NARA's 1950 census website development team has made an important improvement to the name search feature:  Names transcribed by humans now clearly rank higher in search results than names transcribed by optical character recognition (OCR).  What does this mean?  Let's look at a few examples.

 

Example 1:  Thomas Ives, Ohio

  • Before Transcription:  The OCR search results for Thomas Ives in Ohio returned thousands of potential "matches" that reflected the inherent problems of OCR trying to read many different handwriting styles - 442 results for Ives, Iles, Ines, Lives, Iues, Gives, Iver, Ides, Dives, and other "similar" names, plus more than 10,000 persons named Thomas.

 

  • After Transcription:  A human transcribed the name of Thomas A. Ives in Cleveland, Ohio.  Now, a person searching for Thomas Ives in Ohio will receive this result as the first potential match!   Note, however, that the "Matched Name(s)" above and below the image of the census page will still show the original imperfect OCR results:  "Thomas; Sues Thomas a.; Thoms."  The OCR misread "Ives" as "Sues."  Of course, you still need to look at the census page to verify this is the person you wanted and get the information. 

 

 

Search Result for Transcribed Name:  Thomas Ives in Ohio
Search Result for Transcribed Name:  Thomas Ives in Ohio (ED 92-141, Cleveland, Cuyahoga County, Ohio)

 

 

Example 2:  Irma Schneid, Ohio.

  • Before Transcription (May 9, 2022):  A search for Irma Schneid in Ohio returned 4,361 potential matches.  Narrowing the search to Cleveland, Cuyahoga County, Ohio, reduced the number to 447 potential matches, but the "right person" was ranked only 73rd! 
  • After Transcription (check on or after May 11, 2022):  A search for Irma Schneid in Ohio will return Irma L. Schneid, Sheet 78 (image 25), ED 92-141, Cleveland, Cuyahoga County, Ohio, as the first potential match!

 

Important Takeaways: 

  • Thank you for your transcriptions!  They matter!  They significantly improve the search results!
  • Narrowing your name search to include state and county is always better if the name was significantly misread by the OCR, has not yet been transcribed, or contains common names (John, Smith, and so forth!).