Interquartile Range

Turkish equivalent: Çeyrekler AçıklığıDomain: Data Systems

The difference between the upper and lower quartiles, measuring the spread of the middle half of a distribution.

The interquartile range, abbreviated IQR, measures the width of the central half of an ordered dataset. It is the difference between the upper and lower quartiles:

IQR = Q3 − Q1

If Q1=18 and Q3=30, the IQR equals 12 units. It says nothing directly about the number of observations, the mean or the symmetry of the entire distribution; it is the distance between two quantiles.

Resistance is not invariance

Standard deviation uses distances of all observations from the mean. IQR is based on central order statistics and is usually much less affected by a single extreme value. Yet the estimate is not entirely independent of computational choices: alternative sample-quantile conventions may produce different quartiles on small datasets.

From quartiles to displays and rules

Freedman–Diaconis uses 2×IQR/n^(1/3) to estimate histogram width. A box plot commonly spans Q1 to Q3, while the Tukey convention flags values beyond Q1−1.5×IQR and Q3+1.5×IQR as candidates for outlier investigation. Crossing such a fence does not prove a sensor fault, invalid entry or other data defect.

When Q1=Q3, the IQR is zero. This may occur in heavily tied data even when the minimum and maximum differ. Interpreting the spread then requires examining the full empirical distribution.