Tuesday, July 2, 2013

Searching Records From Within An Ancestry Member Tree

I've been creating a presentation on Ancestry Member Trees, so I thought that I should investigate all of the features of this online family tree, which has over 46 million trees and over 46 billion profiles.

Since Ancestry.com recently claimed that 50% of all searches were started from an Ancestry Member Tree, I thought it might be smart to see how it works and if I "like" it.  It's not clear to me if those 50% are clicking on a green shaky leaf, or are actually clicking the "Search records" link.  My guess is that for 90% or more of that 50%, it's the leaves, but I don't know for sure.

I've written about the green "shaky" leaf "Hints" before, so I wanted to see what happens when I click on the "Search records" link.

1)  I decided to search for a female for this study, since they usually have records in more than one surname.  Here's a screen of the person Profile of my 2nd great-grandmother, Lucretia Townsend (Smith) Seaver (1828-1884), who married Isaac Seaver in 1851:


The "Search Records" link is in the red box on the screen above.

NOTE:  Only the Owner and Editors of an Ancestry Member Tree can see the "Search Records" link.

2)  clicking the "Search records" link opened the Ancestry Search tab and provided over 9 million matches:



Whoa...too many.  But let's look at them first (I won't show you the entire list!).  From the top, the Search found these records:

*  Massachusetts Marriage Records, 1840-1915 database - marriage to Isaac Seaver.  Good one.
*  Massachusetts Town and Vital Records, 1620-1988 database - marriage to Isaac Seaver.  Good one.
*  Massachusetts Town and Vital Records, 1620-1988 database - marriage to Isaac Seaver.  Same as above.
*  Massachusetts Death Records, 1841-1915 database - death record.  Good one.
*  Public Member Photos & Scanned Documents database - 1870 census image.  Good one.
*   Public Member Photos & Scanned Documents database - 1851 marriage certificate.  Good one.
*   Public Member Photos & Scanned Documents database - death record.  Good one.
*   Public Member Photos & Scanned Documents database - photo.  good one.
*   Public Member Photos & Scanned Documents database - marriage to Isaac Seaver.  Good one.
*   Public Member Photos & Scanned Documents database - 1860 census image.  Good one.

*   Massachusetts Town and Vital Records, 1620-1988 database - birth record.  Good one.
*  1880 U.S. Federal Census database - Isaac Seaver household.  good one.
*  Massachusetts Marriage Records, 1840-1915 database - marriage to Isaac Hopkins.  Wrong.
*  Web: Massachusetts Find A Grave Index, 1620-2013 - cemetery memorial.  Good one.
*  Massachusetts Town and Vital Records, 1620-1988 database - marriage to Isaac Seaver.  Duplicate.
*  Massachusetts Town and Vital Records, 1620-1988 database - marriage to Isaac Seaver.  Duplicate.
*  Massachusetts Town and Vital Records, 1620-1988 database - marriage to Isaac Seaver.  Duplicate.
*  Massachusetts Town and Vital Records, 1620-1988 database - marriage to Isaac Seaver.  Duplicate.
*  Massachusetts Marriage Records, 1840-1915 database - marriage to Isaac Seaver.  Duplicate
*  Massachusetts Marriage Records, 1840-1915 database - marriage to Phillip Sprague.  Wrong

*  1850 U.S. Federal Census database - Isaac Seaver household.  Good one.
*  1880 U.S. Federal Census database - Isaac L. Smith household.  Wrong.
*  Massachusetts Town and Vital Records, 1620-1988 database - death of Andretia D. Seaver.  Good one, but duplicate.
*  1870 U.S. Federal Census database - Isaac Seaver household.  Good one.
*  Massachusetts Town and Vital Records, 1841-1915 database - marriage to Phillip Sprague.  Wrong
*  1860 U.S. Federal Census database - Isaac Seaver household.  Good one.

The rest of the over 9 million matches are not my Lucretia (Smith) Seaver (I think, it's really hard to tell!).  Even though there were 4 wrong matches, I can understand how the wrong ones showed up because of the name, and birth date, and in one case, the parents names.  In those cases, the results satisfied the search algorithms used.

Of the 26 matches listed above, only 4 were not for the correct person.  Several were duplicates (perhaps caused by alternate names in the index), and some were image links to the same person in another of my Ancestry Member Trees.  4 out of 26 is 15%, so 85% of the matches were good.  That's in line with my previous experience (of 80 to 90% accurate matches).

Unfortunately, there are no other Ancestry Member Trees with an image for Lucretia, but there's always the possibility.

I didn't see much logic in the order of the matches - I would expect to see Record matches first, then Photo matches.  There were no matches for other Ancestry Family Trees (there are several for her).  Funny, but they show up in a "Categories" search!

3)  I was curious to see the Search fields for this search (since I just clicked on the "Search results" link).  So I clicked on the green "Edit search" button (top of left-hand column) and the search fields displayed in a popup window:


You can see that the Search fields are filled in for:

*  First and Last Names (including her married name)
*  Birth Event
*  Death Event
*  Marriage Event
*  Family Member names - father, mother, siblings, spouses, children.

Further down, all of the major collection types (Historical Records, Stories & Publications, Family Trees, Photos & Maps) were checked.

The good news is that it found records for both of her surnames - the maiden name Smith and the married name Seaver.

4)  I decided to see what an Exact search for the current search fields (fully filled in) would produce:


Nothing.  Not any Ancestry Member Trees.  I wonder why?

Of course, I didn't expect any Record Matches with an Exact Match and the current search fields, because there are no records in all of Ancestry.com with all of that information indexed for a person.

5) So how do I check to see if there are any more matching records in those 9+ million matches?  Well, I can do an "Exact Match" search with a modified set of search fields.  I edited the search fields to include only:

*  First name = "lucretia," last name = "smith"
(  Birth in "1828" plus/minus "2" years, location in "Massachusetts, USA"

Here is the search form:


And the results of the "Smith" exact match search:


Only 24 - I can work through those easily - most of them are probably the same as the Not exact search results.

6)  Because Lucretia married, I needed to search for her married name also, so I edited the search fields to include:

*  first name = "lucretia," last name = "seaver"
*  birth in "1929, plus/minus "2" years, location "Massachusetts, USA"

Here is the list of results:


Again, some of these are the same as the search with not exact search fields.

7)  This "Record Search" from within an Ancestry Member Tree only "works" when the user is in his/her own Tree.  You can't search from an entry in someone else's tree.  But the search results seem to be exactly the same as if the user started on the "Search" tab on Ancestry.com.

Do I "like" this?  Yes, it seems to do the job pretty well.  It finds LOTS of matches, but the closest ones are, generally, for the correct person (80 to 90%).  Some people have said that the search uses links to records previously attached to another Ancestry Member Tree that includes the person being searched.  I don't know if that is correct, but it makes sense to me for Ancestry.com to do it.

During the discussions about "Old Search" vs. "New Search" on blogs, Facebook and elsewhere, I kept seeing complaints that "New Search" "...finds millions of matches and I don't have time to sort through them all."  Frankly, "Old Search" finds the same record matches, but doesn't tell you how many there are!  Does knowing how many matches there might be scare some searchers off the hunt?

The URL for this post is:  http://www.geneamusings.com/2013/07/searching-records-from-within-ancestry.html

Copyright (c) 2013, Randall J. Seaver

3 comments:

Smadar said...

Nice review Randy! I agree with you, I "Like" searching from a member tree. I always beging that way. Usually, their top recommendations are applicable and easy to search through. Only after I work through their suggestions, do I go back and modify the search, change spellings, take away locations etc. For example, if an ancestor spend very little time in New York and has a very common name, I might get so many New York listings that are not relevant, and which push the Vermont listings where he lived most of his life way down the list. So I take out New York and voila!

Crista said...

Randy,

That 50% of searches that begin from a tree are actual historical record/tree searches just like you did here, not hint views.

-Crista

Anonymous said...

Personally, I use the hints and determine whether or not they are for my person. If they are I attach them. If they aren't, I ignore them. Then I click on the search records from that person's profile. This usually turns up a couple of extra hints, unless it is a person that had lots of hints. If the search records don't turn up results in databases where I would expect to find an individual, I then edit the search for that specific database. Finally, if I know there are databases that could contain my individual, but didn't hint or turn up in the search records, I navigate to that database using the card catalog or quick links on my Ancestry homepage, and play around with search for that particular database. Admittedly, I've only been researching for a couple of years, but new search makes sense to me, and generally has done a pretty good job. The only problem I've had, which has just been recently, is that, sometimes, using some of the exact search options gives me a result of zero good matches. This must be a glitch in the software, as, later, I can do the exact same search and get results. I'm not certain if it has something to do with traffic or what.