[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: Re: Nonparametric two way ANOVA
At 11:28 05/02/04 -0500, Joseph Wagner wrote:
In my experience, viral loads are lognormally distributed. I would
therefore log-transform the viral loads and use linear regression methods
on the logs to estimate geometric means and their ratios, using the
-eform()- option of -regress-. For instance:
I want to do what I know is a pretty simple calculation but I am constrained
by my dependent variable which is non-normal.
I wish to see if there are differences in viral load stratified by age
category and by race. Viral load is not normal and in my case, I cannot
transform this variable to approximate normal.
I know I could do kruskal wallis or trend tests by race and by age category
individually but I need to do both.
If viral load were normal, I could do:
anova vload agecat race
The Friedman test is the only thing that comes to mind but I don't think
this is what I want.
gene byte baseline=1
xi:regress vload i.agecat i.race baseline, noconst robust eform(GM/Ratio)
The parameter -baseline- will then be a baseline geometric mean viral load,
representing geometric mean viral load for individuals with the reference
age category and race category. The other parameters will be ratios of
geometric mean viral load in other race and age categories, compared to the
baseline category. The effect of race and age are assumed to be multiplicative.
We need to generate the variable -baseline- and use the -noconst- option of
-regress-, because, if we use the -sform- option without the -noconst-
option, then Stata does not print the intercept, which in this case is the
baseline geometric mean. Using the variable -baseline- with the -noconst-
option, we can fool Stata into thinking there is no intercept for it to hide.
I hope this helps.
Lecturer in Medical Statistics
Department of Public Health Sciences
King's College London
5th Floor, Capital House
42 Weston Street
London SE1 3QD
Tel: 020 7848 6648 International +44 20 7848 6648
Fax: 020 7848 6620 International +44 20 7848 6620
or 020 7848 6605 International +44 20 7848 6605
Opinions expressed are those of the author, not the institution.
* For searches and help try: