A percentile is a position in an ordered dataset. For example, an observation at the 75th percentile is at or above roughly 75% of observations according to the method used.
Percentiles are not percentages
A score of 75% and a 75th-percentile rank describe different things. The first is a proportion of a possible total; the second is a relative position among observations.
Methods can disagree
Different statistical software uses different conventions for interpolation, treatment of ties and sample endpoints. A source should state which convention it uses, especially when samples are small.
Population matters
A household income percentile in one country cannot be compared mechanically with an individual earnings percentile in another. The denominator, time period, currency, demographic unit and tax treatment all matter.
A five-value example
For sorted values 10, 20, 30, 40 and 50, the median is 30. Exact percentile results away from the median can differ depending on the chosen statistical method. This is why credible statistical tools should publish their formula and data sources.
When Worlds Daily releases percentile calculators, each tool will display its source vintage, geographic scope, calculation assumptions and limitations.