11 August 2013

Elephind Finding All that the Library of Congress Does?

On Genealogy Tip of the Day's Facebook page, we've discussed Elephind, a webiste which searches free digital newspapers at several sites at once--including the Library of Congress.

And this is where I am confused. I searched Elephind for "rampley" and the Library of Congress digital newspaper site for "rampley." The results were different as shown below with Elephind obtaining 87 hits and the Library of Congress obtaining 168. I am not certain why there is a difference, but this is somewhat frustrating as I liked Elephind's display of the text in which the reference appeared. But if the number of hits fewer than on the Library of Congress site directly, I'll have to go with the LOC site.

Suggestions for what I may be doing wrong would be appreciated-if I'm misinterpreting something I'll happily post a corrected update.





8 comments:

Anonymous said...

It may be that Elephind counts only the first instance of the search term in a document and then moves to the next document, while LOC counts every instance of the search term.

Lisa Suzanne Gorrell said...

I don't know what you're doing "wrong" but I thought I'd try it, too. I searched "Wollenweber" and got 421 hits on Chronicling America and only 208 hits on Elephind.com. I, too, liked how the results are displayed on Elephind.com. If only I could be sure I was getting all the results possible.

Anonymous said...

There's a comment on the page at http://www.bloodandfrogs.com/2013/06/the-ring-of-trust.html that explains the diffrence between results on Chronicling America and Elephind.com. Scroll down and expand the comment by Stefan Boddie, which attempts to explain the "fuzzy searching" used by Chronicling America, and why it returns a lot more results than the "not fuzzy" searching used by Elephind.com. Unfortunately most of the "fuzzy" results don't contain the words you're searching for!

Anonymous said...

It's my understanding that Elephind doesn't yet search all of the content on Chronicling America, although they're gradually linking to more and more items. So that might partially explain why you don't get as many hits.
But also remember that Elephind searches things that aren't in the Chronicling America collection, both in the US and overseas, so it's still worth taking a look.

---
Deb Stock
UK

lois said...

http://blog.eogn.com/eastmans_online_genealogy/2013/08/search-historical-newspaper-archives-with-elephindcom.html

You can also check out above and his comments...seems the advanced search much better?

lois said...

http://blog.eogn.com/eastmans_online_genealogy/2013/08/search-historical-newspaper-archives-with-elephindcom.html

Read above his comments and using the advance search.

Michael John Neill said...

Actually advanced search isn't the problem when only one search term is entered as was done in this case, "rampley." Advanced searches at Elephind may be more effective when multiple search terms are used, but the number of matches for unusual last names should be the same and they weren't in this case. That's what makes me think there's something not quite complete about Elephind.

It may be as some suggested that not all LOC newspapers are indexed in Elephind.

Advanced searches would be suggested for more common names or combinations of multiple search terms.

Results viewed on LOC when going through Elephind also do not highlight the resulting terms on the newspaper page as is done when LOC is searched directly.

Anonymous said...

As someone noted already in one of the comments above the difference between the two is that Chronicling America uses "fuzzy searching", while Elephind.com does not. That means Chronicling America finds newspapers that have "similar" words to those you searched for, as well as finding those that contain the exact words you searched for.

The trouble with fuzzy searching is it returns lots of items that aren't relevant.

So Elephind.com does contain all the newspapers from the LOC, and searching does work as expected. It just doesn't return all the extra "fuzzy" results that the LOC site does.

There's a discussion of this at the link below (scroll down and read the comments).

http://www.bloodandfrogs.com/2013/06/the-ring-of-trust.html