Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Using UK data, evidence is provided that the widely used quadratic earnings function may fare badly as indicated by recent work in the USA. Using data from pooled time series of cross-sections from the General Household Survey the suitability of the quadratic specification alongside higher order polynomials is investigated on samples of men and, for the first time in work of this type, women working full-time. It is found that the quadratic specification understates earnings growth at low levels of experience, especially for those workers with few formal qualifications. The usage of higher order polynomials provides a better fit.