Secondary 3 data handling starts before the graph and before the mean: it starts with what data is being collected, how it is classified and whether the table actually represents the question fairly.
The G3 syllabus includes simple concepts in collecting, classifying and tabulating data before students interpret statistical diagrams and measures.
For the representation stage, read Secondary 3 Statistical Representations.
Start With a Clear Statistical Question
“What do students think?” is too vague.
“How many minutes do Secondary 3 students in this class spend travelling to school on a typical weekday?” is much clearer.
A clear question defines the variable, group and context.
Identify the Variable
A variable is the quantity or category being recorded.
Examples: travel time in minutes, number of siblings, transport mode, test score, height in centimetres.
Different variables need different classifications and representations.
Categorical Versus Numerical Data
Transport mode is categorical.
Travel time is numerical.
Numerical data can be discrete, such as number of siblings, or continuous, such as mass or time measured to suitable precision.
A Table Should Match the Variable
For categorical data, a simple frequency table can list each category and count.
For numerical values with many possible measurements, grouping into intervals may make patterns easier to see.
Worked Example: Frequency Table
Suppose 20 students report transport mode: 8 MRT, 5 bus, 4 walk, 3 car.
A frequency table records category and frequency.
Check total frequency: 8+5+4+3=20.
Tally Marks Are a Collection Tool
Tallies help count observations as they are collected.
Group tallies in fives so totals are easy to audit.
The final frequency column should be numerical and should sum to the number of observations.
Worked Example: Group Continuous Data
Journey times range from 6 to 42 minutes.
Intervals such as 0–10,10–20,20–30,30–40,40–50 may organise the data, provided the boundary convention is clear and each observation belongs to one class.
The intervals should cover all observed values without gaps or ambiguous overlap.
Class Boundaries Need Consistency
If one interval is 10≤t<20 and the next is 20≤t<30, a value t=20 belongs only to the second interval.
Students should know where boundary values are counted.
Too Many Classes Can Hide the Pattern
If every value receives its own interval, grouping adds little.
If intervals are too wide, meaningful variation can disappear.
A sensible classification balances detail and readability.
Too Few Observations Can Give an Unstable Picture
A sample of five students may not represent a cohort of several hundred very well.
A small data set can still answer a small descriptive question, but strong claims about a larger group require more care.
Collection Method Can Affect the Data
If a travel survey is conducted only among students who arrive early, the results may differ from surveying the whole class.
If an online poll is answered only by volunteers, the responding group may differ from the intended population.
These are data-quality concerns. The syllabus may keep this discussion simple, but the reasoning habit is valuable.
Question Wording Can Influence Answers
“How bad is the school commute?” encourages a different response frame from “How long is your school commute?”
When collecting opinion data, neutral wording helps the table reflect responses rather than the survey writer’s expectation.
Measurement Precision Matters
If students report travel time as 23.7 minutes but they only estimated roughly from memory, the decimal may suggest more precision than the data truly has.
Data should be recorded to a precision appropriate to the measurement process.
Missing Data Should Not Be Invented
If two students did not answer, record the missing responses appropriately rather than silently assigning them to a category.
The total frequency should reflect the actual valid observations used in later calculations.
Worked Example: Build a Grouped Table
Data: 7,12,14,18,21,22,24,29,31,33,37,42.
Using intervals 0–10,10–20,20–30,30–40,40–50 gives frequencies 1,3,5,2,1.
Check total=12 observations.
From Table to Representation
A category table can lead to a bar graph or pie chart.
Grouped continuous data can lead to a histogram.
A cumulative-frequency table can lead to a cumulative-frequency curve.
The table is the organised bridge between raw data and the graph.
From Table to Mean
For a frequency table of exact values, mean=Σfx/Σf.
For grouped intervals, class midpoints may be used to estimate the mean because the original individual values within each interval are no longer known.
A Data Audit Before Graphing
- What question is being answered?
- What is the variable?
- What are the units?
- How many valid observations are there?
- Do all frequencies sum correctly?
- Are class intervals clear and non-overlapping?
- Does the collection method fit the intended group?
- Is the recorded precision sensible?
Common Errors
- collecting data before defining the question;
- mixing categories with numerical intervals;
- creating overlapping class intervals;
- forgetting boundary values;
- frequencies not summing to the sample size;
- using overly precise measurements from rough estimates;
- drawing a graph before checking the table;
- making broad claims from a narrow or unrepresentative group.
A Five-Question Independent Check
- Is “transport mode” categorical or numerical?
- Is “height in centimetres” categorical or numerical?
- Why must class intervals avoid overlap?
- A table has frequencies 6,8,5,4. How many observations?
- Why should the total frequency be checked before graphing?
Answers
Question 1: categorical. Question 2: numerical and usually continuous. Question 3: each observation should belong to one clear class. Question 4: 23. Question 5: an incorrect table will produce an incorrect graph and statistics.
Exam-Day Data-Handling Routine
- identify variable and units;
- read class boundaries carefully;
- check total frequency;
- distinguish raw, frequency and grouped data;
- use the table to choose the correct representation or statistic;
- state conclusions only as strongly as the data supports.
How data collection, classification and tabulation Fits a 3-Pax Secondary 3 Mathematics Lesson
At eduKatePunggol, Secondary Mathematics is taught in focused groups of up to three students in 1.5-hour lessons near Punggol MRT. The small class matters because a final answer does not reveal where the reasoning broke.
One student may need a prerequisite repaired. Another may understand the concept but lose marks through setup, units or working. A third may be ready for mixed application, timing or extension. The topic can be shared while the next question differs.
Warm-up retrieval
Begin with one short older question so the current topic remains connected to the wider Mathematics system.
Concept before procedure
Teach the relationship before increasing volume. A remembered formula is useful only when the student knows which quantity it describes and when it applies.
Guided to independent practice
Use prompts while the idea is new, then remove them. A fresh question with no worked example visible is the real independence check.
Mixed practice
Once topical work is stable, remove the chapter heading. The student should recognise the structure rather than wait for the tutor to name the method.
Error review
Record the first wrong move in the Secondary 3 Mathematics error log. “Careless” is too broad. A specific error can be trained.
Clear working
Enough working should remain visible to trace the method. See Should Students Show Working or Do It Mentally?.
Three Secondary 3 Student Pathways
Repair
Find the earliest unstable prerequisite affecting the current topic, repair it narrowly and reconnect it to school work.
Stabilisation
Use delayed retrieval, mixed questions and school-test review to make performance more consistent.
Extension
Reduce unnecessary routine repetition and add explanation, transfer, alternative methods or selected timing.
What Parents Can Bring
- recent school tests and weighted assessments;
- marked homework or worksheets;
- the school’s current topic sequence;
- teacher comments;
- one question the student cannot start;
- one question that is correct but unusually slow;
- the student’s own description of what feels difficult.
The useful question is not only “What mark did my child get?” but “What pattern produced the mark?”
What Progress Should Look Like
- less hesitation on familiar structures;
- clearer working and diagrams;
- fewer repeated mistakes;
- better retrieval after a gap;
- stronger recognition in mixed questions;
- better explanation of why a method applies;
- calmer performance under time.
Responsible tuition does not promise an instant grade. Improvement depends on the size of the gap, consistency of practice, school demands and time before assessment.
Helpful Reading
- Secondary 3 to SEC Mathematics — Build the Exam Runway Before Secondary 4
- Secondary 3 Statistical Representations
- Secondary 3 Topical Practice to Mixed Practice
- Secondary 3 Mathematics Error Log
- How Much Secondary 3 Mathematics Practice Each Week?
- What to Do Between Weekly Secondary 3 Mathematics Tuition Lessons
- When Should Secondary 3 Students Start Full SEC Mathematics Papers?
- G1, G2 and G3 Mathematics Parent Guide
- Punggol Mathematics Article Index
Official 2027 SEC G3 Mathematics Reference
For current G3 Mathematics scope, see the SEAB K310 G3 Mathematics Syllabus for 2027. Schools may sequence topics differently, so match practice to the student’s current school programme and assessment scope.
Families who want to discuss a Secondary 3 Mathematics plan can WhatsApp eduKatePunggol. Please check current class availability and fees directly.
Properly taught kids shine a bright light into the future.
Why the Topic Must Survive a Delay
Same-day success is not enough. A student can follow a worked example while the method is fresh and still lose it several weeks later.
Use a simple sequence: immediate independent question, delayed retrieval, mixed question and later appearance inside a timed or school-paper setting.
Cold-start check
Give a fresh question without notes, examples or a topic heading. Can the student identify the first valid step?
Transfer check
Change the wording, numbers, diagram or representation while preserving the underlying relationship.
Timed check
Add time only after the method is accurate. The clock should reveal fluency, not replace understanding.
Secondary 4 Handoff
Before Secondary 4, the topic should be more than a recently completed chapter. The student should retrieve it after a gap, recognise it inside mixed work and use a checking routine without waiting for a prompt.
That is the standard that turns Secondary 3 knowledge into an SEC runway.
A Worked Mini-Survey From Question to Table
Question: “How many minutes do students in this class spend travelling from home to school on a usual weekday?”
Variable: travel time in minutes.
Suppose responses are 12,18,22,25,27,31,35,36,42,48.
A simple grouped table might use 10≤t<20,20≤t<30,30≤t<40,40≤t<50 with frequencies 2,3,3,2.
Check total frequency=10 before any graph is drawn.
Class Width Should Serve the Question
Intervals of width 1 minute would preserve detail but create many mostly empty classes.
Intervals of width 30 minutes could hide useful variation.
The grouping decision should balance simplicity with the pattern the student needs to see.
Worked Example: Discrete Data
Number of siblings for 15 students: 0,1,1,2,0,3,1,2,2,1,4,0,2,1,3.
A frequency table can use exact values 0,1,2,3,4 because the variable is discrete and has few possible outcomes.
Frequencies are 3,5,4,2,1.
Grouping into intervals such as 0–2 and 3–5 would lose unnecessary detail here.
Worked Example: Categorical Data
Transport modes: MRT, bus, walk, car.
Do not order the categories numerically unless there is a meaningful reason.
A frequency table and bar graph are natural choices because the values are labels rather than measurements.
Data Coding Must Be Defined
If responses are coded 1=MRT, 2=bus, 3=walk, 4=car, those numbers are labels.
The mean of the codes has no meaningful transport interpretation.
Students should distinguish numerical codes from quantitative measurements.
Frequency Density Is Not Needed for Every Histogram
The current G3 syllabus specifies histograms with equal class intervals.
With equal class widths, comparing bar heights is straightforward for frequency in the expected school treatment.
Students should not import a more advanced frequency-density method unless the question or course requires it.
Cumulative Frequency Table
If class frequencies are 2,3,3,2, cumulative frequencies are 2,5,8,10.
The cumulative total should end at the full number of observations.
A wrong final cumulative frequency immediately reveals an earlier counting error.
Two-Way Classification
Some data can be classified by two categorical variables, such as transport mode and whether a student arrives before 7:30 am.
A two-way table organises counts across both classifications.
Row and column totals provide useful checks on the data.
Worked Example: Two-Way Table Reasoning
Suppose 20 students are classified by transport: 12 public transport, 8 other; and by arrival: 14 early, 6 later.
If 9 public-transport students arrive early, then 3 public-transport students arrive later. Early students using other transport=14−9=5, leaving 3 later students using other transport.
Every row and column total should agree with the grand total 20.
Collection Bias: Keep the Claim Proportional to the Method
If only students at one CCA are surveyed about school travel, the sample may not represent all Secondary 3 students.
The data may still accurately describe that CCA group.
The correct conclusion should match the group actually observed.
Response Bias and Question Wording
A question such as “Don’t you agree the bus is too crowded?” encourages a direction of response.
A neutral question such as “How crowded do you usually find the bus?” with defined response choices is less leading.
This reasoning helps students see that data quality begins before calculation.
Measurement Error
A stopwatch time can be affected by reaction time. A ruler reading can depend on scale precision. A self-reported travel time can depend on memory.
Not all data uncertainty is caused by arithmetic.
Missing Data
If a student does not answer, mark the response missing rather than silently treating it as zero.
Zero can be a genuine data value; missing is absence of data.
Outliers During Collection
An unusual value should be checked for possible recording error but not automatically deleted.
A 120-minute commute may be unusual but genuine.
If it was meant to be 12.0 and entered incorrectly, the correction should be based on evidence rather than convenience.
Data Cleaning Before Analysis
- check impossible values;
- check duplicated entries if duplicates are not expected;
- check units;
- check missing values;
- check category spelling or coding;
- check class boundaries;
- check total frequency.
Simple data cleaning prevents a graph or mean from faithfully summarising a mistake.
A Practice Ladder
- classify variable type;
- build a simple frequency table;
- group continuous data;
- build cumulative totals;
- read two-way tables;
- evaluate a collection method;
- choose an appropriate representation;
- state a conclusion no stronger than the data supports.
Frequently Asked Questions
Should all numerical data be grouped?
No. Small discrete data sets can often be shown exactly. Grouping is useful when many values make patterns hard to see.
What is the difference between zero and missing?
Zero is a recorded numerical value. Missing means no valid value was recorded.
Why check total frequency?
It confirms that the table accounts for the intended number of observations before later calculations are built on it.
Can a survey be accurate but not representative?
Yes. It may accurately describe the surveyed group while being a poor basis for claims about a larger population.
What if a value looks strange?
Investigate it. Correct only when there is evidence of an error.
Does a table need units?
When the variable has units, yes. “25” means little without knowing whether it is minutes, centimetres or marks.
Parent Check: Ask Where the Data Came From
Before discussing a graph, ask, “Who was measured, what was measured, and how?”
A student who can answer those questions is much less likely to treat statistics as numbers detached from evidence.
Population and Sample
A population is the whole group the question is interested in. A sample is the subset actually observed.
If the goal is to understand all Secondary 3 students in a school but only 40 students answer a survey, those 40 form the sample.
The quality of the inference depends partly on how the sample was chosen.
A Sample Can Be Convenient but Unrepresentative
Surveying only students in one CCA, one class or one arrival-time window may overrepresent certain experiences.
A convenience sample can still describe the people who answered, but broad conclusions about the whole population need caution.
Random Sampling as a Fairness Idea
A simple random sample aims to give each population member an equal chance of selection.
At school level, the exact technical method may be simple, but the reasoning is important: avoid choosing only the easiest people to ask.
Stratified Thinking
If a cohort contains different levels or groups, a sample can be designed to preserve those proportions where appropriate.
For example, if 60% of a population belongs to Group A and 40% to Group B, a proportionate sample might preserve roughly the same split.
This concept helps students see that representativeness can depend on population structure.
Worked Example: Sampling Bias
A school wants to estimate average travel time. A survey is conducted only among students who arrive before 7:00am.
The sample may overrepresent students with particular transport arrangements or longer safety margins.
The resulting average may not represent all students.
Question Design Can Create Response Bias
Compare “How convenient is our excellent new transport arrangement?” with “How convenient do you find the transport arrangement?”
The first wording nudges the respondent toward a positive frame.
Neutral wording improves the chance that responses reflect the respondent rather than the survey designer.
Categories Should Be Mutually Clear
If transport categories are “MRT”, “bus”, “public transport” and “walk”, MRT and bus both fit inside public transport, creating overlap.
A frequency table needs categories that classify each response consistently.
If multiple responses are allowed, state that explicitly instead of pretending the categories are exclusive.
Worked Example: Design Better Categories
Question: “How do you usually travel to school?”
Possible exclusive categories might be: walk only; private car/taxi; bus as main mode; MRT/LRT as main mode; bicycle/PMD where applicable; other.
The precise categories depend on the survey purpose, but they should be interpretable without double counting.
Class Width Affects the Story
Journey times grouped in 5-minute intervals reveal more detail than 30-minute intervals.
But very narrow classes can make the table noisy and cumbersome.
The grouping should be fine enough to show meaningful structure and broad enough to remain readable.
Open Versus Closed Intervals
Intervals such as 10≤t<20 and 20≤t<30 give every boundary value one clear home.
A table labelled 10–20,20–30 without a stated convention can be ambiguous at exactly 20.
Good mathematical communication removes that ambiguity.
Tallying Without Losing Data
When observations arrive one by one, tally marks provide a live counting system.
After collection, convert tallies into frequencies and check that the frequency sum matches the number of valid observations.
A mismatch means the table should be repaired before any graph or mean is calculated.
Missing and Invalid Responses
Suppose 50 survey forms are returned but three omit the travel-time question.
The valid sample size for travel-time analysis is 47, not automatically 50.
Do not insert zero minutes or another guess for missing values.
Outliers Need Investigation
A reported travel time of 240 minutes in a mostly 10–60 minute data set could be real, a data-entry mistake or an unusual circumstance.
Check the source if possible before deleting or changing it.
Cleaning data should be transparent rather than convenient.
Measurement Versus Category Errors
A travel time entered as 2 hours while the table expects minutes is a unit inconsistency.
A response counted in two mutually exclusive categories is a classification inconsistency.
These are different errors and require different corrections.
From Raw Data to Frequency Table
Raw values: 12,15,18,18,21,24,25,29,33,36.
Grouped into 10≤t<20: four values; 20≤t<30: four values; 30≤t<40: two values.
The grouped table is shorter but loses the exact internal values.
That loss of detail is the trade-off for a clearer overview.
From Frequency Table to Graph
Check the table before choosing the graph.
Category frequencies may lead to a bar graph. Grouped continuous intervals may lead to a histogram. Running totals lead to a cumulative-frequency curve.
The table structure should match the graph structure.
A Data-Quality Checklist
- Is the statistical question clear?
- Is the population clear?
- How was the sample selected?
- Are categories or class intervals unambiguous?
- Are units consistent?
- Are there missing or impossible values?
- Do frequencies sum to the valid sample size?
- Does the table support the graph or statistic being requested?
Strong Student Extension: Redesign a Weak Survey
Give the student a biased survey question and an overlapping category list.
Ask them to rewrite the question neutrally, define a better sampling method and create non-overlapping categories.
This extends statistics beyond calculation into the quality of evidence.
Frequently Asked Questions
Does a large sample guarantee a fair result?
No. A large biased sample can still misrepresent the target population.
Should every survey use random sampling?
Not necessarily, but the selection method should fit the question and its limitations should be understood.
Why not group data immediately?
Raw data may contain useful detail or errors that become harder to see after grouping.
Can categories overlap?
They can if multiple responses are intentionally allowed, but the table must then be designed and interpreted accordingly.
What should happen before drawing a graph?
Check the table totals, intervals, units and validity of the data first.
Secondary 4 Handoff for Data Collection
The student should enter Secondary 4 able to distinguish population from sample, classify data clearly, build reliable tables and recognise when the data-collection method limits the conclusion.
That reasoning protects every later statistic built on top of the data.
Worked Example: Raw Data to Frequency Table
Raw scores: 12,15,18,18,21,22,22,22,27,31.
For exact values, a frequency table can list value and frequency:
12→1, 15→1, 18→2, 21→1, 22→3, 27→1, 31→1.
Total frequency=10.
This table preserves exact scores while removing repeated writing.
When to Group the Same Data
If instead 200 scores span 0 to 100, an exact-value table may be long and noisy.
Grouping into intervals such as 0–10,10–20,… can make the distribution easier to see.
The cost is loss of exact individual values.
Open and Closed Class Boundaries
Intervals should be defined so every value belongs to exactly one class.
Using 10≤x<20 and 20≤x<30 is clear.
Writing “10–20” and “20–30” without convention can be ambiguous for x=20 unless the context establishes the intended boundary rule.
Worked Example: Cumulative Frequency Table
Intervals have frequencies 3,5,7,4,1.
Cumulative frequencies are 3,8,15,19,20.
The final cumulative total matches the total sample size 20.
If it does not, recheck the running addition before plotting a cumulative-frequency graph.
Two-Way Tables as Consistency Checks
Suppose a table classifies students by transport and arrival time.
Row totals, column totals and grand total should all agree.
An unknown cell can often be found by subtracting known entries from a row or column total.
Worked Example: Unknown Cell
A table has 30 students. 18 use public transport. 20 arrive before 7:30. Of the public-transport students, 12 arrive early.
Public transport later=18−12=6.
Other transport early=20−12=8.
Other transport total=12, so other transport later=4.
Later total=6+4=10, consistent with 30−20.
Design Categories That Do Not Overlap
If classifying reading time, categories “0–30 min” and “30–60 min” need a clear convention at 30 minutes.
If classifying transport, categories “bus” and “public transport” overlap because bus is a type of public transport.
Categories should be mutually exclusive when each observation is meant to enter one category only.
Exhaustive Categories
If every observation must be classified, the categories should also cover all relevant cases.
A transport survey with only “MRT”, “bus” and “walk” has no place for car, bicycle or mixed-mode journeys unless an “other” or more complete classification is provided.
Data Coding and Re-Coding
A category code can make electronic tabulation easier, but the code itself does not become a numerical measurement.
If MRT=1 and bus=2, the statement “bus is twice MRT” has no meaning.
Check for Impossible Values
A travel time of −15 minutes is impossible in the ordinary context.
A test score of 105 may be impossible if the maximum is 100.
Data cleaning should flag such values before graphing.
Check for Unit Mixtures
If some heights are recorded in metres and others in centimetres, the table must convert them to one unit before statistics are calculated.
A frequency table can look internally neat while combining incompatible measurements.
Simple Inference Must Stay Within the Evidence
If 18 of 25 surveyed students prefer option A, we can say option A was preferred by most of that surveyed group.
Claiming “all Secondary 3 students prefer A” would exceed the evidence unless the sampling design justifies such a generalisation.
From Data Collection to Representation
A good workflow is:
- define question;
- identify variable and units;
- collect observations;
- clean obvious data issues;
- classify or group;
- tabulate frequencies;
- choose representation;
- calculate statistics;
- interpret in context.
Skipping the early steps can make the later statistics precise answers to a poorly defined question.
What to Put in the Error Log
- class intervals overlap;
- total frequency wrong;
- mixed units;
- missing value treated as zero;
- category codes averaged as if quantitative;
- graph drawn before table checked;
- claim stronger than data supports.
Secondary 4 Handoff for Data Handling
Before Secondary 4, the student should be able to move from raw observations to a clean table and explain why the classification is sensible.
That foundation makes later graphs, averages, quartiles and standard deviation much more reliable because the data entering those calculations has already been organised correctly.
Worked Example: From Question to Table
Question: “How long do students in this class spend travelling to school on a usual weekday?”
Variable: travel time in minutes.
Population for this small study: the class.
Suppose 30 valid responses are collected. Before grouping, inspect the raw values for impossible entries, missing responses and unit inconsistency.
Only after that check should intervals be chosen.
Design Class Intervals Around the Data Range
If all values lie between 8 and 52 minutes, intervals 0–10,10–20,…,50–60 cover the range cleanly.
Using intervals 0–100 and 100–200 would technically cover the values but destroy useful detail.
Using one-minute intervals could create an unnecessarily long table.
Good classification balances visibility and simplicity.
Worked Example: Boundary Discipline
Suppose classes are 0≤t<10,10≤t<20 and 20≤t<30.
A value 10 belongs in the second class, not the first.
A value 20 belongs in the third.
Writing the inequalities once removes ambiguity from every later tally.
Frequency Density Is Not Needed for Every Histogram
When class widths are equal, ordinary frequency can directly determine bar height in the simple histogram treatment used by the current G3 syllabus.
Students should not import unequal-class frequency-density machinery unless their school question actually requires it.
Staying within the scope helps keep ordinary Mathematics distinct from unnecessary extension.
Data Cleaning Should Preserve Evidence
If one response says 240 minutes while all others lie below 60, mark it for checking rather than deleting it instantly.
If it was entered as 240 seconds by mistake, correct the unit only when the source confirms that interpretation.
Good data handling records what happened rather than silently changing inconvenient values.
Worked Example: Missing Response
Fifty students return a survey, but three leave the travel-time item blank.
For travel-time statistics, valid n=47 unless the missing values are handled through a stated method.
Do not divide the total time by 50 merely because 50 forms were returned.
From Classification to Inference
A table can describe the sample accurately and still fail to justify a claim about a larger population.
For example, surveying one high-performing class cannot automatically describe all Secondary 3 students in Singapore.
The strength of the conclusion should match the sampling design.
A Parent-Friendly Data Audit
- What exactly was measured?
- Who was included?
- Who might have been missed?
- Are the units consistent?
- Do the categories overlap?
- Do frequencies add correctly?
- Are any values impossible or suspicious?
- What can the data actually support?
These questions teach statistical thinking before formulas appear.
A One-Week Data-Handling Sequence
- Day 1: identify variables and data types.
- Day 2: build frequency tables from raw data.
- Day 4: design sensible grouped intervals.
- Day 5: audit a biased sample or poorly worded survey.
- Day 7: choose a representation from the cleaned table and explain why.
Frequently Asked Questions
Is more data always better?
More data can improve stability, but poor sampling or bad measurement can still produce misleading conclusions.
What is the difference between population and sample?
The population is the full group of interest; the sample is the subset actually observed.
Why should I inspect raw data before grouping?
Grouping can hide impossible values, duplicates, unit errors or unusual observations.
Can I choose any class intervals I like?
Intervals should cover all valid data, avoid overlap and preserve enough detail for the question.
What should happen before graphing?
Verify the table, sample size, units and class definitions first.
Secondary 4 Handoff for Data Handling
The student should enter Secondary 4 able to move from a clear statistical question to a defensible table, recognise sampling limitations and choose an appropriate representation without being told every step.
That foundation makes later interpretation, standard deviation and graph comparison more trustworthy.
