I need to show that BMI of adults is neither age nor gender specific. Should I do ANCOVA or randomly make age groups and do two way ANOVA?
Asked
Active
Viewed 373 times
0
-
1Categorizing continuous data (eg, age) is always a bad idea. Multiple regression & ANCOVA are ultimately the same thing, but sometimes software present them differently; it might be best to think of this as a basic multiple regression. A bigger issue, though, is that it sounds like you want to affirm the null, which is logically impossible. It may help you to read my answer here: Why do statisticians say a non-significant result means "you can't reject the null", rather than accepting the null hypothesis? – gung - Reinstate Monica Aug 09 '14 at 00:33
-
Thank you. I get my mistake. In this case, I got a significant effect in gender and interaction between age and gender. It would be silly to ask of me. Isin't my conclusion violating the known "fact"? – user53740 Aug 09 '14 at 01:20