Quartiles from cumulative frequency
With grouped data, the original individual values are hidden. A cumulative frequency curve lets you estimate the value below which a chosen number of observations lie.
Enter on the count axis first
Let be the total frequency. The lower quartile, median and upper quartile are , and . Their cumulative counts are:
The original individual values are hidden inside the grouped data, so each read from the curve is an estimate.
Enter on the count axis first
- Find the required cumulative count. Use , or on the cumulative-frequency axis.
- Move horizontally to the curve. The curve converts a cumulative count into an estimated value.
- Move vertically to the value axis. Read and report the value in context. Graph readings are estimates.
Worked example
Read the middle half of 60 values
Enter the count axis at 15, 30 and 45. The illustrated curve gives about 18, 30 and 40.
Common mistake
A percentile uses the same route
For percentile , enter the count axis at:
The 90th percentile is the value below which about 90% of the data lie. First calculate the count, then read from the curve exactly as you did for a quartile.
Worked example
Find the 90th-percentile count
There are 240 observations.
Plot upper boundaries against cumulative totals
- Add frequencies down the table to obtain cumulative frequency.
- Plot each cumulative total at the upper class boundary.
- Include the lower boundary of the first class with cumulative frequency zero.
- Draw one smooth increasing curve rather than a dot-to-dot polygon.
A cumulative frequency graph represents 200 journey times. It gives minutes and minutes. State the two counts used to obtain these readings, find the IQR and interpret it.
Show worked answer
The counts are 50 and 150.
The middle 50% of journey times span about 13 minutes.
