Find your next learning step
If a blank entry has been turned into zero, begin with Chapters 1–5. If you are analysing surveys, see the denominator and response chapters. For data tables, averages and graphs, use the corresponding interpretations, then test your understanding through the 29 solved practice situations.
Full worked data example · 29 practice situations · Mathematics learning hub
A missing value is not automatically zero. Zero records that the measured quantity was zero. Missing means the relevant value was not supplied or observed. For data 5,7 and a blank, the mean of the two recorded values is six, but the mean of all three intended observations cannot be determined without knowing the third.
Replacing the blank with zero would produce (5+7+0)/3=4. The arithmetic is correct for a different dataset, but the input is invented unless a data rule explicitly assigns zero. The distinction affects tables, means, percentages, graphs, frequency counts, probability and standard deviation.
The 2027 SEC G1, G2 and G3 Mathematics syllabuses develop data interpretation and problem solving at different levels. This guide teaches a cross-level reasoning habit and includes optional examples of data cleaning and uncertainty; it does not claim identical examination requirements across subject levels.
- Quick read
- 1. A missing value and a measured zero say different things
- 2. A genuine zero belongs in the mean
- 3. The number of observed values is not always the group size
- 4. Full worked example: four observations and one blank
- 5. Writing an unknown as x preserves information
- 6. A dash or blank needs a legend
- 7. Zero frequency is not a missing measurement
- 8. A missing category should not be added to the total automatically
- 9. Missing survey replies are not negative answers
- 10. A percentage needs an explicit denominator
- 11. Unknown does not mean a value is small
- 12. Worst-case and best-case bounds are conditional tools
- 13. Omitting missing cases may change an average
- 14. A zero is not always a missing-data code
- 15. Do not invent zero readings during a sensor outage
- 16. A graph gap is not necessarily a drop to the axis
- 17. Interpolation is a guess from a model, not an observation
- 18. Removing a value is not the same as replacing it
- 19. Imputation requires a stated assumption
- 20. A low count of observations weakens interpretation
- 21. Missing entries and zero values can affect standard deviation differently
- 22. A zero is meaningful only with its measurement unit
- 23. Beware of a table total that ignores missing rows
- 24. Missingness can differ between groups
- 25. A missing observation cannot be recovered without enough information
- 26. Inclusion-exclusion counts need reliable group totals
- 27. Mathematical models should expose missing inputs
- 28. A precise answer may be ‘not enough information’
- 29. Connect missing-data reasoning across G1, G2 and G3
- 30. A final checklist before reporting any statistic
- 29 solved data-reading questions
- Frequently asked questions
Three different states must stay separate: recorded zero, unavailable observation and not-applicable entry. Use a blank as unknown unless the source defines it otherwise. For an observed-only mean, sum the observed numbers and divide by the observed count. If the question asks for the full intended population mean, show an unknown or bounds when values are missing.
In surveys, use the responding group for a percentage of respondents and do not treat nonresponders as no votes. In graphs, a missing reading is not an observed fall to zero. In standard deviation, an invented zero can change spread. State all assumptions before completing missing entries with estimates.
The number zero is a numerical observation. A missing value means the relevant number was not supplied, not recorded or cannot be determined from the available information. Replacing an unknown value with zero creates a new factual claim that the original data did not make. A correct calculation on that replacement may therefore answer the wrong question.
Worked checkpoint. Imagine two recorded journeys of 5 and 7 kilometres, plus one day with no distance reported. The mean of the two known distances is 6 kilometres. If the blank is filled with zero, the three-value mean becomes 4 kilometres, but that assumes the unrecorded journey distance truly was zero.
The correct response depends on the question: the known-data mean is six, while the mean over all three days is not uniquely determined without further information.
Study action. Read blank as ‘not known’ unless the table legend or a clear instruction explicitly defines the blank as zero.
When a measurement was actually made and its value is zero, it must be treated as a data value rather than removed. Zero can represent no sales, no rainfall, no goals or no waiting time. Deleting it changes the dataset and usually changes the mean.
Worked checkpoint. A shop records daily sales 5,7,0 on three observed days. Its mean sales is (5+7+0)/3=4. The zero is a valid recorded outcome, not a gap. If the third cell had instead been blank because the reporting system failed, the total and three-day mean would remain uncertain.
The numerals may look similar in a spreadsheet, but a blank and the digit 0 carry different information.
Study action. Before excluding a value, ask whether zero was actually measured or simply absent from the records.
Averages are calculated from a defined collection of values and an appropriate count. When some observations are missing, the denominator for the observed-only mean is the number actually observed, not automatically the number invited, scheduled or expected. But the observed-only result must not be relabelled as the full-population mean without justification.
Worked checkpoint. If five students were expected to report results and only three scores 6,8,10 are available, the observed-only mean is (6+8+10)/3=8. Dividing by five would imply that the two missing results contributed nothing. The mean across all five students cannot be determined exactly from the supplied scores.
An observed mean can still be useful, but it describes a smaller dataset and may be biased if missingness is related to performance.
Study action. Label both the observed count and the intended total count in the working.
A teacher’s simplified practice table contains five entries: 4,6,8,10 and one blank. The question asks for the mean of the values recorded and whether the mean of all five intended results is known. Distinguish the two targets before calculating. Do not convert the blank into a result unless the problem states what it means.
Worked checkpoint. The four recorded numbers sum to 28, so their observed mean is 28/4=7. If the missing fifth value is x, the five-result mean is (28+x)/5. This cannot be given as one number without knowing x. If x happened to be zero, the five-result mean would be 5.6, but that is a conditional case, not the known answer.
The expression (28+x)/5 is more informative than inventing a fifth score, because it makes the missing information visible.
Study action. Answer in two sentences: ‘The recorded-value mean is seven. The mean across all five results is undetermined because one score is missing.’
| What is being asked? | Supported result | Why |
|---|---|---|
| Recorded-value mean of 4,6,8,10 | 7 | 28 ÷ 4 observed values |
| Full mean of 4,6,8,10,x | (28+x)/5 | Fifth value unknown |
| If x=0 | 5.6 | Only a conditional completion |
| If x=12 | 8 | Another possible completion |
When a missing value is mathematically relevant, an unknown variable can represent it without asserting a numerical value. This is one of the key benefits of algebra: unknown does not mean zero, and the formula can show exactly what additional information would be needed.
Worked checkpoint. For results 4,6 and a missing x, the full mean is (10+x)/3. If the full mean is later given as five, solve (10+x)/3=5 to obtain x=5. Only that added information determines the missing result.
If no full mean, total or other relationship is supplied, x must remain unknown. Picking zero simply to make arithmetic easy is unjustified.
Study action. Write a variable next to the blank and state the extra condition that would make the calculation solvable.
Data tables sometimes use a dash, an empty cell, ‘N/A’, a shaded box or zero for different reasons. These symbols are conventions, not universal mathematical definitions. A table legend or task instruction decides whether a mark means not observed, not applicable, withheld, negligible or an actual zero.
Worked checkpoint. A timetable may show a dash for a service that does not run, while a sales report may show zero units sold. A science table might use a dash for a missing measurement after a device failure. Identical display marks need not carry identical meanings across sources.
Using an ambiguous symbol without explanation can damage averages and totals even when the arithmetic formula is correct.
Study action. Read the table key first. If it does not explain the mark, describe the uncertainty rather than guessing.
In a frequency table, the value column names an observation and the frequency column says how many times it occurred. A frequency of zero means a known category occurred zero times in the recorded dataset. A blank frequency means the count is not supplied unless an explicit convention says otherwise.
Worked checkpoint. Suppose shoe sizes 5,6,7 have frequencies 2,0,3. The recorded sample includes two size-fives and three size-sevens, with no size-sixes. The total frequency is five. If the size-six frequency were blank rather than a recorded zero, the total count could be unknown.
Numerical zero in the frequency column is meaningful information. It does not create an additional observation with value zero.
Study action. Name which column contains a data value and which contains a count before forming a mean or total.
SECTION 10 OF 34
8. A missing category should not be added to the total automatically
A table may list categories that did not appear, or it may omit some categories entirely. The displayed category labels do not by themselves prove that the data collection was exhaustive. Totals and percentages must be based on the recorded counts and any statement about missing categories.
Worked checkpoint. A class survey lists 12 students choosing football and 8 choosing badminton. Without a statement that all students responded and could choose exactly one, the table does not establish the class size or the number choosing neither. Adding 12+8 is simply a count of responses across those categories, subject to overlap rules.
The missing-information issue is separate from inclusive-or overlap, though the two can occur in one problem.
Study action. Check whether the response options were exhaustive and mutually exclusive before announcing a full-population total.
Someone who did not respond to a yes-or-no question should not automatically be counted as a ‘no’. Nonresponse is a third information state. A reported percentage must state whether its denominator is the number who answered or everyone invited.
Worked checkpoint. Suppose 100 students receive a survey, 80 respond, and 40 say yes. The yes share among respondents is 40/80=50%. It is also true that 40 of the 100 invitees are known yes responses, but we cannot conclude that only forty percent of all invitees would have answered yes.
The twenty nonrespondents might all answer yes, all answer no or split between them. Those possibilities produce different population proportions.
Study action. Write ‘among respondents’ when reporting 50% and do not silently assign an answer to missing responses.
Percentages express a part relative to a particular whole. Missing data can change what that whole is. If the denominator changes between reports, the same numerator may produce different percentages without any arithmetic mistake.
Worked checkpoint. Forty yes responses are 50% of eighty replies and 40% of one hundred invitations. Both calculations are mathematically correct but answer different questions. Neither calculation by itself measures the unknown opinions of twenty nonrespondents.
A reader may infer an unproven conclusion when a percentage is reported without its denominator label.
Study action. State the numerator, denominator and target population in one sentence before giving the percentage.
A missing observation could be lower or higher than the recorded ones. Treating it as zero is especially misleading when the missing value is potentially large. Some circumstances allow a sensible bound, while others provide no numerical range at all.
Worked checkpoint. If recorded scores are 4 and 6 with a third score missing, the full mean is (10+x)/3. If scores are restricted to the interval 0 to 10, the mean could range from 10/3 to 20/3. It is not automatically 10/3 simply because zero is the lowest permitted score.
The interval expresses uncertainty honestly. A bounded answer is different from a uniquely known answer.
Study action. Use the stated minimum and maximum possible values to calculate a range when the question provides valid limits.
When the missing value has an allowed range, computing lower and upper bounds can show how much uncertainty remains. This does not mean the extreme values are equally likely or that one of them is the actual missing value. Bounds describe what is possible under stated constraints.
Worked checkpoint. With known values 4 and 6 and x from 0 to 10, the full mean (10+x)/3 ranges from approximately 3.33 to 6.67. If a later clue says x must be between 5 and 8, the mean range narrows to 5 to 6.
Giving a mid-point guess without an explicit model would add a new assumption. In many classroom questions, showing the range is the strongest justified conclusion.
Study action. Write the restriction on x next to the mean expression, then calculate both extreme outcomes.
One common method reports the mean of available observations only. This is mathematically valid for the observed dataset, but it estimates the full group reliably only under additional conditions about how values become missing. If low or high outcomes are more likely to be unrecorded, the observed mean may misrepresent the intended population.
Worked checkpoint. Suppose three completed quizzes have scores 8,9,10 but two uncompleted quizzes have unknown scores. The observed mean is nine. It is unjustified to assume the two absent results would also average nine, or that they must both be zero, without a relevant scoring rule.
The calculation is not wrong. The uncertainty is about generalising from the observed subset to all expected outcomes.
Study action. Name the observed subset and describe any relevant selection concerns when interpreting its average.
Some datasets deliberately reserve special numeric values such as −1 or 999 to mean ‘not recorded’. Those are codes rather than genuine observations, but that meaning comes from the data dictionary. Other datasets legitimately contain zero and negative measurements, so no digit is universally safe as a missing marker.
Worked checkpoint. Suppose a temperature log defines 999 as a sensor-error code. Including 999 in an average of temperatures around 25°C would create a nonsense result. If the data dictionary instead says 999°C was actually observed by an instrument, that would require separate plausibility and instrument checks.
Replacing or removing a coded value must be based on a documented convention, not an intuition that the number looks strange.
Study action. Look up the field’s stated coding rules before performing any cleaning calculation.
A time series may contain gaps because an instrument stopped working or information was not transmitted. A gap is not a measurement of zero physical quantity. Filling it with zeros can create artificial drops, change totals and produce misleading trends.
Worked checkpoint. A rainfall gauge reports 4 mm on Monday and 6 mm on Wednesday, but Tuesday is missing due to maintenance. The observed total is 10 mm across the two measured days. The total rainfall for all three days is unknown unless Tuesday’s amount is supplied or independently established.
Reporting a three-day total of ten would implicitly assume zero rainfall on Tuesday, which the outage does not prove.
Study action. Mark the missing interval visibly on the timeline rather than drawing a line down to zero as if the reading were observed.
Graphs of measurements may omit points when a reading is unavailable. Connecting the previous value to zero and back can falsely imply an actual event. A visual representation should distinguish missingness from recorded zero and indicate when interpolation is only an estimate.
Worked checkpoint. If a speed sensor reports 20 km/h at 10:00 and 25 km/h at 10:10 but no 10:05 reading, a zero-speed point at 10:05 is unsupported. A dotted connecting line might illustrate an interpolation, but it would still be a model-based estimate.
Graphs can communicate uncertainty well when missing intervals are clearly labelled, while an automatic zero fill conceals it.
Study action. Check whether every plotted point represents an observation or a constructed estimate.
An interpolation formula can estimate a missing value between known measurements, but it does not transform that estimate into a recorded fact. Even linear interpolation assumes the process changed smoothly over the interval, which may not be justified.
Worked checkpoint. Given values 4 at time zero and 8 at time two, a linear interpolation estimates 6 at time one. But a process could have jumped to ten and returned to eight. The endpoints alone do not establish the intermediate value.
If a lesson asks students to interpolate, report the estimate with its assumption. If it asks what was observed, do not include the estimate among original readings without distinction.
Study action. Use a different symbol or label for an estimated point and explain the model producing it.
Some workflows omit incomplete records, while others replace missing values with an estimate. These choices can generate different means, spreads and group sizes. Neither is a universal mathematical law. The appropriate decision depends on the purpose, missingness mechanism and instructions.
Worked checkpoint. For values 4,6 and one blank, complete-case calculation uses {4,6} and produces mean five. Replacing the blank with five produces {4,6,5}, still mean five but a different three-observation dataset and different variation. Replacing it with zero yields mean 10/3.
Equal means after two treatments do not make the resulting datasets equivalent. Their counts and standard deviations can differ.
Study action. Before analysing, state whether data were omitted, imputed or treated as a genuine zero and why.
Filling a missing cell with a mean, median, forecast or model prediction is called imputation. It may be useful in some analyses, but it is a methodological choice with uncertainty, not a statement that the missing value truly equals the inserted number.
Worked checkpoint. If values 4 and 6 are observed and the third is missing, mean imputation would insert five because that is the mean of the known entries. The completed dataset {4,6,5} has mean five, but no evidence proves the unobserved third value was five.
Treating the imputed value as certain can understate uncertainty and distort calculated spread.
Study action. Label imputed values and preserve information about which entries were observed versus estimated.
A statistic computed from only a few nonmissing observations may be numerically accurate for that subset while being less informative about the intended larger group. Missingness reduces available evidence; it does not merely inconvenience a calculator.
Worked checkpoint. A class with thirty enrolled learners but only three submitted survey scores can calculate the mean of those three scores. It cannot confidently describe all thirty learners from that mean alone without assumptions about who responded.
Exam word problems often simplify the data context, but readers should not conflate a defined arithmetic mean with a representative population estimate.
Study action. Report the number of observations contributing to a statistic when it materially affects interpretation.
SECTION 23 OF 34
21. Missing entries and zero values can affect standard deviation differently
Standard deviation describes spread in the numerical observations included in its dataset. Inserting artificial zeros creates new values and can inflate spread when the real measurements are positive. Omitting missing values can also change spread, so the population being described must be clearly stated.
Worked checkpoint. For observed values 4 and 6, the mean is five and the descriptive standard deviation is one. If an unknown third value is incorrectly inserted as zero, the dataset becomes {4,6,0}, with mean 10/3 and standard deviation √56/3≈2.49.
The larger spread does not prove the true original observations were more variable; it shows that the invented zero altered the calculation.
Study action. Inspect the data included in the formula before interpreting σ. The calculator faithfully describes whichever values you supplied.
An actual recorded zero takes the unit of the variable measured, such as zero centimetres, zero minutes or zero dollars. Missingness has no measured numerical value and so cannot be assigned a physical unit as though it were an observation. Clear labels help prevent ambiguous entries.
Worked checkpoint. A traffic counter reports zero vehicles in an interval: that is a real count if the counter worked. A blank because the counter failed says nothing about vehicle count during that interval. A different unit or sensor status would be needed to interpret the record.
The same distinction appears in science practical data, sports scores, attendance records and financial summaries.
Study action. Attach the variable name and measurement status to each observation before any statistical analysis.
A displayed total may summarize only known entries. It should not automatically be treated as the total across every planned category or date. To establish full coverage, the source needs to say which rows and periods are included and how blanks were handled.
Worked checkpoint. If a three-day sales table contains 10, 12 and a blank and displays total 22, that may be a sum of the first two observed values. It does not establish that total sales over three days were exactly twenty-two unless the missing day’s contribution is defined.
A total is therefore a calculation over a stated data scope, not a magical replacement for absent source information.
Study action. Check the interval and row coverage that produced the total before using it in another percentage or mean.
Comparing two groups requires knowing whether they have similar observation coverage and definitions. Different response rates can change the interpretation of group averages even if the calculations on available values are impeccable.
Worked checkpoint. Group A reports ten test scores and Group B reports only three out of ten planned scores. Comparing their observed means may be informative, but the two summaries describe differently complete datasets. The missing seven B values could change the full-group mean substantially.
A numerical difference in averages does not by itself establish a causal difference in learning or performance.
Study action. List the observed and expected counts for both groups before declaring one group stronger.
SECTION 27 OF 34
25. A missing observation cannot be recovered without enough information
Some problems provide totals or means that let an unknown value be solved algebraically. Others do not. A learner should know the difference between calculating an unknown from constraints and inventing an unknown because a blank must be filled somehow.
Worked checkpoint. Four recorded observations are 4,6,8 and x, and their total is stated to be 30. Then x=30−(4+6+8)=12. If no total or mean is supplied, x remains undetermined.
The extra equation supplies information that the blank itself lacks. This is ordinary equation solving, not an arbitrary rule for missing entries.
Study action. Write the known relationship and verify that the recovered x satisfies it before reporting a numerical completion.
Survey questions about overlapping categories may require n(A∪B)=n(A)+n(B)−n(A∩B). That formula assumes the category counts are known for the same universal population. Missing responses or ambiguous inclusion rules can prevent an exact combined total.
Worked checkpoint. A survey reports 12 respondents liking sport A and 9 liking sport B, with 5 known to like both. Among the recorded respondents, the union count is 16. If additional students did not complete the survey, this is not automatically the full-school union.
Overlap correction addresses double-counting, while missing-data treatment addresses absent information. These are distinct issues.
Study action. Define the common population and observation coverage before applying any union formula.
A formula can be correct even when some inputs are unavailable. Reporting an expression in terms of unknowns or a range of possible answers can be more honest than producing an unsupported numerical value. This is especially useful when real-world records are incomplete.
Worked checkpoint. A transportation model calculates total distance D=6+u+4 kilometres for three segments, with the middle segment u not recorded. The known part totals ten, but the full distance remains 10+u. If the context restricts 0≤u≤5, then D lies between ten and fifteen kilometres.
The model explains exactly which missing measurement controls uncertainty. It does not conceal it behind a forced zero.
Study action. Write the unknown symbol and its known bounds, if any, next to the result.
When a question asks for a numerical total, average or prediction that the supplied data cannot determine, the correct mathematical conclusion may be that more information is required. This is not failure to calculate. It is recognising that several different values remain compatible with the known conditions.
Worked checkpoint. From recorded values 4,6 and one unknown x, the full mean is (10+x)/3. If x=0 the mean is 10/3; if x=5 it is 5. Both possibilities fit the supplied incomplete data, so the full mean is not uniquely determined.
A counterexample with two different possible completions is enough to show that no single answer follows from the information given.
Study action. State two permitted completions with different results when explaining why the requested total is underdetermined.
G1 learners can practise distinguishing a reported zero from a blank and calculating means using explicitly observed values. G2 learners can represent unknown entries algebraically and interpret tables. G3 learners can examine how missing observations affect mean, standard deviation, grouped information and claims about samples.
Worked checkpoint. These are mathematical reasoning connections rather than a claim that all missing-data techniques are separate required units at every subject level. The official 2027 SEC K110, K210 and K310 documents list their respective statistics and problem-solving content.
The Clementi teaching reference’s emphasis on diagnosing the first unsupported step applies directly: before correcting arithmetic, confirm whether the underlying number was ever observed.
Study action. Choose the explanation depth appropriate to the learner’s level and current classroom topic.
Identify the variable and population of interest. Mark each recorded numerical observation, each genuine zero, and every missing or not-applicable entry. Determine the actual observed count and the question’s intended denominator. If the full total is unknown, show what can be calculated from observed values and what cannot be claimed.
Worked checkpoint. For the table 4,6,8,10 and a blank, the four-value observed mean is seven. The full five-value mean is (28+x)/5. If the task supplies x, use it; if x is bounded, report a range; if nothing else is known, do not declare a unique five-value mean.
A transparent result names its scope. ‘Mean of available observations’ is distinct from ‘mean of every expected observation’.
Study action. End with the sentence ‘No zero was inserted for a missing observation unless the data source or question explicitly justified that choice.’
Read each situation as a data definition before attempting a number. Name the observed entries, the unknown entries and the intended group. A correct response may be a number, an algebraic expression, a permitted interval or a statement that more information is needed.
Case 1. Data 5,7,blank: observed mean?
Answer: 6.
Reason: Two actual numeric values contribute.
Case 2. Same three entries: known full mean?
Answer: No.
Reason: The third value could change it.
Case 3. Data 5,7,0: full mean?
Answer: 4.
Reason: Zero is a recorded third value.
Case 4. Data 4,6,8,10,blank: observed mean?
Answer: 7.
Reason: 28 divided by four observed values.
Case 5. Same five entries: full mean in terms of x?
Answer: (28+x)/5.
Reason: The fifth result is unknown.
Case 6. If that fifth result is zero, full mean?
Answer: 5.6.
Reason: Only under that additional assumption.
Case 7. Data 6,8,10 from five expected records: observed mean?
Answer: 8.
Reason: 24 divided by three.
Case 8. Is five the denominator for that observed mean?
Answer: No.
Reason: Only three values are supplied.
Case 9. Four results 4,6,8,x total 30: x?
Answer: 12.
Reason: 30 minus known sum eighteen.
Case 10. Three results 4,6,x with full mean five: x?
Answer: 5.
Reason: (10+x)/3=5.
Case 11. Three results 4,6,x with x between 0 and 10: full mean range?
Answer: 10/3 to 20/3.
Reason: Use the minimum and maximum allowed x.
Case 12. Survey: 100 invited, 80 respond, 40 yes. Yes share among respondents?
Answer: 50%.
Reason: 40/80.
Case 13. Same survey: known yes replies as fraction invited?
Answer: 40%.
Reason: 40/100, without classifying nonrespondents.
Case 14. Same survey: may we call the other 60 all no?
Answer: No.
Reason: Twenty invitations had no response.
Case 15. Sizes 5,6,7 have frequencies 2,0,3: total observed count?
Answer: 5.
Reason: Zero frequency contributes no observation.
Case 16. If frequency for size 6 is blank instead of 0, total count known?
Answer: Not necessarily.
Reason: That count is not provided.
Case 17. Rainfall Monday 4, Tuesday blank, Wednesday 6: three-day total known?
Answer: No.
Reason: Known recorded subtotal ten only.
Case 18. Rainfall Monday 4, Tuesday recorded 0, Wednesday 6: total?
Answer: 10 mm.
Reason: Every day has an observed value.
Case 19. Series values at times 0 and 2 are 4 and 8: linear midpoint estimate?
Answer: 6.
Reason: Model-based interpolation, not an observed fact.
Case 20. Does that estimate prove midpoint reading was six?
Answer: No.
Reason: Other paths between endpoints are possible.
Case 21. Data 4,6: descriptive σ?
Answer: 1.
Reason: Mean five and squared deviations one each.
Case 22. Replace missing third value with invented 0: σ of {4,6,0}?
Answer: √56/3≈2.49.
Reason: The extra zero changes the dataset.
Case 23. Table value 0 with frequency 3: zero observations included?
Answer: Yes.
Reason: Three genuine zero-valued observations.
Case 24. Table value 3 with frequency 0: value three observed?
Answer: No.
Reason: The category has no observations.
Case 25. Three days sales 10,12,blank and subtotal 22: complete total 22?
Answer: Not established.
Reason: Subtotal may omit the unknown day.
Case 26. A missing quiz score is automatically a failing zero?
Answer: Only if an explicitly applicable scoring policy assigns that score.
Reason: An unreported score and a graded zero are distinct states.
Case 27. If all missing entries are deleted, is full-population mean guaranteed?
Answer: No.
Reason: Selection of observed cases may not represent missing cases.
Case 28. Can an empty table cell mean N/A?
Answer: Yes, if the legend defines that convention.
Reason: Missing, not applicable and zero have separate meanings.
Case 29. If two permissible missing values produce different full means, is mean unique?
Answer: No.
Reason: Different completions show the result is underdetermined.
Afterwards, classify the first error: confusing a blank with a measured zero, using the wrong denominator, disregarding missing respondents, misreading a frequency table, inventing an interpolation or reporting a full-population conclusion from an observed-only subset. Each mistake requires a different repair. Do not merely repeat the calculation with the same unsupported assumption.
Should a blank cell be counted as zero in a Mathematics average?
Not automatically. Only use zero if the question or data definition says it was an observed zero or a required scoring value. A truly missing observation remains unknown.
Can I calculate the mean of available values?
Yes, by dividing their sum by the number actually observed. Label it as the observed-value mean; it need not equal the mean of every intended observation.
What is the difference between zero frequency and missing frequency?
Zero frequency says a category occurred no times. Missing frequency means the count has not been supplied unless the table defines a special convention.
Is filling missing entries with the observed mean always correct?
No. Mean imputation is an analytical choice and does not establish the true unobserved values. It can also distort variation.
What if the missing value could be anywhere in a known interval?
Represent it by a variable and derive lower and upper bounds for the requested quantity. The result may be a range rather than one number.
Is missing-data treatment assessed identically in all G1, G2 and G3 Mathematics syllabuses?
No. These are cross-level data-reading and reasoning habits. Consult the specific SEAB K110, K210 or K310 syllabus and current school programme for required topics.
For official assessed content consult SEC 2027 G1 Mathematics K110, G2 Mathematics K210 and G3 Mathematics K310. For connected teaching read the complete Mathematics explainer directory and the Clementi Mathematics teaching reference.