A mathematical expression containing one or more terms, each of which consists of a coefficient and a variable(s) raised to some power. Thus a + bx is a linear polynomial and a + bx + cx2 is a quadratic polynomial in x. A polynomial of order m includes terms with powers of x up to xm. Polynomials will typically provide excellent fits to the calibration data but it is difficult to find a case where polynomials have contributed to forecasting in the social or management sciences. On the contrary, they are associated with poor accuracy. Do not use polynomials unless there is a very strong a priori case.