percentielen

Percentielen, or percentiles in English, are statistical measures that describe the position of a value within a data distribution. A percentile indicates the value below which a given percentage of observations fall. The p-th percentile is the value below which p percent of the data lie; for example, the 25th percentile is the value below which 25% of observations fall, the 50th percentile being the median, and the 90th percentile marking the value below which 90% fall.

Several conventions exist to compute percentiles from a finite data set. A common approach is the nearest-rank method: sort data in ascending order and choose the smallest value for which at least p% of observations are at or below it. More refined methods use linear interpolation between adjacent order statistics, yielding the p-th percentile as a weighted average of neighboring data points. Different software packages may adopt different conventions, so results may vary slightly, especially for small samples.

Percentiles are used to summarize distributions and to compare individuals to a reference population. They appear in growth charts, standardized test scores, and income distributions. They are related to but distinct from percentile ranks; the rank indicates the percentage, while the percentile is the corresponding value.

Percentiles relate to other split points such as quartiles (25th, 50th, 75th) and deciles (every 10 percent). When interpreting percentiles, it is important to consider sample representativeness and the method used to compute them, since ties and small samples can affect values.

Example: In a height distribution, the 90th percentile of 1.75 m means that 90% of people have a height of 1.75 m or less.