0
$\begingroup$

I am having trouble understanding how SUDAAN calculates adjusted column percents and I'm hoping someone can explain.

Let's say I have 3 variables, gender(M/F), happyness (binary happy/sad), and age group (so 18-30, 31-50 etc.). Age group has 5 levels. From the Census we also have a total population for each age group to adjust to.

I want to cross happyness by gender and then adjust the column percents (because it is more useful to know what percent of men are happy rather than what percent of people are happy men) to the Census levels of age group.

I understand how to calculate the age adjusted (not column) percentage. To find the adjusted percent for happy men I would calculate the percent of happy men that are in each age group, multiply that by the percent for that age group in the Census data (so if happy men ages 18-30 are 6% and ages 18-30 on the Census are 5.5% I would multiply $.055*.06$) then I sum the adjusted percents for the individual age groups, and I get the adjusted percent of happy men out of the total population.

How would I calculate an age adjustment for the percent of men that are happy?

Kind Regards

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

the second script on this page

http://www.asdfree.com/search/label/national%20health%20and%20nutrition%20examination%20survey%20%28nhanes%29

precisely replicates the cdc's direct method of age adjustment

http://www.cdc.gov/nchs/data/databriefs/db92.pdf

with very well-documented and easy-to-reproduce R code that you should be able to apply to any of the other census survey datasets listed on asdfree.com

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.