Tinkering with Mean and Median

Aim

Investigate, using dot plots and counters, how the mean and the median of a collection change when values are included or removed, and how they behave when every value is increased, decreased or multiplied by a fixed number.

Apparatus

  • A printed or hand-drawn number line strip from 0 to 35 (one per group)
  • Counters, bindis, buttons or sticker dots to place on the strip
  • Paper slips, one data value written on each
  • Coloured pens, a ruler and a notebook
  • A calculator (optional, for checking only)

Procedure

  1. Put one counter above each of 5, 7, 8, 8, 12 on the strip. Find the mean and the median and record them.
  2. Include 20 with a counter of a new colour. Recalculate the mean and the median.
  3. Remove 20 and include 2 instead. Record the new mean and median.
  4. Include a value equal to the mean, 8. Note what happens to the mean and, separately, to the median.
  5. Go back to the original data. Remove 12 and record; restore it and remove 5 instead.
  6. Find two values, then three values, that can be included without changing the mean. Give two solutions for each.
  7. Slide every counter 10 steps right (add 10), then subtract 3. Predict the new mean first, then check.
  8. Double every value and place the counters afresh. Predict, then check. Also multiply by 5. Record in Table 2.
  9. Use algebra to say what happens to the mean when k is added to every value, and when every value is multiplied by k.

Interactive 3D Experiment

What to observe

  • Including a value greater than the mean raises the mean, while the median moves little or not at all; a smaller value lowers it.
  • Including a value equal to the mean leaves the mean unchanged. Two values can be included without changing the mean if their total is twice the mean.
  • Adding 10 to every value adds 10 to the mean and the median; multiplying every value by a number multiplies both by it.
  • The mean responds to every value and is pulled by extreme values; the median depends only on the middle of the ordered data.

Notes

  • Math Spark: a single extreme value can drag the mean a long way but hardly moves the median. This is why reports on income usually quote the median: one very rich resident can lift the mean income of a whole town, while the middle person’s income stays where it was. ‘Median’ comes from the Latin medianus, ‘in the middle’.
  • Arrange the values in ascending order before finding the median, every time the data changes.
  • When the number of values is even, the median is the average of the two middle values, not either one of them.
  • Recalculate the mean with the new number of values; dividing by the old count is a common error.
  • Place each counter exactly above its value on the strip, stacking counters for repeated values.
  • Use a different colour of counter for each newly included value so that the change can be traced.
  • Predict the new mean before calculating it, and only then check.
  • Return to the original data set before starting each new investigation.
  • Use the calculator only for checking, after the arithmetic has been done by hand.