Sec 3 G3 Maths Tutor Bukit Timah

Secondary 3 G3 Mathematics Tutor Bukit Timah | Histograms, Stem-and-Leaf Diagrams and Misleading Data — Did you know? The hardest part of some Secondary 3 Mathematics questions is deciding which relationship belongs on the page. A learner may follow a worked solution perfectly and then freeze when the next problem looks different. This quick-first guide explains the underlying idea, offers checked practice and helps parents see the difference between copied steps and genuine independent progress.

The quick answer, before the long explanation

Frequency is a count; total frequency is the number of observations.
With equal-width classes and a frequency vertical axis, histogram bar heights show frequencies.
A stem-and-leaf diagram needs a key and preserves actual values.
Grouped-data means calculated using midpoints are estimates.
Check graph axes, class widths and whether visual differences accurately represent numeric differences.

This Secondary 3 G3 Mathematics tutor Bukit Timah article is written for readers who want clear explanations of a topic in the official Singapore–Cambridge SEC 2027 G3 Mathematics syllabus K310. K310 describes the full secondary examination course, not a guarantee that all schools teach this content in the same Secondary 3 term. Use the student’s latest school scheme of work alongside the current SEAB syllabus index.

Jump to: How it works · 50 checked examples · Mistakes to fix · Six-week plan · Parent questions.

What the mathematics is actually telling you

The best chart depends on the question

Did you know two correct charts can answer different questions from the same information? A bar chart is useful for comparing separate categories. A line graph helps reveal change over time. A histogram for equal-width class intervals describes a numerical distribution across continuous intervals. A stem-and-leaf diagram keeps individual values visible while organising them.

The SEC K310 G3 Mathematics syllabus includes bar graphs, line graphs, pictograms, pie charts, dot diagrams, histograms with equal class intervals, stem-and-leaf diagrams, cumulative frequency diagrams and box-and-whisker plots. This article concentrates on choosing, reading and checking basic displays, with special attention to equal-width histograms and stem-and-leaf notation.

Frequency measures how often, not how large

In a frequency table, each row or class has a count of observations. The total frequency is the total number of data points, not the sum of the measured values. To find a weighted mean from a frequency table, multiply each value by its frequency, add the products and divide by the total frequency.

In a grouped table, an estimated mean uses class midpoints to represent all values in each class. It is an estimate because the exact observations within each interval are not known. Do not report the midpoint-based result as an exact mean of the unobserved raw data. This distinction is a useful mark of statistical understanding.

Histograms with equal intervals are readable by frequency

A basic SEC G3 histogram question using equal class widths can be read from frequencies represented by the bar heights when the vertical scale is frequency. Adjacent numerical intervals touch because they describe a continuous axis; gaps in a bar chart often separate categories. Confirm both the class limits and the vertical-axis label before reading a bar.

The K310 content description specifies histograms with equal class intervals. That is why this article does not present unequal-width frequency-density histograms as a required feature of the syllabus. If a school enrichment activity uses unequal intervals, the interpretation must instead respect the way area and frequency are related; do not transfer an equal-width reading habit without checking.

Stem-and-leaf diagrams preserve the numbers

A stem-and-leaf diagram splits a value into leading digits and a final leaf. With the key 2|4 = 24, the stem 2 and leaf 4 mean twenty-four. A row 2 | 1 3 6 represents 21, 23 and 26. The leaves should usually be ordered for quick interpretation.

Unlike a grouped histogram, a stem-and-leaf display still lets you reconstruct each actual observation. This makes it easy to find the smallest and largest values, count observations and locate a median when the data are ordered. A key is essential: without it, 2|4 could be read as 2.4 rather than 24.

Cumulative frequency answers how many up to here

Cumulative frequency adds the frequencies from the lowest group through the current boundary. It should not decrease as you move right through successive increasing upper boundaries. A cumulative frequency diagram can help estimate medians and quartiles for a grouped distribution, but the values are estimates derived from the display.

For a simple three-group table with frequencies 4, 7 and 3, the cumulative totals are 4, 11 and 14. The first 11 observations fall within the first two groups. The final cumulative total must equal the number of observations, providing an immediate consistency check before any plotted curve is drawn.

Axes can make a small difference look enormous

If two categories have values 92 and 96, the genuine absolute difference is four. A bar chart that starts at 90 rather than zero may make the second bar look several times as tall as the first, even though their actual values differ by only about four percent. The visual impression depends on the chosen axis range.

A truncated axis is not always forbidden, particularly in specialist line charts, but it must be visible and justified. For simple bar comparisons, including zero often provides a more faithful magnitude comparison. Look at the numerical tick labels rather than believing the visual height alone.

Mean and median react differently to an outlier

Consider scores 4, 5, 6, 7 and 28. The median is the middle value, six. The mean is ten because the large final observation pulls the arithmetic average upwards. Neither statistic is automatically ‘wrong’, but they describe different aspects of the distribution. Choose the one suited to your question.

If one student has a much higher value than the rest, the mean can be sensitive to that exceptional value. The median is more resistant. A good report can state both and describe the spread rather than announcing that one single measure tells the whole story.

A statistical answer needs interpretation

A correct calculation can still be misleading if the conclusion overstates what the data show. A frequency table for one class does not automatically describe every school. A line graph of sales over a few weeks does not establish why the sales changed. A chart showing association is not proof of causation.

For a learner, the important habit is to identify what the data actually represent, which graph or statistic addresses the question, and what conclusion remains uncertain. This mirrors the K310 emphasis on interpreting data and explaining why a representation may mislead readers.

A tiny diagnostic before more worksheets

Choose three questions. The first is an ordinary example, the second looks similar but contains a feature that changes the method, and the third is embedded in a real-world description or unfamiliar diagram. Ask the student to say what is given, which quantity is unknown, and what a correct answer should represent. These three decisions reveal more about the actual difficulty than a long list of final answers alone.

A useful tutor identifies whether the obstacle lies in reading the question, remembering a prerequisite, choosing between methods, carrying out arithmetic or checking results. Each obstacle needs a different response. Some students need a clearer definition, others a worked contrast, and others a disciplined checking routine. Repeating the same comfortable exercise fifty times is rarely an effective substitute for that diagnosis.

The SEC 2027 K310 assessment objectives emphasise techniques, problem-solving in varied contexts, and mathematical reasoning and communication. Students therefore benefit from writing reasons and interpreting results, not merely arriving at a number. A strong answer makes its assumptions visible, keeps notation coherent and can be understood by another reader. The worked exercises below include both reasoning and a separate check for that purpose.

50 worked exercises and independent checks

These are original learning exercises, not an official examination paper or predictions of upcoming questions. Do not attempt all fifty at once. Choose a short set based on the exact mistake you need to repair; read the question before revealing its method and check. After one or two accurate attempts, mix different clusters so the student must select a method without a chapter heading giving the answer away.

A frequency table and its mode

Here the focus is a frequency table and its mode. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 1. Question: Three categories A, B and C have frequencies 2, 5 and 3. Find the total and modal category. Method: Total frequency = 2+5+3 = 10 observations. Category B has the largest frequency 5, so the mode is B. Check: The total counts observations, not a sum of category names or category scores. Watch out: A mode is the most frequent category, not the numerical average of counts.

Try it next: Give a real-world example where the three categories might be transport choices. Try hiding the finished working and saying the first mathematical decision aloud. If the method is only recognisable with the example visible, return to the definition before increasing speed. Build a second question with the same relationship but different numbers, and ask another student to explain its opening step.

Mean from an ungrouped frequency table

Here the focus is mean from an ungrouped frequency table. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 2. Question: Scores 1, 2 and 3 have frequencies 2, 3 and 1. Find their mean. Method: Weighted total = 1×2 + 2×3 + 3×1 = 11. Total observations 6; mean = 11/6 ≈ 1.833. Check: The mean lies between the minimum value 1 and maximum value 3. Watch out: Do not divide the weighted sum by the three distinct scores; divide by total frequency.

Try it next: Calculate the modal score and explain why it need not equal the mean. A numerical check can confirm arithmetic, but a stronger check also asks whether the result has the correct meaning. Name the variable, unit or quantity shown by your answer. Compare an exact result with a sensible approximation. When do you need the exact form, and when does rounding help communication?

Estimate a mean from grouped data

Here the focus is estimate a mean from grouped data. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 3. Question: Equal-width score classes 0–<10, 10–<20 and 20–<30 have frequencies 2, 4, 1. Estimate the mean using class midpoints. Method: Class midpoints are 5, 15, 25. Estimated total = 5×2+15×4+25×1 = 95; divide by 7 observations. Estimated mean ≈ 13.57. Check: The estimate lies between 0 and 30 and respects the frequency weighting. Watch out: This is an estimate, since actual individual scores inside each class are not supplied.

Try it next: Explain why the midpoint method would become unnecessary if the original raw scores were available. Record the first mistake, not merely the final incorrect total. A useful error log states which rule was misapplied and what evidence will prevent repetition. Sketch a diagram or quick table that makes the relationship visible without changing what the question says.

Read an equal-width histogram

Here the focus is read an equal-width histogram. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 4. Question: An equal-width histogram has adjacent classes 0–<5, 5–<10 and 10–<15. Its vertical axis is frequency, with bar heights 3, 6, 4. Find modal class and total frequency. Method: All class widths are five units, so the frequency-labelled heights give counts directly. The middle class 5–<10 has largest frequency 6; total frequency = 3+6+4 = 13. Check: The largest frequency is indeed 6 and the sum of all three frequencies is 13. Watch out: Check the actual vertical-axis label. An unequal-width histogram needs additional care not required for this equal-width example.

Try it next: Compare why histogram bars touch but separate category bars may have gaps. After practising the example, change one given value and predict what should happen before recalculating. Transfer to new numbers matters more than repeated copying. Explain the most tempting incorrect method and identify the first reason it fails.

Read an ordered stem-and-leaf diagram

Here the focus is read an ordered stem-and-leaf diagram. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 5. Question: A stem-and-leaf display has key 2|1 = 21. Rows are 2|1 3 6 and 3|1 4 7 9. State the median and number of values. Method: It represents 21, 23, 26, 31, 34, 37, 39. Seven observations have a fourth (middle) value 31, so median = 31, and there are 7 values in all. Check: Reading the key reconstructs the original two-digit numbers. Watch out: Do not interpret 2|1 as a decimal unless the key says so.

Try it next: Identify the smallest and largest observations from the two ordered rows. Set the exercise beside a different-looking question. Explain which feature tells you whether the two need the same technique or a different one. Make a one-sentence rule that will still work when the problem is presented in an unfamiliar story.

Cumulative frequency totals

Here the focus is cumulative frequency totals. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 6. Question: Consecutive increasing classes have frequencies 4, 7, 3. Give cumulative totals after each class. Method: The running totals are 4, 4+7=11, then 11+3=14. The final value 14 is the total number of observations. Check: The cumulative values 4, 11, 14 never decrease. Watch out: Cumulative frequency is a running count, not a fresh isolated bar frequency.

Try it next: Explain how a cumulative curve could locate roughly half the observations. Ask whether a shortcut depends on a condition such as equal intervals, regularity, a nonzero coefficient or a particular unit. Conditions are part of mathematics. Check whether the same calculation can be reversed to recover the given information.

Spot a truncated-axis bar-chart trap

Here the focus is spot a truncated-axis bar-chart trap. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 7. Question: Two categories have values 92 and 96. A bar chart starts the vertical axis at 90. What numerical comparison should be stated despite the visual exaggeration? Method: The true absolute difference is 96−92 = 4. Relative to the first value, this is 4/92×100 ≈ 4.35%. A baseline starting at 90 makes the visual bar lengths disproportionately different. Check: The values are both near a hundred; the numbers, not the cropped appearance, show a modest difference. Watch out: Do not infer percentage increase by comparing the lengths of bars on a truncated vertical scale.

Try it next: Explain when a clear axis break and detailed labels may be useful, and why the chart still needs caution. A clear solution should be readable by another learner. Give reasons beside key steps, include any necessary bracket and label the final quantity or conclusion. State which pieces of the answer are facts and which are assumptions about the situation.

Pie-chart angle from frequency

Here the focus is pie-chart angle from frequency. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 8. Question: A survey has category frequencies 5, 7 and 8. What pie-chart angle represents the first category? Method: Total is 20. The first category’s fraction is 5/20 = 1/4. Its sector angle is ¼×360° = 90°. Check: All pie-chart sectors together must make 360°, and the first is exactly one-quarter. Watch out: Do not multiply a frequency directly by 360° without dividing by the total first.

Try it next: Work out the other two angles and confirm their sum with 90° is 360°. Return to this idea after two days with no worked example in sight. Independent retrieval after a delay shows which parts of the reasoning are truly accessible. Imagine the examiner gives the final number but asks for justification; decide what working would demonstrate the method.

Mean versus median with an outlier

Here the focus is mean versus median with an outlier. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 9. Question: A sorted five-value data set is 4, 5, 6, 7, 28. Find mean and median. Method: The middle (third) value is 6, so median = 6. Total is 50; dividing by five gives mean = 10. The very large last value raises the mean relative to the median. Check: Check 10 is between the smallest 4 and largest 28, while median is the central ordered value. Watch out: An extreme value can pull the mean; do not claim the median must move by the same amount.

Try it next: Replace the largest value by 8 and compare the two means. Try working backwards from the answer to see whether the original condition is recovered. If a real-world result is impossible, check the modelling assumptions. Identify one place where an incorrect sign, unit or label could reverse the interpretation of the answer.

Compare two groups without overclaiming

Here the focus is compare two groups without overclaiming. Before calculating, write the quantity the question requests and say why the first step is justified. A student who can explain that choice is better prepared for unfamiliar wording than one who relies on recognising a repeated page layout.

Worked example 10. Question: Data set A is 2, 2, 3, 4, 8 and B is 2, 3, 3, 4, 5. Compare their medians and ranges. Method: Both sets have median 3. A has range 8−2 = 6, while B has range 5−2 = 3. So typical middle value matches, but A is more spread by the range measure. Check: Each median is the third of five ordered values; each range is maximum minus minimum. Watch out: Same median does not imply identical spread or shape.

Try it next: Explain how an extreme observation could affect a mean without changing the median. Look at the scale, size or sign of the result before trusting a calculator. A rough estimate often catches a mathematically impossible or mislabelled output. Try the problem again after a short break and compare whether fewer hints or corrections were needed.

A frequency table and its mode

Worked example 11. Question: Three categories A, B and C have frequencies 3, 6 and 4. Find the total and modal category. Method: Total frequency = 3+6+4 = 13 observations. Category B has the largest frequency 6, so the mode is B. Check: The total counts observations, not a sum of category names or category scores. Watch out: A mode is the most frequent category, not the numerical average of counts.

Mean from an ungrouped frequency table

Worked example 12. Question: Scores 1, 2 and 3 have frequencies 3, 4 and 2. Find their mean. Method: Weighted total = 1×3 + 2×4 + 3×2 = 17. Total observations 9; mean = 17/9 ≈ 1.889. Check: The mean lies between the minimum value 1 and maximum value 3. Watch out: Do not divide the weighted sum by the three distinct scores; divide by total frequency.

Estimate a mean from grouped data

Worked example 13. Question: Equal-width score classes 0–<10, 10–<20 and 20–<30 have frequencies 3, 5, 2. Estimate the mean using class midpoints. Method: Class midpoints are 5, 15, 25. Estimated total = 5×3+15×5+25×2 = 140; divide by 10 observations. Estimated mean ≈ 14.00. Check: The estimate lies between 0 and 30 and respects the frequency weighting. Watch out: This is an estimate, since actual individual scores inside each class are not supplied.

Read an equal-width histogram

Worked example 14. Question: An equal-width histogram has adjacent classes 0–<5, 5–<10 and 10–<15. Its vertical axis is frequency, with bar heights 4, 7, 5. Find modal class and total frequency. Method: All class widths are five units, so the frequency-labelled heights give counts directly. The middle class 5–<10 has largest frequency 7; total frequency = 4+7+5 = 16. Check: The largest frequency is indeed 7 and the sum of all three frequencies is 16. Watch out: Check the actual vertical-axis label. An unequal-width histogram needs additional care not required for this equal-width example.

Read an ordered stem-and-leaf diagram

Worked example 15. Question: A stem-and-leaf display has key 3|1 = 31. Rows are 3|1 3 6 and 4|1 4 7 9. State the median and number of values. Method: It represents 31, 33, 36, 41, 44, 47, 49. Seven observations have a fourth (middle) value 41, so median = 41, and there are 7 values in all. Check: Reading the key reconstructs the original two-digit numbers. Watch out: Do not interpret 3|1 as a decimal unless the key says so.

Cumulative frequency totals

Worked example 16. Question: Consecutive increasing classes have frequencies 5, 8, 4. Give cumulative totals after each class. Method: The running totals are 5, 5+8=13, then 13+4=17. The final value 17 is the total number of observations. Check: The cumulative values 5, 13, 17 never decrease. Watch out: Cumulative frequency is a running count, not a fresh isolated bar frequency.

Spot a truncated-axis bar-chart trap

Worked example 17. Question: Two categories have values 93 and 97. A bar chart starts the vertical axis at 90. What numerical comparison should be stated despite the visual exaggeration? Method: The true absolute difference is 97−93 = 4. Relative to the first value, this is 4/93×100 ≈ 4.30%. A baseline starting at 90 makes the visual bar lengths disproportionately different. Check: The values are both near a hundred; the numbers, not the cropped appearance, show a modest difference. Watch out: Do not infer percentage increase by comparing the lengths of bars on a truncated vertical scale.

Pie-chart angle from frequency

Worked example 18. Question: A survey has category frequencies 6, 8 and 10. What pie-chart angle represents the first category? Method: Total is 24. The first category’s fraction is 6/24 = 1/4. Its sector angle is ¼×360° = 90°. Check: All pie-chart sectors together must make 360°, and the first is exactly one-quarter. Watch out: Do not multiply a frequency directly by 360° without dividing by the total first.

Mean versus median with an outlier

Worked example 19. Question: A sorted five-value data set is 5, 6, 7, 8, 29. Find mean and median. Method: The middle (third) value is 7, so median = 7. Total is 55; dividing by five gives mean = 11. The very large last value raises the mean relative to the median. Check: Check 11 is between the smallest 5 and largest 29, while median is the central ordered value. Watch out: An extreme value can pull the mean; do not claim the median must move by the same amount.

Try it next: Replace the largest value by 9 and compare the two means. Try working backwards from the answer to see whether the original condition is recovered. If a real-world result is impossible, check the modelling assumptions. Identify one place where an incorrect sign, unit or label could reverse the interpretation of the answer.

Compare two groups without overclaiming

Worked example 20. Question: Data set A is 3, 3, 4, 5, 9 and B is 3, 4, 4, 5, 6. Compare their medians and ranges. Method: Both sets have median 4. A has range 9−3 = 6, while B has range 6−3 = 3. So typical middle value matches, but A is more spread by the range measure. Check: Each median is the third of five ordered values; each range is maximum minus minimum. Watch out: Same median does not imply identical spread or shape.

A frequency table and its mode

Worked example 21. Question: Three categories A, B and C have frequencies 4, 7 and 5. Find the total and modal category. Method: Total frequency = 4+7+5 = 16 observations. Category B has the largest frequency 7, so the mode is B. Check: The total counts observations, not a sum of category names or category scores. Watch out: A mode is the most frequent category, not the numerical average of counts.

Mean from an ungrouped frequency table

Worked example 22. Question: Scores 1, 2 and 3 have frequencies 4, 5 and 3. Find their mean. Method: Weighted total = 1×4 + 2×5 + 3×3 = 23. Total observations 12; mean = 23/12 ≈ 1.917. Check: The mean lies between the minimum value 1 and maximum value 3. Watch out: Do not divide the weighted sum by the three distinct scores; divide by total frequency.

Estimate a mean from grouped data

Worked example 23. Question: Equal-width score classes 0–<10, 10–<20 and 20–<30 have frequencies 4, 6, 3. Estimate the mean using class midpoints. Method: Class midpoints are 5, 15, 25. Estimated total = 5×4+15×6+25×3 = 185; divide by 13 observations. Estimated mean ≈ 14.23. Check: The estimate lies between 0 and 30 and respects the frequency weighting. Watch out: This is an estimate, since actual individual scores inside each class are not supplied.

Read an equal-width histogram

Worked example 24. Question: An equal-width histogram has adjacent classes 0–<5, 5–<10 and 10–<15. Its vertical axis is frequency, with bar heights 5, 8, 6. Find modal class and total frequency. Method: All class widths are five units, so the frequency-labelled heights give counts directly. The middle class 5–<10 has largest frequency 8; total frequency = 5+8+6 = 19. Check: The largest frequency is indeed 8 and the sum of all three frequencies is 19. Watch out: Check the actual vertical-axis label. An unequal-width histogram needs additional care not required for this equal-width example.

Read an ordered stem-and-leaf diagram

Worked example 25. Question: A stem-and-leaf display has key 4|1 = 41. Rows are 4|1 3 6 and 5|1 4 7 9. State the median and number of values. Method: It represents 41, 43, 46, 51, 54, 57, 59. Seven observations have a fourth (middle) value 51, so median = 51, and there are 7 values in all. Check: Reading the key reconstructs the original two-digit numbers. Watch out: Do not interpret 4|1 as a decimal unless the key says so.

Cumulative frequency totals

Worked example 26. Question: Consecutive increasing classes have frequencies 6, 9, 5. Give cumulative totals after each class. Method: The running totals are 6, 6+9=15, then 15+5=20. The final value 20 is the total number of observations. Check: The cumulative values 6, 15, 20 never decrease. Watch out: Cumulative frequency is a running count, not a fresh isolated bar frequency.

Spot a truncated-axis bar-chart trap

Worked example 27. Question: Two categories have values 94 and 98. A bar chart starts the vertical axis at 90. What numerical comparison should be stated despite the visual exaggeration? Method: The true absolute difference is 98−94 = 4. Relative to the first value, this is 4/94×100 ≈ 4.26%. A baseline starting at 90 makes the visual bar lengths disproportionately different. Check: The values are both near a hundred; the numbers, not the cropped appearance, show a modest difference. Watch out: Do not infer percentage increase by comparing the lengths of bars on a truncated vertical scale.

Pie-chart angle from frequency

Worked example 28. Question: A survey has category frequencies 7, 9 and 12. What pie-chart angle represents the first category? Method: Total is 28. The first category’s fraction is 7/28 = 1/4. Its sector angle is ¼×360° = 90°. Check: All pie-chart sectors together must make 360°, and the first is exactly one-quarter. Watch out: Do not multiply a frequency directly by 360° without dividing by the total first.

Mean versus median with an outlier

Worked example 29. Question: A sorted five-value data set is 6, 7, 8, 9, 30. Find mean and median. Method: The middle (third) value is 8, so median = 8. Total is 60; dividing by five gives mean = 12. The very large last value raises the mean relative to the median. Check: Check 12 is between the smallest 6 and largest 30, while median is the central ordered value. Watch out: An extreme value can pull the mean; do not claim the median must move by the same amount.

Try it next: Replace the largest value by 10 and compare the two means. Try working backwards from the answer to see whether the original condition is recovered. If a real-world result is impossible, check the modelling assumptions. Identify one place where an incorrect sign, unit or label could reverse the interpretation of the answer.

Compare two groups without overclaiming

Worked example 30. Question: Data set A is 4, 4, 5, 6, 10 and B is 4, 5, 5, 6, 7. Compare their medians and ranges. Method: Both sets have median 5. A has range 10−4 = 6, while B has range 7−4 = 3. So typical middle value matches, but A is more spread by the range measure. Check: Each median is the third of five ordered values; each range is maximum minus minimum. Watch out: Same median does not imply identical spread or shape.

A frequency table and its mode

Worked example 31. Question: Three categories A, B and C have frequencies 5, 8 and 6. Find the total and modal category. Method: Total frequency = 5+8+6 = 19 observations. Category B has the largest frequency 8, so the mode is B. Check: The total counts observations, not a sum of category names or category scores. Watch out: A mode is the most frequent category, not the numerical average of counts.

Mean from an ungrouped frequency table

Worked example 32. Question: Scores 1, 2 and 3 have frequencies 5, 6 and 4. Find their mean. Method: Weighted total = 1×5 + 2×6 + 3×4 = 29. Total observations 15; mean = 29/15 ≈ 1.933. Check: The mean lies between the minimum value 1 and maximum value 3. Watch out: Do not divide the weighted sum by the three distinct scores; divide by total frequency.

Estimate a mean from grouped data

Worked example 33. Question: Equal-width score classes 0–<10, 10–<20 and 20–<30 have frequencies 5, 7, 4. Estimate the mean using class midpoints. Method: Class midpoints are 5, 15, 25. Estimated total = 5×5+15×7+25×4 = 230; divide by 16 observations. Estimated mean ≈ 14.38. Check: The estimate lies between 0 and 30 and respects the frequency weighting. Watch out: This is an estimate, since actual individual scores inside each class are not supplied.

Read an equal-width histogram

Worked example 34. Question: An equal-width histogram has adjacent classes 0–<5, 5–<10 and 10–<15. Its vertical axis is frequency, with bar heights 6, 9, 7. Find modal class and total frequency. Method: All class widths are five units, so the frequency-labelled heights give counts directly. The middle class 5–<10 has largest frequency 9; total frequency = 6+9+7 = 22. Check: The largest frequency is indeed 9 and the sum of all three frequencies is 22. Watch out: Check the actual vertical-axis label. An unequal-width histogram needs additional care not required for this equal-width example.

Read an ordered stem-and-leaf diagram

Worked example 35. Question: A stem-and-leaf display has key 5|1 = 51. Rows are 5|1 3 6 and 6|1 4 7 9. State the median and number of values. Method: It represents 51, 53, 56, 61, 64, 67, 69. Seven observations have a fourth (middle) value 61, so median = 61, and there are 7 values in all. Check: Reading the key reconstructs the original two-digit numbers. Watch out: Do not interpret 5|1 as a decimal unless the key says so.

Cumulative frequency totals

Worked example 36. Question: Consecutive increasing classes have frequencies 7, 10, 6. Give cumulative totals after each class. Method: The running totals are 7, 7+10=17, then 17+6=23. The final value 23 is the total number of observations. Check: The cumulative values 7, 17, 23 never decrease. Watch out: Cumulative frequency is a running count, not a fresh isolated bar frequency.

Spot a truncated-axis bar-chart trap

Worked example 37. Question: Two categories have values 95 and 99. A bar chart starts the vertical axis at 90. What numerical comparison should be stated despite the visual exaggeration? Method: The true absolute difference is 99−95 = 4. Relative to the first value, this is 4/95×100 ≈ 4.21%. A baseline starting at 90 makes the visual bar lengths disproportionately different. Check: The values are both near a hundred; the numbers, not the cropped appearance, show a modest difference. Watch out: Do not infer percentage increase by comparing the lengths of bars on a truncated vertical scale.

Pie-chart angle from frequency

Worked example 38. Question: A survey has category frequencies 8, 10 and 14. What pie-chart angle represents the first category? Method: Total is 32. The first category’s fraction is 8/32 = 1/4. Its sector angle is ¼×360° = 90°. Check: All pie-chart sectors together must make 360°, and the first is exactly one-quarter. Watch out: Do not multiply a frequency directly by 360° without dividing by the total first.

Mean versus median with an outlier

Worked example 39. Question: A sorted five-value data set is 7, 8, 9, 10, 31. Find mean and median. Method: The middle (third) value is 9, so median = 9. Total is 65; dividing by five gives mean = 13. The very large last value raises the mean relative to the median. Check: Check 13 is between the smallest 7 and largest 31, while median is the central ordered value. Watch out: An extreme value can pull the mean; do not claim the median must move by the same amount.

Try it next: Replace the largest value by 11 and compare the two means. Try working backwards from the answer to see whether the original condition is recovered. If a real-world result is impossible, check the modelling assumptions. Identify one place where an incorrect sign, unit or label could reverse the interpretation of the answer.

Compare two groups without overclaiming

Worked example 40. Question: Data set A is 5, 5, 6, 7, 11 and B is 5, 6, 6, 7, 8. Compare their medians and ranges. Method: Both sets have median 6. A has range 11−5 = 6, while B has range 8−5 = 3. So typical middle value matches, but A is more spread by the range measure. Check: Each median is the third of five ordered values; each range is maximum minus minimum. Watch out: Same median does not imply identical spread or shape.

A frequency table and its mode

Worked example 41. Question: Three categories A, B and C have frequencies 6, 9 and 7. Find the total and modal category. Method: Total frequency = 6+9+7 = 22 observations. Category B has the largest frequency 9, so the mode is B. Check: The total counts observations, not a sum of category names or category scores. Watch out: A mode is the most frequent category, not the numerical average of counts.

Mean from an ungrouped frequency table

Worked example 42. Question: Scores 1, 2 and 3 have frequencies 6, 7 and 5. Find their mean. Method: Weighted total = 1×6 + 2×7 + 3×5 = 35. Total observations 18; mean = 35/18 ≈ 1.944. Check: The mean lies between the minimum value 1 and maximum value 3. Watch out: Do not divide the weighted sum by the three distinct scores; divide by total frequency.

Estimate a mean from grouped data

Worked example 43. Question: Equal-width score classes 0–<10, 10–<20 and 20–<30 have frequencies 6, 8, 5. Estimate the mean using class midpoints. Method: Class midpoints are 5, 15, 25. Estimated total = 5×6+15×8+25×5 = 275; divide by 19 observations. Estimated mean ≈ 14.47. Check: The estimate lies between 0 and 30 and respects the frequency weighting. Watch out: This is an estimate, since actual individual scores inside each class are not supplied.

Read an equal-width histogram

Worked example 44. Question: An equal-width histogram has adjacent classes 0–<5, 5–<10 and 10–<15. Its vertical axis is frequency, with bar heights 7, 10, 8. Find modal class and total frequency. Method: All class widths are five units, so the frequency-labelled heights give counts directly. The middle class 5–<10 has largest frequency 10; total frequency = 7+10+8 = 25. Check: The largest frequency is indeed 10 and the sum of all three frequencies is 25. Watch out: Check the actual vertical-axis label. An unequal-width histogram needs additional care not required for this equal-width example.

Read an ordered stem-and-leaf diagram

Worked example 45. Question: A stem-and-leaf display has key 6|1 = 61. Rows are 6|1 3 6 and 7|1 4 7 9. State the median and number of values. Method: It represents 61, 63, 66, 71, 74, 77, 79. Seven observations have a fourth (middle) value 71, so median = 71, and there are 7 values in all. Check: Reading the key reconstructs the original two-digit numbers. Watch out: Do not interpret 6|1 as a decimal unless the key says so.

Cumulative frequency totals

Worked example 46. Question: Consecutive increasing classes have frequencies 8, 11, 7. Give cumulative totals after each class. Method: The running totals are 8, 8+11=19, then 19+7=26. The final value 26 is the total number of observations. Check: The cumulative values 8, 19, 26 never decrease. Watch out: Cumulative frequency is a running count, not a fresh isolated bar frequency.

Spot a truncated-axis bar-chart trap

Worked example 47. Question: Two categories have values 96 and 100. A bar chart starts the vertical axis at 90. What numerical comparison should be stated despite the visual exaggeration? Method: The true absolute difference is 100−96 = 4. Relative to the first value, this is 4/96×100 ≈ 4.17%. A baseline starting at 90 makes the visual bar lengths disproportionately different. Check: The values are both near a hundred; the numbers, not the cropped appearance, show a modest difference. Watch out: Do not infer percentage increase by comparing the lengths of bars on a truncated vertical scale.

Pie-chart angle from frequency

Worked example 48. Question: A survey has category frequencies 9, 11 and 16. What pie-chart angle represents the first category? Method: Total is 36. The first category’s fraction is 9/36 = 1/4. Its sector angle is ¼×360° = 90°. Check: All pie-chart sectors together must make 360°, and the first is exactly one-quarter. Watch out: Do not multiply a frequency directly by 360° without dividing by the total first.

Mean versus median with an outlier

Worked example 49. Question: A sorted five-value data set is 8, 9, 10, 11, 32. Find mean and median. Method: The middle (third) value is 10, so median = 10. Total is 70; dividing by five gives mean = 14. The very large last value raises the mean relative to the median. Check: Check 14 is between the smallest 8 and largest 32, while median is the central ordered value. Watch out: An extreme value can pull the mean; do not claim the median must move by the same amount.

Try it next: Replace the largest value by 12 and compare the two means. Try working backwards from the answer to see whether the original condition is recovered. If a real-world result is impossible, check the modelling assumptions. Identify one place where an incorrect sign, unit or label could reverse the interpretation of the answer.

Compare two groups without overclaiming

Worked example 50. Question: Data set A is 6, 6, 7, 8, 12 and B is 6, 7, 7, 8, 9. Compare their medians and ranges. Method: Both sets have median 7. A has range 12−6 = 6, while B has range 9−6 = 3. So typical middle value matches, but A is more spread by the range measure. Check: Each median is the third of five ordered values; each range is maximum minus minimum. Watch out: Same median does not imply identical spread or shape.

Five errors that deserve a specific correction

1. Dividing a weighted total by the number of distinct rows instead of total frequency.

Do not simply cross out an incorrect answer and copy a clean one. Locate the first line where the interpretation or operation stopped being valid, explain the reason, and correct that decision. Then tackle a different-looking example to show that the repair survives changed numbers and wording. A short error log that records the cause is more useful than a long list of unexplained corrections.

2. Treating a grouped midpoint estimate as the exact raw-data mean.

3. Using an unlabeled stem-and-leaf display or forgetting its key.

4. Trusting cropped bar heights more than labelled numerical values.

5. Confusing a single category frequency with a cumulative frequency.

How a focused Bukit Timah small-group lesson could help

Small-group tuition works best when the learner has a real opportunity to attempt a question, describe an approach, receive precise feedback and try again without the same hint. The tutor can model one decisive step and then ask each pupil to choose a method independently. This is especially useful when school practice spans several subtopics and pupils must distinguish them for themselves.

Instead of measuring a lesson by the number of pages completed, look at whether the student can explain the first step and check the final answer in a new setting. For a three-student format, every student should still produce independent working and receive individual corrections. Parents should confirm the actual class size, timetable, tutor fit and chapter sequence rather than assume that a particular format guarantees a grade.

For further background on eduKateSG’s foundations approach, see the unchanged Secondary 1 Mathematics Tutor Clementi small-group tutorial guide. It is a different year-level resource and does not replace the Secondary 3 G3 syllabus. A local year-specific introduction is available at Secondary 3 G3 Mathematics Tuition Bukit Timah.

A practical six-week improvement plan

Week 1 — Diagnose the first sticking point

Choose two recent marked questions and one similar unseen question. Do not begin by counting only incorrect answers. Find where the student lost the meaning of a word, a formula, an axis, a unit or an algebraic operation. Write one specific target for the next lesson. For histograms, stem-and-leaf diagrams and misleading data, keep one dated original attempt and one later independently completed problem as evidence of real progress. The purpose is to improve correct mathematical decisions, not simply extend revision time.

Week 2 — Rebuild the relationship

Read a definition in plain language and solve a carefully chosen standard question. Ask the learner to explain why the relationship fits the given conditions, then try one new example without the model visible. Separate a conceptual misunderstanding from a slip in calculation. For histograms, stem-and-leaf diagrams and misleading data, keep one dated original attempt and one later independently completed problem as evidence of real progress. The purpose is to improve correct mathematical decisions, not simply extend revision time.

Week 3 — Contrast neighbouring cases

Place two look-alike questions side by side, one requiring the familiar method and the other needing a different first move. Ask the student to identify the deciding feature before doing arithmetic. A method learnt in isolation is not yet reliable in a mixed assessment. For histograms, stem-and-leaf diagrams and misleading data, keep one dated original attempt and one later independently completed problem as evidence of real progress. The purpose is to improve correct mathematical decisions, not simply extend revision time.

Week 4 — Make checking independent

Choose a reverse calculation, substitution, sketch, unit check or table comparison that can expose a wrong step. Explain what the check proves and what it does not. A second copy of the same mistaken keystroke is not an independent check. For histograms, stem-and-leaf diagrams and misleading data, keep one dated original attempt and one later independently completed problem as evidence of real progress. The purpose is to improve correct mathematical decisions, not simply extend revision time.

Week 5 — Mix chapters and retrieve later

Revisit the corrected idea after a gap of several days. Mix it with problems from other chapters, remove formula prompts and use a reasonable time limit only after the first step is secure. Note which misreadings or sign errors return. For histograms, stem-and-leaf diagrams and misleading data, keep one dated original attempt and one later independently completed problem as evidence of real progress. The purpose is to improve correct mathematical decisions, not simply extend revision time.

Week 6 — Compare unseen work with the starting point

Use a short new set with similar conceptual demand but different numbers and wording. Compare independent method choice, accuracy, clarity of written explanation and ability to spot an implausible answer. Choose the next one or two precise improvement targets. For histograms, stem-and-leaf diagrams and misleading data, keep one dated original attempt and one later independently completed problem as evidence of real progress. The purpose is to improve correct mathematical decisions, not simply extend revision time.

What parents can see without taking over the homework

Try three questions: ‘What is the problem asking?’, ‘Which fact or relationship makes the first operation legal?’, and ‘How will you know the result makes sense?’ These prompts encourage the student to explain a decision instead of asking a parent to supply the next step. Keep the conversation short and interested, especially if the child has already spent a long time on the task.

Progress may be visible before it appears in a major assessment mark. A learner starts a mixed question more confidently, recognises a wrong unit, spots a sign mistake or gives a clearer reason on the page. Compare independent work from different days and use school feedback to choose specific next targets. Aim for methods that remain useful in Secondary 4 rather than temporary fluency produced by repeated hints.

Frequently asked questions

Why do histogram bars touch?

They represent neighbouring intervals on a numerical axis, unlike unrelated categories in an ordinary bar chart. For practical revision, test the explanation on one fresh example with no prompt about which formula or diagram rule to use.

Does the K310 syllabus require unequal-class-width histograms?

Its stated histogram content specifies equal class intervals. Check the current official document and school lessons for the applicable assessment. For practical revision, test the explanation on one fresh example with no prompt about which formula or diagram rule to use.

What does 3|7 mean on a stem-and-leaf chart?

If the key states 3|7 = 37, it represents thirty-seven. A different key could change the interpretation. For practical revision, test the explanation on one fresh example with no prompt about which formula or diagram rule to use.

Why is a grouped mean an estimate?

The exact individual values inside each interval are unknown, so class midpoints stand in for them. For practical revision, test the explanation on one fresh example with no prompt about which formula or diagram rule to use.

How do you check cumulative frequency?

The running total should never decrease, and its last value should equal the total number of observations. For practical revision, test the explanation on one fresh example with no prompt about which formula or diagram rule to use.

Can a bar chart be technically labelled but still misleading?

Yes. A truncated vertical axis can exaggerate the visual difference between values. For practical revision, test the explanation on one fresh example with no prompt about which formula or diagram rule to use.

Should median or mean always be preferred?

Neither universally. The median is more resistant to extreme values, while the mean accounts for every numerical value. For practical revision, test the explanation on one fresh example with no prompt about which formula or diagram rule to use.

What is the quickest data-check habit?

Read the labels, sample size, units, interval widths and any axis break before comparing columns. For practical revision, test the explanation on one fresh example with no prompt about which formula or diagram rule to use.

For the authoritative content and assessment framework, refer to the 2027 G3 Mathematics K310 syllabus document and the SEAB G3 syllabus listing. A reading guide can clarify an idea, but it does not supersede the current school programme or examination instructions.

The takeaway

Did you know the most useful sign of progress is often quieter than a perfect worksheet? A student pauses, identifies the mathematical structure, selects a valid technique, shows coherent working and independently checks whether the result fits the question. That is the habit worth developing in histograms, stem-and-leaf diagrams and misleading data. Begin with one misconception, correct it carefully and return after a delay to see whether the improvement remains.