Hi everyone. I am new to Statalist.
I am using two timepoints/datasets from a cross-sectional public survey to calculate change in prevalence of a variable (binary variable: 1=used, 0=did not use).
Will it be appropriate to use prtest to calculate the difference in prevalence and generating a p-value across the two timepoints?
Of what I know, in order to do a trend analysis, I will need >=3 timepoints.
In the literature, it seems more common to use a logistic regression and generating odd ratios to compare prevalences. Sorry if I am mixing two different things.
I am using Stata 14.2. I will appreciate any help. Thank you.
Omar.
I am using two timepoints/datasets from a cross-sectional public survey to calculate change in prevalence of a variable (binary variable: 1=used, 0=did not use).
Will it be appropriate to use prtest to calculate the difference in prevalence and generating a p-value across the two timepoints?
Of what I know, in order to do a trend analysis, I will need >=3 timepoints.
In the literature, it seems more common to use a logistic regression and generating odd ratios to compare prevalences. Sorry if I am mixing two different things.
I am using Stata 14.2. I will appreciate any help. Thank you.
Omar.