Secondary 4 statistics becomes easier when students separate the data from the display. A cumulative-frequency curve, box plot, histogram or summary statistic is not the data itself. It is a way of seeing structure inside the data.

Students often lose marks because they read the wrong scale, confuse frequency with cumulative frequency, place quartiles in the wrong order or calculate correctly but interpret the graph incorrectly.

At eduKatePunggol, Secondary 4 Mathematics is taught in focused groups of up to three students in 1.5-hour lessons near Punggol MRT. This topic guide supports the wider Secondary 4 SEC Mathematics year plan. The examples below are original teaching examples.

Frequency and cumulative frequency are different questions

Frequency tells us how many observations are in a category or interval. Cumulative frequency tells us how many observations are at or below a given boundary.

Suppose the frequencies in four intervals are 5, 8, 12 and 5. The cumulative frequencies are:

5, 13, 25, 30.

The final cumulative frequency is the total number of observations, 30.

A student who copies the original frequencies directly into a cumulative-frequency table has not made a calculation error. The meaning of the column is missing.

Worked example 1: build cumulative frequency from a table

An invented grouped table has frequencies 4, 7, 9 and 10.

Running totals give:

4, 11, 20, 30.

The second cumulative frequency is 4 + 7 = 11. The third is 11 + 9 = 20. The fourth is 20 + 10 = 30.

The values should never decrease. If a cumulative-frequency column goes from 20 down to 18, the student should know immediately that something is wrong.

Quartiles divide ordered data into four parts

For a cumulative-frequency graph, the lower quartile, median and upper quartile correspond to approximately 25%, 50% and 75% of the total frequency.

If there are 80 observations, the corresponding cumulative-frequency levels are:

• Q1 at cumulative frequency 20;
• median at cumulative frequency 40;
• Q3 at cumulative frequency 60.

From each cumulative-frequency level, move horizontally to the curve and then vertically to the data axis. Because the graph is drawn, the result may be an estimate rather than an exact value.

Worked example 2: interquartile range

Suppose a cumulative-frequency graph gives Q1 ≈ 18 and Q3 ≈ 42.

The interquartile range is:

IQR = Q3 − Q1 = 42 − 18 = 24.

The IQR describes the spread of the middle 50% of the data. A larger IQR means that the central half is more spread out.

A box plot compresses five important values

A standard box plot uses:

• minimum;
• lower quartile;
• median;
• upper quartile;
• maximum.

The box runs from Q1 to Q3. The line inside marks the median. The whiskers extend toward the minimum and maximum according to the convention being used in the question.

Students should read the scale before interpreting the plot. A box that looks physically wider on the page does not automatically have a larger numerical IQR if the axes differ.

Worked example 3: compare two box plots

Suppose Group A has median 52 and IQR 12. Group B has median 47 and IQR 20.

Group A has the higher median, so its central value is higher. Group A also has the smaller IQR, so the middle half of its data is less spread out.

A good comparison states both the measure and the interpretation. “Group A is better” may be too vague unless the question provides a context and asks for that judgment.

Do not confuse range and interquartile range

The range is maximum − minimum. It uses the two extremes.

The interquartile range is Q3 − Q1. It uses the middle half of the data.

Because they describe different parts of the distribution, they can tell different stories. A dataset may have a large overall range because of an extreme value while the middle 50% remains tightly grouped.

Worked example 4: estimate a percentile from cumulative frequency

Suppose a dataset contains 200 observations and we want the 90th percentile.

Calculate 90% of 200:

0.90 × 200 = 180.

On the cumulative-frequency graph, locate 180 on the cumulative-frequency axis, move to the curve, then read the corresponding data value.

The exact numerical answer depends on the graph. The important method is converting the percentile into the correct cumulative-frequency level first.

Statistics questions often test reading before calculation

A student can lose marks even with perfect arithmetic if the axis scale is read incorrectly or the question asks for a comparison rather than a calculation.

• Check axis labels and units.
• Notice whether the graph uses class boundaries.
• Decide whether the answer is exact or estimated.
• Identify whether the question asks for centre, spread or proportion.
• Write a comparison using the requested statistic.

The question-reading guide is useful when correct calculations are being applied to the wrong request.

---

How we diagnose statistics mistakes

Table error: cumulative totals are built incorrectly.

Graph-reading error: the scale, boundary or axis is misread.

Quartile error: the student uses 25, 50 and 75 instead of 25%, 50% and 75% of the actual sample size.

Comparison error: the student gives values but does not interpret centre or spread.

Calculator error: arithmetic is correct in setup but entered incorrectly.

These categories point to different repairs and help prevent “statistics is weak” from becoming an unhelpfully broad label.

Why the three-student format helps

In a small group, the tutor can ask one student to build cumulative totals, another to read quartiles from a graph and another to compare two distributions. The shared concept is data interpretation, while each learner practises the part that is currently unstable.

The tutor can also see whether a wrong answer comes from arithmetic, graph reading or interpretation before assigning more questions.

What a 90-minute lesson could look like

An illustrative lesson could begin with ten minutes of table reading, twenty minutes on cumulative frequency and quartiles, twenty minutes of guided graph work, twenty minutes of independent mixed statistics questions and twenty minutes for comparison writing, error review and continuation work.

Repair, stabilisation and extension

Repair: use small tables and obvious running totals. Practise reading scales before drawing curves or box plots.

Stabilisation: mix cumulative frequency, quartiles, IQR and box-plot comparisons so the required statistic must be identified.

Extension: ask students to compare distributions critically, explain what a measure does not tell us and connect percentile positions to cumulative-frequency readings.

Try a short independent set

• Frequencies are 6, 9, 5 and 10. Write the cumulative frequencies.
• For 120 observations, find the cumulative-frequency levels for Q1, median and Q3.
• A box plot has Q1 = 24 and Q3 = 41. Find the IQR.

Answers: 6, 15, 20, 30; 30, 60, 90; and 17.

Repeat later with different sample sizes so quartile positions are calculated rather than memorised. If the student succeeds only when the headings tell them exactly what to do, continue mixed practice using the topic-mastery framework.

What progress should look like

• cumulative frequencies never decrease;
• quartile positions are based on the total frequency;
• graph scales are read accurately;
• range and IQR are not confused;
• comparisons state both centre and spread where relevant;
• estimated graph values are presented with appropriate care.

Punggol class details and consultation inputs

eduKatePunggol Secondary Mathematics lessons are 1.5 hours in groups of up to three students near Punggol MRT. Confirm current availability, fees and meeting arrangements directly.

Bring the student’s subject level, examination year and recent statistics papers or school graphs. If the student calculated a quartile correctly but interpreted it incorrectly, keep that original response intact so the diagnosis remains visible.

Frequently asked questions

Why must cumulative frequency increase?

Because each new value is a running total that includes all earlier observations plus the next group.

Is the median always exactly halfway between Q1 and Q3?

No. The median is the 50th-percentile position, but its numerical distance from Q1 and Q3 depends on the distribution.

Why use IQR instead of range?

IQR describes the spread of the middle half of the data and is less directly influenced by extreme endpoints than the range.

Read the data story, not just the graph

Return to the Secondary 4 Mathematics year plan for the wider revision sequence. For paper review when statistics errors recur, use the full-paper error-map guide.

Build the running total, locate the quartiles, read the scale and explain what the numbers mean. Families can WhatsApp eduKatePunggol with recent work to discuss a suitable next step.