# Chapter 13, Table 12: Using SPSS Statistics

### Testing Polynomial Contrasts via SPSS point and click

For the hypothetical data contained in Table 13.2, the linear and quadratic D variables were formed by making use of the appropriate coefficients from Appendix Table A.10. Because the eight participants were measured at three occasions, both a linear and a quadratic effect can be tested. The question of interest in this instance is: “Is there a linear and/or quadratic trend exhibited by the group over time?” Recall that in the book (pages 646-647) it was shown that the D variables for linear and quadratic effects led to an omnibus F test of 19.148, which was a value previously obtained for the omnibus effect. Because the particular values chosen for the D variables do not matter (unless it leads to a linear combination of columns), we will focus only on the tests of the individual contrasts when analyzing the data given in Table 12.

Because columns one and two already represent the linear and quadratic effect respectively, all that needs to be done is to test mean of the column in order to determine if it differs from zero.

#### 4. After moving the quadratic particular trend to the Test Variable(s) box, clicking OK will produce the results.

Note: a more direct way to carryout tests of polynomial trends is to specify polynomial contrasts from within the Repeated Measures menu. When defining the levels of the repeated factor (as we did early in this chapter), the Contrastsbutton can be clicked (bottom left-hand side of the dialog box) and then Polynomial specified. This requires directly inputting the raw data (rather than the linear composite from making use of the coefficients in Appendix Table A. 10).