If you search online for newspaper articles, you know that it is often challenging. This is primarily because OCR is not perfect and that old newsprint can often have creases, tears, ink spots and other errors. This creates an index that often is filled often with a mumbo jumbo of letters that do not result in accurate names and the like when the index is created. Thus your searches may not include all the articles that you seek.
This can easily be remedied by YOU! Just change the letters in the search criteria and DELIBERATELY misspell. Refer to this collection of letter pairs as shown in the image below. You will get much better results, guaranteed. I have seen as much as a 20% increase in VALID results.
Read and print out the following image for a Quick Reference Sheet Summary or if you prefer, download it in pdf format at the link below.
Click below to download the pdf file
| misspell_tipsheet.pdf |
Visit the Newspapers! Page for many more details, tricks, tutorials, webinars and articles intended to allow you to become a master newspaper researcher!
Really interesting! Have you found that these ocr errors are somewhat consistent from site to site, or do some sites exhibit some of these errors more than others?
Also, how did you get the 20% number?
Bill, It varies from site to site. Some seem to be better than others. I got the 20% number from my own search statistics. I took a couple of my surnames in my tree spelled exactly correctly and then started varying the letters per the letter pairs that I mentioned and it cam to be about 20%. It depends on the complexity of the surname and the letters that make up that surname.