Dot Plots, Median & Outliers

Aim

To organise data on a dot plot, find the median, identify outliers, and see how the mean and the median change when an outlier is present.

Apparatus

  • Graph paper or pre-drawn number-line strips
  • Ruler, pencil and coloured sketch pens
  • Sticky notes or small counters to mark data points

Procedure

  1. Collect a data set: story books read by 9 members (Fig. 2.1); later, family heights and another set.
  2. Tap the number line at each value to place a dot. Stack dots when a value repeats.
  3. Tap the dot plot to sort the values. Tap the middle value — the median (even count, the two middle ones).
  4. Work out the mean. The mean (solid line) and the median (dashed line) are marked on the dot plot.
  5. Look for a value far from the rest — an outlier. Is it at the low or the high end?
  6. Drag the outlier off the line. Find the median and the mean again. What changes?
  7. Compare the sets: when is mean < median, mean > median, mean ≈ median?

Interactive 3D Experiment

What to observe

  • Story books: median 5, mean 6.22. The outlier 20 pulls the mean above the median.
  • Without 20 the mean falls to 4.5, but the median only moves from 5 to 4.5.
  • Family heights have no outlier: mean 156.17 cm ≈ median 156.5 cm.
  • An outlier at the low end (1) pulls the mean below the median: 11 < 13.
  • The median is the middle value of ordered data. It is more reliable than the mean when there is an outlier.

Notes

  • Math Spark: in a real Test innings many batters scored 0, so the median was 0 though the team scored 407. News reports use the median for incomes and house prices: a few huge values pull the mean up, not the median.
  • Mark the number line with equal spacing before plotting any dots.
  • Place one dot for each value, stacking repeats neatly in a straight column.
  • Arrange the values in increasing order before picking out the median.
  • When the number of values is even, take the average of the two middle values.
  • Do not discard an outlier without a reason — record it and say how it affects the mean and the median.
  • Recount the dots against the original data list before drawing conclusions.