2027 G3 SEC Mathematics K310 | Learner’s Guide Vol 0198. A histogram may look like a row of rectangles, but its bars represent a distribution, not simply a collection of named categories. Students often know how to add the frequencies and still misread the class boundaries, invent individual values or turn a large bar into a false claim about a whole population. This chapter is about reading equal-width histograms so carefully that the calculation, the comparison and the final sentence agree.
All school-event data in this guide are invented for mathematical practice. The measurements are not records of any actual students and do not imply medical or fitness outcomes. Questions and solutions are original teaching examples, not official SEAB examination questions.
1. The precise 2027 syllabus target
The SEAB 2027 SEC G3 subject directory, last updated 26 October 2026, lists Mathematics as K310, replacing the former reference code 4052. The official K310 syllabus covers statistical diagrams including histograms with equal class intervals, cumulative-frequency diagrams, box-and-whisker plots, grouped means and questions about misleading displays. This guide stays within that stated histogram scope. Variable-width frequency-density techniques are not presented here as a required K310 skill.
Paper 1 and Paper 2 each last 2 hours 15 minutes, carry 90 marks and contribute 50% each. An approved calculator may be used in both. The syllabus expects clear essential working; omitting it can lose marks. Non-exact numerical answers generally use three significant figures, and degree angles one decimal place, unless the question specifies otherwise. The final problem in Paper 2 focuses on mathematics in a real-world situation. A histogram can therefore form part of a larger decision problem, not merely a stand-alone graph-reading task.
Earlier Vol 0194 dealt with upper-tail and percentile claims. Vol 0190 handled standard deviation. Here the distinctive task is to understand a frequency distribution presented in touching equal-width rectangles, reconstruct its table and state what can and cannot be concluded.
2. Three questions to ask before any calculation
First: what is measured on the horizontal axis? It may be journey duration, packet mass, distance, temperature or another numerical quantity. Identify its unit. Second: what does the vertical axis measure? In the equal-interval histograms studied here, the bar heights show frequencies when the vertical axis is labelled frequency. Third: how wide is each group, and which boundaries belong to which interval? Without those answers, a bar that looks tall can conceal a wrong category or an off-by-one count.
A histogram groups numerical observations into consecutive intervals. The bars touch because the scale is continuous and adjacent intervals share a boundary. A bar chart of club names has separate named categories and normally leaves gaps between bars. That visual distinction is useful but not decisive if a printer has drawn the diagram badly; the axis labels and meaning of the data control interpretation.
Use the chart only after reading its heading and any note about grouping. If the horizontal scale says ‘minutes’, do not answer in ‘students’ when asked for a time. If the vertical scale says ‘number of sessions’, do not describe the frequencies as numbers of people unless the data collection states one session per person.
3. A complete worked distribution
Imagine a school counts the durations of forty fictional lunchtime activity sessions, to the nearest minute, and groups them into equal-width intervals of ten minutes. The table is: 10–under 20 minutes: 4 sessions; 20–under 30: 10; 30–under 40: 12; 40–under 50: 9; 50–under 60: 5. The intervals are written in half-open form to avoid counting a boundary observation twice. An observation of exactly 30 minutes belongs to the 30–under 40 interval, not the preceding group.
The frequency total is 4 + 10 + 12 + 9 + 5 = 40. All five intervals have the same width, ten minutes. A text description of the corresponding histogram would be five touching rectangles across those consecutive intervals, with heights 4, 10, 12, 9 and 5 on a vertical axis labelled frequency. You do not need a photograph or illustration to check the mathematical structure.
From the distribution we know the most frequent interval is 30–under 40 minutes, with twelve sessions. We cannot claim the most common individual duration was exactly 35 minutes; a grouped table does not show exact individual values. We also cannot infer that any participant completed exactly 59 minutes. The bar says five observations lie somewhere in the last interval, no more.
4. Frequency, relative frequency and proportion
Frequency is the number of observations in a group. Relative frequency is its share of the total: group frequency divided by total frequency. For the 30–under 40 group, relative frequency is 12 ÷ 40 = 0.30, or 30%. For the 10–under 20 group it is 4 ÷ 40 = 0.10, or 10%. These statements are about the recorded sessions, not the whole school.
The relative frequencies for the five groups are 0.10, 0.25, 0.30, 0.225 and 0.125. Their sum is 1, as it should be: 0.10 + 0.25 + 0.30 + 0.225 + 0.125 = 1. You can use this as a numerical self-check. If your proportions add to 0.95, you have lost a group or made a calculation error.
Keep number of sessions separate from number of participants. If one person contributed three sessions, forty records need not mean forty people. In an examination question, a dataset’s unit of observation matters. The mathematics may be unchanged, but an interpretation such as ‘twelve pupils walked between thirty and forty minutes’ is unjustified unless each record corresponds to a different pupil.
5. The difference between a mode and a modal class
The word ‘mode’ usually names the most frequent individual observation in ungrouped data. A histogram shows grouped frequencies; it allows identification of a modal class, the interval with the largest count. Our fictional modal class is 30–under 40 minutes. The word ‘modal’ does not turn the interval into one exact value.
If someone writes ‘the mode was 35 minutes’, they have used the class midpoint as if it were a measured result. The midpoint is useful in an estimated grouped mean, but it is not automatically the mode. There could be twelve different durations in that interval. The modal class would remain unchanged while the exact mode might change or not exist.
Describe ties accurately. If two intervals both have frequency twelve and every other interval has fewer than twelve, they are joint modal classes. Do not choose whichever bar looks marginally taller in a poorly reproduced diagram. Read the stated scale and the count.
6. Reconstruct a table from a histogram
When an examination question gives only a histogram, begin by reproducing class intervals and frequencies in a small table. Do not calculate the mean directly from a picture without first deciding which observations each bar represents. A six-row scratch table needs columns for interval, frequency and midpoint. Add cumulative frequency if the question later asks ‘at most’ or ‘more than’.
Do a sum check: frequencies must add to the stated total if all observations are included. If the question says fifty observations and your reconstructed bars add to forty-eight, inspect the vertical reading or a missing class. If the question gives no total, calculate it from the bars rather than assuming a convenient number.
Suppose a frequency axis is labelled 0, 2, 4, 6 and a rectangle reaches halfway between 4 and 6. The frequency is 5, provided the scale is linear. But if a bar falls between two ticks and the diagram lacks detail, avoid inventing an exact count beyond what is reasonably readable. Examination diagrams should be interpreted using their printed scales, not estimated by the thickness of a line.
7. Find a grouped mean without pretending it is exact
The grouped mean uses each interval midpoint as a representative value. The midpoints of our five ten-minute intervals are 15, 25, 35, 45 and 55 minutes. Multiply each midpoint by its frequency: 15 × 4 = 60; 25 × 10 = 250; 35 × 12 = 420; 45 × 9 = 405; 55 × 5 = 275. The sum of the products is 1410.
Divide by the total frequency forty: estimated mean = 1410 ÷ 40 = 35.25 minutes. The calculation is correct for the midpoint model. The actual mean of the raw, ungrouped observations might differ because their individual values are unknown. Write ‘estimated mean’ when you use grouped midpoints and the question calls for an estimate.
Checking units helps. We have minutes times session count in the numerator and session count in the denominator, leaving minutes. This is not a claim about healthy exercise duration; it is a numerical description of an invented distribution. The useful habit transfers to grouped mass, temperature and journey-time questions.
8. A median class is not necessarily a median value
For forty ordered observations, the median lies between the twentieth and twenty-first values. Cumulative frequencies for our five groups are 4, 14, 26, 35 and 40. Thus both middle observations fall in the 30–under 40 group. We know the median class, but not the precise median of the original data.
Do not substitute the midpoint 35 minutes and declare the exact median. If the question explicitly requests an estimate from a cumulative-frequency graph, the method and expected precision may be different; follow the question’s instruction. This volume’s point is the distinction between what grouped data settles and what it leaves open.
The same issue appears with quartiles. If the relevant rank lies within a particular class, you know the class containing that rank. You do not magically recover the actual observation’s value from a bar height. Vol 0103’s cumulative-frequency workshop explains the separate graph-estimation method.
9. Estimate rather than invent a boundary count
Question: ‘How many of the forty sessions lasted at least thirty minutes?’ The 30–under 40, 40–under 50 and 50–under 60 groups contain 12 + 9 + 5 = 26 sessions. The threshold aligns with a class boundary, so the answer is exact from the grouped table: twenty-six of the recorded sessions, or 65%.
Now change the question: ‘How many lasted at least thirty-five minutes?’ That threshold cuts through the 30–under 40 group. The histogram does not tell us how its twelve observations are distributed inside that interval. We can state that fourteen sessions were in the groups beginning at forty minutes, while twelve more belong to the partly included group. Depending on the actual data, the count meeting the threshold lies somewhere from fourteen to twenty-six.
Any narrower estimate needs an explicit additional assumption, such as approximately uniform distribution inside the interval. That is an approximation, not an exact reconstructed count. An examiner may give a method for interpolation; if not, be careful about the certainty of your claim.
10. Compare two distributions only when the bases fit
Consider a second fictional set of forty sessions with the same five equal intervals and frequencies 2, 6, 18, 10 and 4. Its modal class is also 30–under 40 minutes, but its tallest bar is eighteen, not twelve. The second distribution is more concentrated in that central interval, although the extent of overall spread requires careful comparison.
Because both samples have forty sessions and identical class widths, their frequencies can be compared directly. For example, 18 of 40 versus 12 of 40 were in the 30–under 40 interval: 45% versus 30%. That is a fifteen-percentage-point difference in these recorded datasets, not a causal effect of any teaching or health programme.
If the first group had only twenty sessions, a bar of twelve would represent 60%, not 30%. Simply comparing raw bar heights between samples of unequal size can mislead. Convert to proportions when the question asks which group has a larger share, and name the sample base.
11. How histogram claims become misleading
A poster might claim, ‘Almost everyone stayed longer than thirty minutes.’ Our original table says twenty-six of forty lasted at least thirty minutes, or 65%. That is a majority, but whether ‘almost everyone’ is appropriate is a matter of careful language; 65% does not support a claim of near-universality. Prefer the count or proportion.
A display can also be misleading if the scale is inconsistent or if neighbouring intervals have unequal widths but are presented as though equal. The K310 syllabus explicitly names histograms with equal class intervals. When studying an equal-width example, verify that the horizontal scale actually preserves the widths before reading frequency from height. If the question provides unequal widths, do not quietly apply the equal-width method; recognise that it may require ideas beyond this specific syllabus scope.
Other traps include omitting the total sample size, relabelling ‘sessions’ as ‘students’, changing the unit, zooming a small variation into a dramatic headline or using grouped midpoints as though they were raw measurements. Each changes the story without changing the numbers. The mathematical cure is to record definitions before interpreting the picture.
12. Draw an equal-width histogram accurately
Start with the numerical class intervals on the horizontal axis. Mark every boundary in the right numerical order. Give the horizontal axis a quantity and unit such as ‘Session duration (minutes)’. On the vertical axis, label ‘Frequency’ and choose an even numerical scale large enough for the highest frequency, twelve. Draw touching rectangles whose widths correspond to the equal class widths and whose heights are 4, 10, 12, 9 and 5.
Do not draw gaps if consecutive intervals cover a continuous numerical variable. Do not give the rectangles unrelated widths or stretch the 30–40 bar merely because its frequency is largest. Its height represents twelve; its width still represents ten minutes. If working on graph paper, use a ruler and show values clearly.
After drawing, ask a classmate to recover the original frequency table from the graph alone. If they cannot read the endpoints or the height of a bar, improve the labels. A graph succeeds when a reader can reproduce its mathematical meaning, not when its colours look attractive.
13. Reverse-engineer an unknown frequency
Suppose a fictional histogram has five equal groups with frequencies 4, 10, x, 9 and 5, and you are told that there are forty sessions in total. The known frequencies add to twenty-eight, so x = 40 − 28 = 12. This is not guessing the missing bar from visual similarity; it is using the total count.
Now imagine the information says that the third interval represents 30% of forty observations. Then x = 0.30 × 40 = 12. The two methods agree. If instead a headline claims that the third class has half of all observations, check the count: twelve out of forty is 30%, not 50%.
A more challenging variant says the last two classes together contain fourteen sessions. Since the fourth group has nine, the fifth must have five. Again, the arithmetic is straightforward once the phrase ‘last two’ is mapped to the right intervals.
14. How to answer ‘explain’ questions after the calculation
An explanation should state both the quantity and its implication. ’26’ alone answers a count but not necessarily a question asking whether a claim is supported. A stronger response is: ‘Twenty-six of forty sessions lasted at least thirty minutes, which is 65%; therefore the data support a majority but do not establish that nearly every session met the threshold.’
Keep observed patterns separate from causes. If one group has longer session durations, a histogram cannot on its own show that a new timetable caused the change. Duration distributions may differ because of how participants were selected, because conditions changed or because the number of observations differs. This is mathematical reasoning about evidence, not a demand to invent hidden explanations.
When a conclusion needs a comparison, use the same statistic for both groups. Compare 30–under 40 shares with 30–under 40 shares, not one group’s count with the other’s percentage. Then state what the numbers allow you to conclude.
15. Worked practice: twenty distinct interpretation decisions
Attempt each exercise before reading its answer. These are original examples using fictional data. Some deliberately contain questions with insufficient information; spotting that is part of mathematical performance.
Practice 1. A class has four sessions in 10–under 20 and ten in 20–under 30. How many are under thirty? Worked answer: Fourteen sessions. Both groups lie below thirty; no assumptions about exact values inside them are needed.
Practice 2. Is an observation of exactly thirty in 20–under 30? Worked answer: No. The upper boundary is excluded by the half-open interval. It belongs to 30–under 40.
Practice 3. The bars have heights 4, 10, 12, 9 and 5. Find the total. Worked answer: Forty. Add all group frequencies once, checking that adjacent intervals do not overlap.
Practice 4. Find the relative frequency of the last group. Worked answer: Five divided by forty is 0.125, or 12.5%. Label it as a share of the observed sessions.
Practice 5. Which is the modal class? Worked answer: 30–under 40 minutes, frequency twelve. That does not determine an exact mode.
Practice 6. Could the raw median be 38 minutes? Worked answer: Yes. Both central observations are in the 30–under 40 class; a value such as 38 is possible but not established.
Practice 7. Is the raw median exactly 35 minutes? Worked answer: Not necessarily. Thirty-five is a midpoint, not evidence of the individual central observations.
Practice 8. How many sessions lasted at least forty minutes? Worked answer: Nine plus five equals fourteen, or 35% of forty.
Practice 9. How many lasted at least thirty minutes? Worked answer: Twelve plus nine plus five equals twenty-six, or 65% of forty.
Practice 10. Can we determine exactly how many lasted thirty-seven minutes? Worked answer: No. The grouped frequency does not expose the count of any individual duration inside its interval.
Practice 11. What is the grouped mean estimate? Worked answer: Midpoint-product sum 1410 divided by total forty gives 35.25 minutes.
Practice 12. Is the grouped mean guaranteed to equal the raw mean? Worked answer: No. Midpoints represent whole classes; individual observations may sit at different values.
Practice 13. The vertical frequency scale marks 0, 5, 10 and 15. A bar reaches 10. What is its frequency? Worked answer: Ten, assuming the axis really is frequency and the scale is linear.
Practice 14. Two samples have 40 and 20 observations; each has a bar of 10. Same share? Worked answer: No. Ten of forty is 25%; ten of twenty is 50%.
Practice 15. A chart changes its axis label from ‘number of sessions’ to ‘number of people’. Is the conclusion safe? Worked answer: Only if the data guarantee one session per person. Otherwise it misstates the unit observed.
Practice 16. A frequency of twelve is called ‘twelve percent’. What is wrong? Worked answer: A frequency is a count. Here twelve corresponds to 30% of forty, not 12%.
Practice 17. What fraction is in the 20–under 30 group? Worked answer: 10/40 = 1/4. It is also 25%.
Practice 18. Could 26 sessions of length at least 35 be true? Worked answer: It is possible if all twelve observations in the 30–under 40 class meet 35, but it is not certain.
Practice 19. Could only 14 sessions meet the threshold of 35? Worked answer: Yes, if none of the twelve observations in 30–under 40 reached 35. The grouped table does not rule it out.
Practice 20. What is a valid summary of the original distribution? Worked answer: The largest frequency, twelve of forty, lies in the 30–under 40-minute group; this is a grouped observation, not an individual-value claim.
Practice 21. Why should histogram bars touch? Worked answer: They represent adjoining numerical intervals of a measured quantity; the boundary of one is the boundary of the next.
Practice 22. What does the equal class width allow us to read here? Worked answer: With a frequency vertical axis, rectangle height is directly comparable with frequency because every horizontal interval has the same width.
Practice 23. The question asks for a health benefit shown by the histogram. Can you calculate one? Worked answer: No. The histogram records durations, not health outcomes. A conclusion about benefit needs relevant measurements.
Practice 24. The last interval contains five observations. Can you say all lasted fifty-five minutes? Worked answer: No. Fifty-five is its midpoint. The five records may have different durations between fifty and sixty.
16. Longer applications for examination transfer
Application A: change the unit. A different dataset records walking-distance intervals in metres rather than session duration in minutes. Rebuild axis labels accordingly, then interpret one bar. The calculations can look identical while the interpretation changes completely. This is a reason to attach the unit to every final sentence.
Application B: unequal sample size. Group A has 24 observations and Group B has 40. Each group records twelve values in the same class. Group A’s share is 50%; Group B’s is 30%. Write a two-sentence comparison naming both denominators. Explain why identical heights in two frequency histograms do not imply equal relative frequency.
Application C: top-end threshold. A group asks how many original sessions lasted at least 45 minutes. The threshold cuts through the 40–under 50 class, so you cannot know the exact total from grouped frequencies. At least the five sessions in the final class meet the threshold; at most those five plus nine in the overlapping class do. State the range five to fourteen and the assumptions needed for a narrower estimate.
Application D: misleading claim. A fictional poster says ‘three times as many students attended’ because its tallest bar increased from six to eighteen between periods. What else must be checked? Start with whether the bar is the same duration interval, whether both samples count sessions or students, whether totals are equal, and whether the groups were selected comparably. A histogram gives distributional facts but cannot by itself answer every causal question.
Application E: repair the table. A learner writes the groups as 10–20, 20–30, 30–40 and so on, without explaining which group contains observations at the boundaries. Rewrite using 10–under 20, 20–under 30 and so forth. The purpose is to make double-counting impossible.
17. A reliable problem-solving checklist
Before reading: identify numerical quantity, unit, interval boundaries, interval width, vertical meaning and total sample. During calculation: copy each group, add frequencies, convert to proportions only when needed, use midpoint estimates only when justified and watch thresholds that cut inside intervals. After calculating: ask whether you have an exact count, an estimate, a bound or insufficient information.
A simple diagnostic routine is to produce a sentence with three parts: quantity, base and qualification. ‘Twelve sessions out of forty were between thirty and forty minutes in this invented sample.’ Each part stops a common error. A calculation without the relevant base is vulnerable to misinterpretation.
For students who struggle with grouped data, return to the PSLE learning guide for frequency-table and graph-reading foundations, then the Secondary 4 mathematics grouped-data owner. Use Vol 0099 when you need to check a deceptive axis; this volume adds the separate grouped-frequency histogram reasoning.
18. Questions students and tutors often ask
Is a histogram the same as a bar chart? No. A histogram groups a numerical variable into consecutive intervals; a bar chart typically compares separate categories. The meaning of each axis matters more than decorative appearance.
Can histogram data determine the exact raw mean? No, not when only grouped frequencies remain. Midpoint-based calculations give a grouped estimate. Individual observations would be needed for the exact raw-data mean.
Do I have to learn frequency density for unequal interval widths? The official 2027 K310 syllabus specifically lists histograms with equal class intervals. Study the published scope and any clarification from your school rather than treating a more advanced technique as automatically examinable.
Why does an answer sometimes say ‘cannot determine’? Because a class interval preserves only a group count. A threshold inside a group or an exact individual value may be invisible in the available evidence.
Can a histogram prove one activity is better for health? Not by itself, especially when its variable is duration or attendance. A fair claim needs a relevant outcome and adequate study design.
What should a tutor mark first? The interpretation of class boundaries and observed unit. Once those are correct, arithmetic and conclusions become much more reliable.
19. The deeper lesson
A histogram is a compact model of many observations. Its value comes from revealing broad patterns quickly; its cost is the loss of individual detail. Learn to enjoy that trade-off rather than fight it. The rectangles tell you something useful about frequency, typical intervals and spread, but they do not authorise any story you wish to tell about the people behind the measurements.
Progress from reading a frequency to reconstructing a table; from table to total; from total to relative frequency; from frequency to an estimated grouped mean; and from calculation to a cautious conclusion. That path is easier to learn and harder to misuse than a collection of memorised rules. For official content boundaries, check the 2027 K310 syllabus linked above.
