I have IQ scores and various levels of a different factor that is presumed to influence IQ. Subjects' IQ scores (on average) are different at different levels of this factor. I can show that IQ increases with increasing levels of this factor, but how do I compute the "p for trend" that I commonly read about? Is there more than one way to do this? And is it done automatically in any SPSS procedure? Thanks for any advice on this.
If your factor is a coarsened quantitative variable, and if the distance from one category to the next is (approximately) the same all the way along the scale, you could use the POLYNOMIAL contrast option available via the ONEWAY procedure in SPSS. Try Googling on to find more info and examples. (If the distances between categories are not constant, you can still use polynomial trend analysis, but you would be advised to modify the coefficients accordingly. The chapter on multiple comparison procedures in Dave Howell's Statistical Methods for Psychology explains how to do that, IIRC.)
HTH.
Hi Jerry, apply either the specified contrast as described by Bruce Weaver or Jonckheere-Terpstra test for trend. However, as the IQ scores are known to be normally distributed, the ANOVA with the contrast would work out well.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis