Re: Child marriage prevalence discrepancies [message #7084 is a reply to message #7083] |
Thu, 20 August 2015 20:24 |
Reduced-For(u)m
Messages: 292 Registered: March 2013
|
Senior Member |
|
|
My first thought would be that it could be an age-reporting issue in the data. Are you back-calculating their age at marriage from their age at survey time? One thing to try might be to drop women with age flags (they flag it if you can't remember certain parts of your age and they have to help you narrow in on it). You could also look at a histogram of reported ages and/or birth years to see if there are any funny spikes. In general I think that it is hard to trust some of the age reporting for people born more than 10 or 20 years ago - they just don't know very well. But maybe I'm totally off, so this is just a suggestion for something to check.
|
|
|