Learning curve

For other uses, see Learning curve (disambiguation).

A learning curve is a graphical representation of how an increase in learning (measured on the vertical axis) comes from greater experience (the horizontal axis); or how the more someone (or thing) does something, the better they get at it.^[1]

Fig 1: Learning curve for a single subject, showing how learning improves with experience

Fig 2: A learning curve averaged over many trials is smooth, and can be expressed as a mathematical function

The term learning curve is used in two main ways: where the same task is repeated in a series of trials, or where a body of knowledge is learned over time. The first person to describe the learning curve was Hermann Ebbinghaus in 1885, in the field of the psychology of learning, although the name wasn't used until 1909.^[2][3] In 1936, Theodore Paul Wright described the effect of learning on production costs in the aircraft industry.^[4]This form, in which unit cost is plotted against total production, is sometimes called an experience curve.

The familiar expression "a steep learning curve" is intended to mean that the activity is difficult to learn, although a learning curve with a steep start actually represents rapid progress.^[5][6]

In psychologyEdit

The first person to describe the learning curve was Hermann Ebbinghaus in 1885. His tests involved memorizing series of nonsense syllables, and recording the success over a number of trials. The translation does not use the term learning curve—but he presents diagrams of learning against trial number. He also notes that the score can decrease, or even oscillate.^[6][7][8]

The first known use of the term learning curveis from 1909: "Bryan and Harter (6) found in their study of the acquisition of the telegraphic language a learning curve which had the rapid rise at the beginning followed by a period of retardation, and was thus convex to the vertical axis."^[3][6]

Psychologist Arthur Bills gave a more detailed description of learning curves in 1934. He also discussed the properties of different types of learning curves, such as negative acceleration, positive acceleration, plateaus, and ogive curves. (Fig 1)^[9]