Re: st: RE: Matching Names

[David Bell <dcbell@iupui.edu>]

Dear Max,

I agree that 52000 is a lot of cases.  I've never had to deal with  
that many, but your method depends on your tolerance of bad/missed  
matches.  In my case, where we had to decide if a person named by  
respondent A is the same as a person with a similar name described by  
respondent B (where we had recorded gender, race/ethnicity,  
approximate age), we found that all mechanical matching algorithms  
were pretty bad.

My advice is to do it in two parts.  Match as much as you can (say  
using -soundex- as Kieran suggests), then eyeball the matches (along  
with any auxiliary information you have on them (like age, gender, or  
whatever is in your dataset).  Then do the same for the non-match  
side.  It's a lot of work, but you only have to do it once.  Of  
course, if you don't care about mismatches, go the mechanical route.   
I'd still eyeball at least a subset to get a mismatch error rate  
estimate.

Dave
====================================
David C. Bell
Professor of Sociology
Indiana University Purdue University Indianapolis (IUPUI)
(317) 278-1336
====================================




On Aug 7, 2008, at 6:10 PM, Kieran McCaul wrote:


> This is a big problem.
>
> You might want to investigate using soundex to help with matching  
> the misspelt names but, depending on the version of soundex that you  
> use, it may not be particularly useful.
>
> Michael Blasnik wrote an egen function to implement a soundex  
> algorithm a while ago for Stata 7.
> http://ideas.repec.org/c/boc/bocode/s420901.html
>
> You could try that.
>
>
>
>
> ______________________________________________
> Kieran McCaul MPH PhD
> WA Centre for Health & Ageing (M573)
> University of Western Australia
> Level 6, Ainslie House
> 48 Murray St
> Perth 6000
> Phone: (08) 9224-2140
> Phone: -61-8-9224-2140
> email: kamccaul@meddent.uwa.edu.au
> http://myprofile.cos.com/mccaul
> _______________________________________________
>
>
> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu 
> ] On Behalf Of Max Perez Leon
> Sent: Friday, 8 August 2008 5:03 AM
> To: statalist@hsphsun2.harvard.edu
> Subject: st: Matching Names
>
>
> Hello statalist users,
>
> I am having a big problem trying to merge to datasets with names.  
> The problem is
> that there are tons of typos in both datasets. Examples bellow:
>
> DATASET 1: --------------------- DATASET 2:
>
> NAMES--------------------------- NAMES
>
> LUIS P�REZ --------------------- LUIS P�REZ
> WILLIAM SMITH ------------------ WILLIAM SMITHSS
> JORGE F. CHOCAN ---------------- JORGE F CHOCANOS
> P. BROWN ----------------------- PAUL BROWN
> ENRIQUETA GAUDENCIA------------- ENRIQUETA G
>
> I could do it by hand but I have 52568 obs and more to come. I am  
> trying to
> establish a method using regular expressions so that I can merge  
> correctly the
> datasets.
> Any help will be very much appreciated,
>
> Thanks for your time,
> Max Perez Leon
> PUCP-IEP
>
>
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/