ABSTRACT
Testing contrasts of levels of a quantitative factor is often referred to as trend analysis. Another term that is frequently used to describe this form of analysis is the method of orthogonal polynomials. The test for nonlinearity is often referred to as a test for deviations from linearity. This phrase holds the key for understanding how to test for nonlinearity. Just as the test of the linear trend can be conceptualized as a test of a contrast, tests of the other beta parameters can also be formulated in terms of contrasts. It may be helpful to further examine our numerical example to gain a better understanding of nonlinear trends, especially how both linear and nonlinear trends might exist in the same data. To understand the meaning of the quadratic trend, it is helpful to remove the effects of the linear trend from the data.
