Friday, September 6, 2013

Ancestry.com "New Search" Results With Wild Card Names

In the present series, I have compared "Old Search" results with "New Search" results n Another Ancestry.com "Old Search" vs. "New Search" Comparison (assuming not exact matches on names), in Ancestry "Old Search" and "New Search" Comparison with Exact Matches (with exact matches on names); both studies used my third-great-grandfather, Henry Carringer (1800-1879) as the subject.  

In this post, I want to show the power and usefulness of using wild cards for names and using the flexibility of the locality settings for events.

What is a "Wild Card?"  On Ancestry.com, the user can use "?" to represent one character in a first or last name, and "*" to use any number of characters (including none) in a first or last name.  However, there have to be three non-wild characters in the name, and one of the non-wild characters has to be to first or last character in the name.  So, for the Carringer name:

*  The following wild card names will return results:  Car*, Car*ger, C*ger, C*g?r, C*r*n*r, ?a*ng*r, C?r*ng*, ???*ger, etc.  
*  The following wild card names would not return results:  ?arring*, C*r, ???*er, etc.

Why would a user want to use a wild card for a name?  My main reason is to find records with misspelled or misindexed names.  However, if a user does not add other search terms to narrow the search a bit, the user may be inundated by record matches that are not their target person.  So there is a "middle ground" for using wild cards - use the narrowing search terms but keep them not exact so that you don't limit the search too much but you drive the best matches to the top of the Results list.  

I will use Henry Carringer (1800-1879) as the example again.  In the previous posts, I found 7 records in the Ancestry databases that applied to him.  Let's see if I can find him again with one search, and with his records near the top of the Results list:

My usual practices include not using many vowels in my wild card searches, and not using double consonants or a long string of consonants.  Users should think about how the name sounds - what consonants are heard when it is spoken, and which letters can be misinterpreted when handwritten.

A)  I decided to start with a relatively simple set of name wild cards:

*  First name:  He*n*r* (because he was of German descent, there may be records for "Henrich" or "Heinrich")
*  Last name:  C*ger (because one or more of the consonants in the name may be missing).

I also narrowed my search by adding these Facts:

*  Birth Fact, Year: 1800 plus/minus 2 years (not exact); Birthplace: Pennsylvania, USA (default settings, which means not exact)
*  Death Fact:  Year: 1879 plus/minus 2 years (not exact); Birthplace: Iowa, USA (default settings, which means not exact)

I chose to make these Fact settings "not exact" so as to avoid the situation found in the "Exact Match" post - if the record does not have the Fact, the results are not matched.

I also chose only "Historical Records" and "Stories and Publications" in the "Collection Priority" section of the search fields.

Here is the screen with the filled-out search fields:


Here are the results for this specific search using wild card names (2 screens shown):



There were 9,697 matches found for this search.  I have shown the "Records View" above; a user could choose the "Categories View" and then checked each database listed.

Here are the database records I found for my Henry Carringer (1800-1879) with the approximate number on this list of 9,697 matches:

1)  1870 U.S. Federal Census: Henry Caringer, born 1801 in PA
2)  Web, Iowa Find A Grave Index, 1800-2012:  Henry Carringer, born 1801, died 1879
3)  Iowa Cemetery Records, 1662-1999: Henry Carringer, born 1798, died 1876
6)  1860 U.S. Federal Census: Henry Carringer, born 1800 in PA
9)  Web, Rootsweb Cemetery Index, 1800-2010: Henry Carringer, born 1801, died 1879
73)  Selected U.S. Federal Census Non-Population Schedules, 1850-1880:  Henry Carringer, 1870
140)  1840 U.S. Federal Census:  Henry Carringer
153)  1830 U.S. Federal Census:  Henry Carriger

I looked through the first 500 matches and stopped.

B)  My second search using name wild cards was similar.  I added a letter to the last name - the search terms I used were:

*  First name:  He*n*r* (because he was of German descent, there may be records for "Henrich" or "Heinrich")
*  Last name:  Ca*ger (because one or more of the consonants in the name may be missing).

I also narrowed my search by adding these Facts:

*  Birth Fact, Year: 1800 plus/minus 2 years (not exact); Birthplace: Pennsylvania, USA (default settings, which means not exact)
*  Death Fact:  Year: 1879 plus/minus 2 years (not exact); Birthplace: Iowa, USA (default settings, which means not exact)

I chose to make these Fact settings "not exact" so as to avoid the situation found in the "Exact Match" post - if the record does not have the Fact, the results are not matched.

I also chose only "Historical Records" and "Stories and Publications" in the "Collection Priority" section of the search fields.

Here are the search Results (again in the "Records View", first 3 screens shown:





Here are the database records I found for my Henry Carringer (1800-1879) with the approximate number on this list of 599 matches:

1)  1870 U.S. Federal Census: Henry Caringer, born 1801 in PA
2)  Web, Iowa Find A Grave Index, 1800-2012:  Henry Carringer, born 1801, died 1879
3)  Iowa Cemetery Records, 1662-1999: Henry Carringer, born 1798, died 1876
4)  1860 U.S. Federal Census: Henry Carringer, born 1800 in PA
5)  Web, Rootsweb Cemetery Index, 1800-2010: Henry Carringer, born 1801, died 1879
8)  Selected U.S. Federal Census Non-Population Schedules, 1850-1880:  Henry Carringer, 1870
15)  1840 U.S. Federal Census:  Henry Carringer
16)  1830 U.S. Federal Census:  Henry Carriger

I looked through all 599 matches and stopped.

C)  It is apparent that the second search with name wild cards found all of the matches on the first page of matches.

Interestingly, this name wild card search found the 1870 Non-Population Schedule entry for Henry Carringer, which did not have the birth date or birthplace indexed.  That match was found in the list for the last post, but was not on the list for the first post (I didn't go far enough down the list).

Interestingly, neither search found the 1850 U.S. Federal Census for Henry Casinger because the age was indexed as 36 rather than 50.  I thought that the results would go further afield than just the range of years I requested with the box checked as not exact, but they didn't.  If I had expanded the birth year range to, say, 1795 to 1815, the 1850 census result would have been found; however, there would have been more total matches.

D)  The conclusions I draw from this study include:

*  Different wild card searches produce different results.  In general, the more characters you use, the fewer matches you receive.
*  If the user uses too many alphabetical characters, all of the relevant results may be limited even with wild cards.
*  Narrowing the search using "not exact" dates and locations should drive the relevant matches for the target person to the top of the "Records View" match list.
*  Narrowing the search using "exact" dates and locations may miss relevant records.
*  For results with many matches, the "Categories View" might be easier to work with than the "Records View"
*  If the results do not find an expected record for the target person (such as the 1850 U.S. census for Henry Carringer), then search that database separately and use all of the tricks (exact name, wild cards, expand the date range, remove birthplace, first name only in a county, etc.) one at a time or in combination.  You may find them!

The URL for this post is:

Copyright (c) 2013, Randall J. Seaver


No comments: