominator mistake without requiring another memorised formula for every situation.

Worked Example 9: Find a Missing Frequency

A frequency table records values 10, 20 and 30 with frequencies 2, k and 4 respectively. The mean is 22. Find k.

Value x	Frequency f	Contribution fx
10	2	20
20	k	20k
30	4	120

Total frequency is k + 6, not three. The weighted total is 140 + 20k. Therefore (140 + 20k)/(k + 6) = 22. Multiplying gives 140 + 20k = 22k + 132, so k = 4.

The completed table contains ten observations with total 220. Its mean is 22. The variable k represents how many copies of twenty are present, not a new data value. That is why it appears in both the numerator and denominator.

A plausible frequency must be a nonnegative integer. If an exact equation produces a negative count or an unavoidable fraction, examine whether the data are inconsistent or whether a supplied mean was rounded. Do not round an impossible exact frequency into a new table without justification.

Worked Example 10: Two Missing Values Need More Information

The four values 4, 8, x and y have mean 10. Also, y is four greater than x. Find x and y. The mean gives a total of forty, so x + y = 28. The additional condition gives y − x = 4.

Adding the equations gives 2y = 32, so y = 16 and x = 12. Check the complete list: 4 + 8 + 12 + 16 = 40.

Without the difference condition, the mean would determine only the sum x + y. The pairs (12, 16), (13, 15) and (14, 14) would all fit. More calculation cannot recover information that the mean has not retained. The correct response to insufficient conditions is to state what is determined and what remains free.

Mean and Median Conditions Can Agree or Conflict

Consider the ordered list 4, 9, x, 14, 20. If its mean is 12, the total must be sixty, so 47 + x = 60 and x = 13. This is consistent with the stated order, since thirteen lies between nine and fourteen. Its median is thirteen.

If the same question also claimed its median was eleven, the conditions would conflict. The central position forces x = 11 while the total forces x = 13. No list of that stated form can satisfy both. A mean equation should therefore be checked against every extra condition, not treated as the only source of truth.

When a list is not stated to be ordered, do not automatically call its middle written entry the median. Order is a mathematical condition, not an assumption about how numbers happen to appear on the page.

Standard Deviation Needs the Sum of Squares Too

For the descriptive standard deviation used here, σ = √[Q/N − (S/N)²], where S is the sum and Q is the sum of squared observations. This matches the descriptive formula in the official syllabus formula sheet. Use the corresponding calculator statistic, not a different sample-estimation formula with divisor N − 1.

The formula can be understood by expanding squared deviations from the mean. The average of (x − mean)² is the average of x² minus the square of the mean. The square root returns the measure of spread to the original data units.

For example, if the data are lengths in centimetres, Q has units cm², Q/N − mean² also has units cm², and σ has units cm. The sum of squares Q is not the square of the sum S. Squaring a total introduces cross-products that do not belong in the total of individual squares.

Worked Example 11: Correct Both the Mean and Standard Deviation

Five recorded observations have S = 49 and Q = 657. One observation was entered as 21 instead of 12. Find the corrected mean and descriptive standard deviation. The count remains five.

The corrected sum is 49 − 21 + 12 = 40. The corrected sum of squares is 657 − 21² + 12² = 657 − 441 + 144 = 360. Thus the corrected mean is 40/5 = 8.

The corrected variance is 360/5 − 8² = 72 − 64 = 8. Therefore σ = √8 = 2√2 ≈ 2.83. The five correct values could be 4, 6, 8, 10 and 12, whose deviations from eight are −4, −2, 0, 2 and 4. Their squared deviations total forty, confirming variance 40/5 = 8.

The sum changed by −9, but the sum of squares changed by −297. It is wrong to subtract nine or eighty-one from Q. A recording correction replaces the old square with the new square: subtract 21² and add 12². The square of the difference is not the difference of the squares.

Keep the corrected mean exact inside the standard-deviation formula. Rounding the mean before squaring can introduce avoidable error, particularly when the two terms in Q/N − mean² are close. A negative computed variance is a signal to inspect arithmetic, input or rounding, not a real negative spread.

Worked Example 12: Shift and Scale Every Observation

A data set has mean 7 and standard deviation 1.5. Every value x is transformed to y = 2x + 3. Find the new mean and standard deviation. The new mean is 2 × 7 + 3 = 17.

Each transformed deviation from seventeen is (2x + 3) − 17 = 2(x − 7). All deviations double, so the new standard deviation is 3. Adding three shifts all observations together but does not change their relative spacing.

In general, adding a constant changes the mean by that constant and leaves standard deviation unchanged. Multiplying all values by a factor a multiplies standard deviation by |a|. The absolute value matters because spread is nonnegative even if a negative scale factor reverses the ordering of the data.

This is useful for consistent unit conversions and coded-data exercises. It is not permission to transform only the mean while leaving other observations or definitions unchanged. The rule applies because every observation undergoes the same stated transformation.

Worked Example 13: Why Standard Deviations Cannot Simply Be Averaged

Group A contains 8 and 12. Its mean is ten and descriptive standard deviation is two. Group B contains 18 and 22. Its mean is twenty and its standard deviation is also two. What is the standard deviation of the combined four observations?

The combined mean is fifteen. Deviations from fifteen are −7, −3, 3 and 7. Their squares total 49 + 9 + 9 + 49 = 116. Divide by four and take the square root: the combined standard deviation is √29 ≈ 5.39, not two.

Each original group was tightly clustered around its own mean, but the group means were far apart. Combining them creates additional spread around the new overall mean. Averaging the original standard deviations ignores that separation.

The same ledger method handles more advanced pooling. If a group has count N, mean m and descriptive standard deviation σ, its sum of squares is N(σ² + m²). Recover each group’s sum and sum of squares, combine those totals, then recalculate the overall spread. This is a guided extension of the correction method, not a need to memorise an unrelated pooling rule.

Worked Example 14: Grouped Data Produce Estimates

A grouped table has intervals 0 ≤ x < 10, 10 ≤ x < 20 and 20 ≤ x < 30, with frequencies 2, 4 and 4. Use midpoints 5, 15 and 25. The midpoint-weighted total is 2(5) + 4(15) + 4(25) = 170, giving estimated mean 17.

The midpoint-based sum of squares is 2(5²) + 4(15²) + 4(25²) = 50 + 900 + 2500 = 3450. Estimated variance is 3450/10 − 17² = 56. Estimated standard deviation is √56 ≈ 7.48.

These values summarise a midpoint model of the unknown observations inside each interval. They are not guaranteed to be the exact mean or spread of the original raw data. The table retains frequencies but not every individual value. The word “estimated” communicates a limitation of the available information, not uncertainty about the arithmetic.

A Rounded Mean May Leave a Range of Missing Values

Suppose five real-valued measurements have a mean reported as 12.4 to the nearest 0.1, and four known observations total fifty. Under the usual positive-number round-half-up convention, the true mean lies in 12.35 ≤ mean < 12.45.

The true total therefore lies in 61.75 ≤ S < 62.25. Subtract fifty to get 11.75 ≤ x < 12.25 for the missing measurement. Treating the reported mean as exactly 12.4 would choose x = 12, but the rounded information alone permits other real values in that interval.

This is an accuracy extension. In many school exercises, an unqualified given mean is intended as exact. Here the phrase “reported to the nearest 0.1” explicitly changes the information. Read that distinction before constructing an exact equation. Integer restrictions or further measurements could narrow the possibilities again.

Independent Practice: Rebuild the Record Before Dividing

Unless stated otherwise, treat given means and totals as exact. Use the descriptive standard deviation with divisor N. Show the updated count and total even when the arithmetic is short.

1. Six observations have mean fourteen. Five are 9, 11, 13, 16 and 18. Find the sixth.

2. Ten observations have mean 68 and fifteen have mean 76. Find the combined mean.

3. Twenty-five observations have mean 72.8. Ten of them have mean 68. Find the mean of the remaining fifteen.

4. Twenty observations were calculated to have mean 35. One value was recorded as 54 instead of 34. Find the corrected mean.

5. Nine observations have mean twenty. Remove one observation of 36. Find the new mean.

6. Seven observations have mean eighteen. Add an observation of thirty. Find the new mean.

7. Values 5, 10 and 15 have frequencies 2, k and 3. The mean is 10.5. Find k.

8. Four recorded values are 3, 5, 9 and 10. The nine should be six. Find the corrected mean and standard deviation.

9. A data set has mean twelve and standard deviation three. Every value is replaced by y = 4x − 5. Find the new mean and standard deviation.

10. Grouped intervals 0 ≤ x < 10, 10 ≤ x < 20 and 20 ≤ x < 30 have frequencies 2, 3 and 5. Estimate the mean using midpoints.

11. The values 5, 9, x and y have mean ten, and y − x = 6. Find x and y. Explain what would remain unknown without the difference condition.

12. Two groups each have standard deviation zero, but their means are ten and twenty. Explain why their combined standard deviation need not be zero.

Explained Answers

1. Required total = 6 × 14 = 84. Known total = 67. The missing observation is 17. Substitution into the six-value total verifies the result.

2. Group totals are 680 and 1140. Combined mean = 1820/25 = 72.8. The larger group has the higher mean, so the result is closer to 76 than to 68.

3. Whole total = 25 × 72.8 = 1820. Subtract 680 to leave 1140 across fifteen observations. Their mean is 76.

4. Recorded total = 700. Corrected total = 700 − 54 + 34 = 680. Count stays twenty, giving corrected mean 34.

5. Original total = 180. Remaining total = 144 and count = eight. New mean = 18. Both total and count change.

6. Original total = 126. New total = 156 and count = eight. New mean = 19.5. Averaging eighteen and thirty directly would give the wrong weights.

7. The equation is (55 + 10k)/(5 + k) = 10.5. Thus 55 + 10k = 52.5 + 10.5k, giving k = 5. Total frequency is ten and total value is 105.

8. Corrected values are 3, 5, 6 and 10. Their sum is 24 and sum of squares is 170. Mean = 6; variance = 170/4 − 36 = 6.5; standard deviation = √6.5 ≈ 2.55.

9. New mean = 4(12) − 5 = 43. New standard deviation = 4(3) = 12. The subtraction shifts the distribution without changing its spread.

10. Midpoint-weighted total = 2(5) + 3(15) + 5(25) = 180. Total frequency is ten, so estimated mean = 18. The actual raw mean is not recoverable exactly from these intervals alone.

11. The mean gives x + y = 26. Together with y − x = 6, this gives x = 10 and y = 16. Without the difference condition, only their sum would be fixed.

12. Zero spread inside each group means its values equal its own mean. Once combined, the tens and twenties differ from the overall mean. The separation between group means creates nonzero overall spread when both groups are present.

How to Check a Statistical Correction

Check the action first: did an observation enter, leave or change? Then inspect N, S and, when needed, Q. A single corrected value changes Q by new² − old², not by the square of the net change. A frequency change contributes that many copies to the count and that many weighted values to the total.

Next check direction and bounds. Adding a value above the mean should raise it; removing a high value should lower it. A positive-size combined mean should lie between the group means. Variance cannot be negative. These checks are fast, but they complement rather than replace the complete computation.

Finally, state what remains unknown. A mean does not reveal each individual score. A grouped table does not preserve exact raw values. Two standard deviations do not by themselves determine pooled spread without group counts and means. Good statistical reasoning includes respecting what the summary has discarded.

Teacher and Parent Prompts

Ask the learner what total the mean implies before suggesting an equation. When a correction fails, ask whether the count should change. When two means are combined, ask how many observations each represents. These prompts isolate the underlying decision instead of immediately repeating the calculator procedure.

For a useful transfer exercise, keep the original data but change “corrected” to “deleted”, or replace a missing observation with a missing frequency. The student must alter the count-and-total record rather than apply the same pattern mechanically. For stronger learners, compare raw-data calculation with the sum-and-squares method and explain why they agree.

Questions Students Often Ask

Can a corrected mean be a decimal even when the data are integers? Yes. An integer total divided by an integer count need not be an integer. The individual values and their mean are different kinds of quantities.

Why is standard deviation recalculated after a correction? The corrected value changes both its own contribution to the sum of squares and the mean about which spread is measured. Updating only one part leaves an inconsistent calculation.

Does a higher mean imply every observation is higher? No. A mean is a total-based summary. Two groups can overlap substantially even when their means differ. Individual comparisons require more detailed information.

Why can a missing-value question have no unique answer? The supplied summary may constrain only a total or a range. Additional observations, ordering conditions or exact rather than rounded information may be needed to isolate one value.

Continue the Secondary 3 Learning Route

Continue with Distance-Time and Speed-Time Graphs for averages over complete intervals; Probability with Unknown Quantities and Changing Sample Spaces for changing denominators; and Geometric Algebra, Constraints and Maximum-Area Problems for equations whose solutions require additional conditions.

A statistical answer is secure when the count, total, spread calculation and interpretation refer to the same corrected collection. Return to the Secondary Mathematics Hub.