bell curve

The smooth, idealised version of a hump-shaped histogram, drawn so that a distribution can be summarised by two numbers: where the middle is and how wide the spread is. It is a model laid over data, not the data, which makes it clean to reason with and dangerous when the real set is lopsided, has two humps, or has a tail that never thins out.

The smooth, idealised version of a hump-shaped histogram, drawn so that a distribution can be summarised by two numbers: where the middle is and how wide the spread is. It is a model laid over data, not the data, which makes it clean to reason with and dangerous when the real set is lopsided, has two humps, or has a tail that never thins out.

Example

Best when the set really is single-humped and you want to talk about middle and spread as two numbers.