Higher: estimating medians from cumulative-frequency graphs
Higher: draw calibrated read-off guides to estimate a median from a supplied cumulative-frequency graph, including nonunique plateau readings.
Exam scope and assessment notes
The first four lessons are shared Foundation work inherited by Higher; the last four are additional Higher spread, quartile and cumulative-frequency reading work. Keep the actual repeated observations and frequencies, state units, and distinguish an average from spread. Grouped midpoint means and within-class graph readings are estimates. The chosen discrete quartile rule is median of each half, excluding the overall middle observation when the count is odd; it is not claimed as a universal Pearson algorithm. Draw the horizontal and vertical guides yourself, use each graph’s own total, and retain a full plateau interval when an inverse is not unique. Fixed numerical agreement does not mark reasoning or award marks or mastery. Specialist and formal-assessment approval remain pending. This is the Issue 2 linear 4MA1 scope: November 2026/June 2027 entries and regional-centre eligibility differ from modular awards. Literal second-edition Book2 Handling Data4/5/6 and Hub tags guide navigation only; no Book1 6.2 tag or verified chapter/Hub body is inferred. Handling Data6 map208–222 conflicts with proof chapter207/summary221/next222; Handling Data4 map39–60 extends beyond summary59 before next61. Graphical representation6.1 and probability6.3 remain separate.
Before you begin
- Read a calibrated cumulative-frequency graph.
- Interpret a median and half of a total.
- Draw horizontal and vertical Cartesian guides.
Work on paper. Try each question before revealing help, and explain why your method works.
Quick readiness check
A complete cumulative-frequency graph ends at cumulative count 36. At what height should the median read-off start?
What you will learn
- Read the completed graph’s total N and locate its median height N/2.
- Construct horizontal and vertical read-off guides and report a graph-based estimate with units.
- Distinguish a unique model intersection from a plateau interval and explain why grouped graph readings are not exact individual-data medians.
Use the supplied straight-guide cumulative-frequency graph for 40 journey times, with boundary points (0,0), (10,6), (20,18), (30,32), (40,40). Draw read-off guides to estimate the median time to 1 decimal place. Explain why this is not an exact individual-data median. On the supplied calibrated diagram, show each requested horizontal and vertical read-off guide; keep the measurement unit and any complete overlap interval.
Enter your own actual observations, positions, products, operations and guide endpoints. No count, total, role, result, quartile, graph height, unit or precision is filled for you. Recording is an attempt, not a correctness result.
Original supplied observations and tables
Only literal observations, frequencies, intervals and supplied cumulative guides are shown initially. The display does not sort observations, compute a total, average, quartile, rank or recovered value, or prefill a learner operation.
All metric panels share one stable camera. Measurement and cumulative-count axes have separately labelled scales; an observation index is not a planar data coordinate. Pan, zoom and whole-view turns change only the frame.
Supplied cumulative guide G1
given: Literal supplied values only.
Supplied cumulative guide G1
given: Supplied values only.
Blue denotes literal supplied guides, amber actual entered work, purple dirty retained work, and green a separately gated reference. Wrong and partial values are never repaired. The drawing bound is±1,000,000 coordinates, with at most512 scalar entries and128 cumulative points; these are tool bounds, not mathematical laws. Native clipping does not clamp values. No completed target is silently substituted.
The selected median-of-halves convention excludes the overall middle observation for odd lists. Grouped means and guide read-offs remain estimates. A zero-frequency plateau can give an interval rather than a unique reading. Fixed guide comparisons use the separate canonical700×500 policy with each source total; the tolerance is not a current-screen pixel or official marking tolerance. Text presence, operation consistency and numerical self-check do not approve reasoning, statistical claims, marks, mastery, teacher/specialist or formal assessment. No probability or new boxplot teaching credit is granted.
G1 median: draw horizontal guide, every vertical read-off and any full overlap
No actual work is recorded for this action.
State explicitly whether your reading has an interval overlap.
Only a selected CF endpoint uses the board. The reading bounds, ranks, list positions and operation rows are separate scalar work. Add each vertical leg yourself; a second leg or overlap is never invented.
My quantities, conclusions and explanation
For a numerical list, use comma-separated values with no brackets or one balanced enclosing pair, or type none when you mean no mode. State the actual unit and exact, estimate or one-decimal-place precision yourself.
Retain the stated data, units, signs and precision. Numerical agreement is a self-check only.
Retain the stated data, units, signs and precision. Numerical agreement is a self-check only.
Retain the stated data, units, signs and precision. Numerical agreement is a self-check only.
Retain the stated data, units, signs and precision. Numerical agreement is a self-check only.
A written attempt needs actual reasoning; text presence does not approve the argument.
A written attempt needs actual reasoning; text presence does not approve the argument.
These are original fixed cases for self-checking, not automatic marking, mastery or teacher/specialist/formal-assessment approval. The stated median-of-halves convention is a chosen convention. Grouped midpoint means and within-class CF readings remain estimates; a plateau can give a full interval. A physical instrument and calibrated graph are separate from the screen. No probability credit follows.
Find the correct height on the graph
A cumulative-frequency graph records the number of observations up to successive value boundaries. Its final cumulative value supplies the total N when the diagram is stated to be complete. The median is estimated at cumulative height N/2. That height belongs on the cumulative-frequency axis, not halfway along the horizontal axis.
Check both axis labels and their tick spacing before drawing. The horizontal axis carries the measurement unit, while the vertical axis carries cumulative counts. A lower boundary need not have horizontal value zero. An unfinished graph with an unknown final total cannot supply a justified median height merely because its last visible count looks like a convenient total.
Make the read-off construction visible
From N/2 on the cumulative axis, draw a horizontal guide until it reaches the supplied curve. From that intersection draw a vertical guide to the value axis. Mark the reading and its unit. The supplied examples use piecewise straight guides, so their model intersections can also be checked by proportional movement along the relevant segment.
The exact intersection of a chosen straight segment is a check on that model, while a reading from a drawn graph is approximate. Neither recovers the actual middle observations hidden inside grouped intervals. Use the stated plotting scale and requested rounding; extra calculator digits do not make an uncalibrated or poorly drawn graph more precise.
Recognise an overlap instead of inventing one answer
If the median-height guide meets a rising part of the curve at one place, there is a unique horizontal read-off for that supplied model. If it lies along a horizontal plateau, the intersections occupy a whole interval. Mark both ends and retain all the possible graph-level readings rather than silently selecting the midpoint or first endpoint.
A plateau reflects no increase in cumulative frequency across that value interval. It does not reveal the exact values on either side of the gap. Selecting a single discrete-data median would need the relevant observations or another explicit rule. The task is a reasoned graphical estimate, not an exact automatic mark or a substitute for learning how to construct the original cumulative graph.
Sources for this lesson
- Pearson Mathematics A linear specification · Issue 2 · Foundation printed24/physical30; Higher printed39/physical45; inheritance printed27/physical33
- Pearson 2026/2027 entry manual · centre regions physical3; Mathematics A physical25
- Pearson second-edition modular book/Hub map · literal navigation only · physical9,10,11
Pause and explain
A complete CF guide has total 20 and a horizontal plateau at height 10 from x = 6 to x = 9. What does its median-height intersection give?
Worked example
Use the supplied straight-guide cumulative-frequency graph for 40 journey times, with boundary points (0,0), (10,6), (20,18), (30,32), (40,40). Draw read-off guides to estimate the median time to 1 decimal place. Explain why this is not an exact individual-data median. On the supplied calibrated diagram, show each requested horizontal and vertical read-off guide; keep the measurement unit and any complete overlap interval.
Show the worked solution
- Check the supplied axis labels and tick intervals. Read the completed cumulative total for each panel. Draw a horizontal guide at each requested fraction of its total, then a vertical guide from the actual intersection to the measurement axis; if a height overlaps a plateau, show both ends of the overlap.
- The final cumulative value is N = 40; the median guide starts at N/2 = 20 on the cumulative-frequency axis.
- It meets the guide between (20,18) and (30,32). Linear-model reading is 20 + (20 − 18)/(32 − 18) × 10 = 150/7 minutes.
- A carefully drawn graph gives about 21.4 minutes. The original within-class values are unknown, so even the exact straight-guide interpolation is a grouped-data estimate.
Answer Use N = 40 and median height 20. A drawn estimate is about 21.4 minutes. The straight-guide model intersection is 150/7 minutes; it does not recover the exact median of the hidden individual observations.
Try a topic test
Choose Easy, Normal or Difficult. Save your answers and open worked review after finishing.
Test this topic →From guided practice to a new situation
Write your answer and reasoning first. Use a hint only if you are stuck. The model solution is for self-checking; your written work is not automatically marked.
Use the supplied CF straight guide through (5,0), (15,8), (25,24), (35,36), (45,48), for mass in grams. Draw median read-off guides and estimate the median to 1 decimal place. Explain why the value-axis start need not be zero. On the supplied calibrated diagram, show each requested horizontal and vertical read-off guide; keep the measurement unit and any complete overlap interval.
Give me a hint
Read the final cumulative total, then take half and follow the supplied guide to the value axis.
Compare my reasoning
- Check the supplied axis labels and tick intervals. Read the completed cumulative total for each panel. Draw a horizontal guide at each requested fraction of its total, then a vertical guide from the actual intersection to the measurement axis; if a height overlaps a plateau, show both ends of the overlap.
- Read the completed graph’s final count 48, then halve that total to obtain the median height 24 on the cumulative-frequency axis.
- The guide passes through (25,24), giving a drawn estimate of about 25.0 g.
- The cumulative count is zero at mass 5 g, the supplied lower boundary; the axis need not begin at mass zero. The median still estimates the original data rather than revealing their central observations.
Use N = 48 and median height 24. A drawn estimate is about 25.0 grams. The straight-guide model intersection is 25 grams; it does not recover the exact median of the hidden individual observations.
Look for these in your work
- Read N and use N/2.
- Construct actual read-off guides on calibrated axes.
- Explain estimate status or nonuniqueness.
On the supplied CF guide with points (0,0), (4,9), (8,27), (12,45), (16,60), estimate the median length to 1 decimal place using drawn guides, and show the rank you used. On the supplied calibrated diagram, show each requested horizontal and vertical read-off guide; keep the measurement unit and any complete overlap interval.
Give me a hint
Use half the cumulative count, not half the horizontal range, and keep the graph reading approximate.
Compare my reasoning
- Check the supplied axis labels and tick intervals. Read the completed cumulative total for each panel. Draw a horizontal guide at each requested fraction of its total, then a vertical guide from the actual intersection to the measurement axis; if a height overlaps a plateau, show both ends of the overlap.
- The total is 60, so the median is read at cumulative frequency 30.
- Between (8,27) and (12,45), the straight-guide model gives 8 + 3/18 × 4 = 26/3 cm.
- The graph reading is about 8.7 cm. Use 30 on the cumulative axis, not the midpoint 8 of the horizontal scale.
Use N = 60 and median height 30. A drawn estimate is about 8.7 cm. The straight-guide model intersection is 26/3 cm; it does not recover the exact median of the hidden individual observations.
Look for these in your work
- Read N and use N/2.
- Construct actual read-off guides on calibrated axes.
- Explain estimate status or nonuniqueness.
The supplied CF straight guide passes through (0,0), (10,15), (20,15), (30,30). Draw the median-height guide. Does the graph determine a unique horizontal read-off? State the full intersection set and explain what extra assumption would be needed to select one number. On the supplied calibrated diagram, show each requested horizontal and vertical read-off guide; keep the measurement unit and any complete overlap interval.
Give me a hint
Draw the entire horizontal overlap at the required height before deciding whether a unique answer is available.
Compare my reasoning
- Check the supplied axis labels and tick intervals. Read the completed cumulative total for each panel. Draw a horizontal guide at each requested fraction of its total, then a vertical guide from the actual intersection to the measurement axis; if a height overlaps a plateau, show both ends of the overlap.
- The total is 30 and the required height is 15. The graph is horizontal at this height from x = 10 to x = 20.
- Every point of the closed interval [10,20] lies on that guide, so the graph-level inverse is not a unique value. Show the entire overlap.
- Selecting 15 merely because it is the interval midpoint adds a rule that was not given. Recovering the exact discrete median would require the relevant individual observations or an explicitly supplied convention/model.
The median-height guide is at cumulative frequency 15 and overlaps the full interval [10,20] minutes. The supplied guide gives no unique graph-level median read-off; neither an endpoint nor the midpoint is automatically justified.
Look for these in your work
- Read N and use N/2.
- Construct actual read-off guides on calibrated axes.
- Explain estimate status or nonuniqueness.
Common mistakes
- Using half the horizontal axis maximum instead of N/2 on the cumulative axis.
- Calling a straight-guide interpolation the exact median of the hidden individual observations.
- Choosing a plateau midpoint without a stated rule or inventing a final total from a cropped graph.
What can you explain now?
Why can an exact straight-segment calculation still give only an estimated median for the original observations?
Compare with the explanation
The straight segment is a chosen guide through grouped cumulative information; the individual observations inside the intervals remain unknown.
Exact arithmetic describes that guide’s intersection, while the original data median and a finite-resolution drawn reading have different information limits.
After trying it yourself, choose your next review. This is your self-assessment.
Your review choice appears on Today. Sign in to sync it across devices.
Make it stick
Describe the entire median read-off, including how you obtain N, calibrate the axes, draw both guides and handle a plateau.
If you needed the solution, close it and solve the problem again from a blank page. A correct answer is stronger when you can explain the reason for each step.