Read beyond the noise.

The interquartile range shows how widely the middle half is spread

A reader comparing dot plots with quartile marks on paper.
AI-generated photographic illustration; not documentary photography.

Two services can report the same median waiting time while offering very different experiences. In one, most waits cluster near the middle. In the other, the middle group is spread across a much wider interval. A median alone cannot show that difference.

The interquartile range, or IQR, subtracts the lower quartile from the upper quartile. These mark the twenty-fifth and seventy-fifth percentiles. NIST describes the IQR as a measure of spread that is less affected by extreme values than the standard deviation.

A worked example with invented figures

Imagine Service A has a lower quartile of 18 minutes and an upper quartile of 26 minutes. Its IQR is 8 minutes. Service B’s quartiles are 10 and 34 minutes, giving an IQR of 24 minutes. Both could have a median of 22 minutes.

These are illustrative numbers, not performance data. They show why “the typical wait is 22 minutes” can conceal a difference in consistency. B’s central half occupies a wider span, even though the midpoint agrees with A’s.

Keep the tails in the conversation

The IQR does not describe every wait. Someone outside the middle half can have a substantially shorter or longer experience. A small IQR therefore does not prove that no one faced a serious delay.

When reading a report, ask how many observations were included, what period they cover and whether the services measured waiting time from the same starting point. A comparison between registration-to-appointment and referral-to-appointment is not repaired by adding more statistics.

Also retain the reported quartiles, rather than copying only their difference. An IQR of eight minutes could describe several very different locations on the time scale. The median, quartile endpoints and information about unusual waits tell a more useful story together than any one figure presented as the complete verdict.