• Practice
  • About 16 min
  • 5 practice questions

Histograms with varying class widths

See why a histogram with unequal class widths misleads if heights equal frequencies, and adjust the heights in proportion to the smallest class size so that areas match the frequencies.

If something looks wrong, please tell us.

By the end of this lesson you can

  • Explain why unequal widths need adjusted heights.
  • Calculate adjusted lengths: (frequency ÷ class width) × minimum class width.
  • Draw and read a histogram whose classes have different widths.

Sometimes a few values are spread thinly over a wide range, so it makes sense to use wider classes there. For example, a teacher may group marks as 0-20, 20-30, ..., 50-60 and then 60-100 because few students scored very low or very high. If we draw such a histogram with heights equal to the frequencies, the picture is wrong.

The problem

Here are the marks of 63 students (made-up data for practice).

Marks of 63 students (made-up data for practice)
Marks0-2020-3030-4040-5050-6060-100Total
Number of students681215101263

The classes 30-40 and 60-100 both have 12 students. But the 60-100 rectangle is 4 times as wide, so with height 12 its area is 4 times as large, and it looks as if far more students scored above 60. Remember: in a histogram, area must be proportional to frequency.

The fix: adjust the heights

Adjusting heights for varying widths

  1. Find the smallest class size Here it is 10 (the classes 20-30 up to 50-60).
  2. For each class compute adjusted length = (frequency ÷ class width) × smallest width
  3. Draw rectangles Width = actual class width; height = adjusted length.
  4. Label the vertical axis 'Number of students per 10 marks interval'.
Adjusted lengths (made-up data for practice)
MarksFrequencyWidthAdjusted length
0-206206/20 × 10 = 3
20-308108/10 × 10 = 8
30-40121012
40-50151015
50-60101010
60-100124012/40 × 10 = 3

Now the 60-100 rectangle has height 3 and width 40, and the 30-40 rectangle has height 12 and width 10. Both have area 120 units, equal areas for equal frequencies. The picture is honest.

Adjusted length = (frequency ÷ class width) × minimum class width

— Varying widths

Worked example 2: ages at a health camp

Ages of 75 people at a camp (made-up data for practice): 0-5: 10, 5-10: 15, 10-20: 30, 20-40: 20. Smallest width = 5. Adjusted lengths: 0-5: 10; 5-10: 15; 10-20: 30/10 × 5 = 15; 20-40: 20/20 × 5 = 5. The vertical axis reads 'people per 5-year interval'. Notice that although 10-20 has the largest frequency (30), its people are spread over 10 years, so its rectangle is only as tall as that of 5-10.

Reading such a histogram

  • Height tells you how crowded a class is (frequency per standard width), not the frequency itself.
  • To get a class's frequency back, multiply its height by (its width ÷ the standard width). For 60-100: 3 × 40/10 = 12.
  • Equal heights in classes of different width mean different frequencies.

Key takeaways: when class widths differ, heights equal to frequencies make wide classes look too big. Choose the smallest class width, scale every height by (frequency ÷ width) × smallest width, and label the axis 'per … interval' so the areas match the frequencies.

Check your understanding

  1. 1. In a histogram the smallest class width is 10. What is the adjusted length for the class 20-50 with frequency 18?

    1. 18
    2. 54
    3. 6
    4. 1.8
    Show answer

    Answer: C. Width of 20-50 is 30. Adjusted length = 18/30 × 10 = 6.

  2. 2. Classes are 0-25 (frequency 15), 25-30, 30-35 and so on, with smallest width 5. What is the adjusted length for 0-25?

    1. 15
    2. 3
    3. 75
    4. 5
    Show answer

    Answer: B. Width of 0-25 is 25. Adjusted length = 15/25 × 5 = 3.

  3. 3. Why are heights adjusted when the class widths are unequal?

    1. So that the areas of the rectangles are proportional to the frequencies
    2. So that all rectangles have the same height
    3. So that the histogram has gaps
    4. So that the total frequency changes
    Show answer

    Answer: A. In a histogram area shows frequency. With unequal widths, heights equal to frequencies make wide classes look too large, so heights are scaled to keep area proportional to frequency.

  4. 4. In an adjusted histogram (standard width 10), the class 40-70 has height 7. What is its frequency?

    1. 7
    2. 70
    3. 10
    4. 21
    Show answer

    Answer: D. Frequency = height × (width ÷ standard width) = 7 × 30/10 = 21.

  5. 5. Classes 0-10 and 10-40 both have frequency 12 (smallest width 10). How do their adjusted rectangles compare?

    1. Same height, same area
    2. 0-10 is 3 times as tall; both have the same area
    3. 10-40 is taller
    4. 0-10 is 3 times as tall and has 3 times the area
    Show answer

    Answer: B. Adjusted heights: 0-10 gives 12; 10-40 gives 12/30 × 10 = 4. So 0-10 is 3 times as tall, but areas are 12 × 10 = 120 and 4 × 30 = 120, equal, because the frequencies are equal.

Sign in to keep track of the lessons you finish.

Next lesson → Frequency polygons from a histogram

Privacy · Terms · Refunds · Delivery