Using Ancestry: Exact Searches, by Juliana Smith

Exact Search.bmpIn last week’s newsletter, I began a series of articles that will take a look at the various search options available at Ancestry. This week’s installment will focus on Exact Search. The Exact Search stands in stark contrast to the Ranked Search, which we discussed last week. Where the Ranked Search is a bit “fuzzy,” the Exact Search is precise. It does exactly what we tell it to do–much like we’d all like our children to behave!

The danger here is that one piece of information not entered exactly as it appears in the database can throw the whole search. With Exact Search, less is more–the less information you add, the more hits you get. That’s why it’s best when doing an exact search, you want to start with just a few basic facts and then narrow your search slowly until you get a manageable number of hits. Rotate in and out different pieces of information, based on which search terms are more or less likely to be correct. For example, a given name is probably less likely to be misinterpreted by a transcriber than a surname, particularly an uncommon surname.

How Do I Do an Exact Search?
To start an Exact Search, simply check the box at the top of the search template that says, “Exact Matches Only.” As I mentioned last week, you want to be aware what type of search you’re performing because this option is “sticky”–it will default to the last search type you used. I try to make it a habit when I’m entering my search criteria to double check and make sure I want the type of search that is selected. Sometimes I swear gremlins come in and change it on me.

Exploring the “Less is More” Concept
To illustrate the “less is more” concept, let’s do a “play-along search” like the one we did last week. From the Advanced Search page, enter:

John Szucs, born in 1906 in the U.S.A. in Ohio

(You can click on the image above to view the results of this search.) All of these records pertain to my grandfather. There are two census records for him living in Cleveland, Cuyahoga County, Ohio; his death record; and a passenger arrival record showing him and my grandmother returning from an anniversary trip to Europe.

Now just for the heck of it, let’s add his county of birth–Jefferson County. Since I know that’s where he’s born, that can only help, right? Wrong. Adding Jefferson County eliminates all of the hits except his death record. Why? Because none of the other records index the birth county. The census will show the county he lived in for those years, but they only give the state in which he was born.

Making Exact Search More Flexible
There are several ways we can make the Exact Search more flexible. First, you could de-select the “Exact” box only on certain fields, but this will take you back into the world of Ranked Search, where you’ll see the results ranked by best match again, rather than the list of hits by database that you see when you do an Exact Search.

There is also a Soundex option available from the homepage that is only available when you have the Exact Matches Only box checked. (When you’re doing a Ranked Search, it will automatically include Soundex options, as well as some other name variations.) Soundex is also available on many of the individual database search pages as well.

Let’s do another “play-along” example, using the same gentleman I used in last week’s article–my great-grandfather, Raymond Dyer. On the homepage, enter

Raymond Dyer, Soundex (from the drop-down box that follows the name),
lived in U.S.A. and New York.

Last week with one Ranked Search we were able to locate Raymond in four of the six enumerations that he was alive for and that are currently available–1880, 1910, 1920, and 1930. With this one other search using Soundex, we pick him up in the remaining censuses–1870 and 1900. In 1870, his last name was indexed as Dyre, and in 1900 it is listed as Dyar.

Soundex Tip: Check the Soundex codes for common variations you’ve found for the surname you are researching so you know that your Soundex Searches are covering all the bases. For example, my great-grandfather’s last name appears on Mekalski on some records, and as Menkalski. I believe that the name was spelled with a diacritic over the e, which according to “First Names of the Polish Commonwealth” by William Hoffman and George W. Helon, the nasal “e” with the diacritic sounds like “en.” Checking the codes for these two variations, I find that Mekalski is coded M242 and Menkalski is M524. Because of the different codes, a Soundex Search for Mekalski wouldn’t turn up variations of Menkalski, so I’d have to do two searches to capture both. It’s a good idea to keep a list of name variations and the corresponding codes handy, so that you don’t miss any during your searches. You can get Soundex Codes for names by using the RootsWeb Soundex Converter. 

Wildcard Searches
Another way to loosen up Exact Searches is with wildcards. Ancestry currently allows for the use of either an asterisk (*) or question mark (?) in searches, although at present, neither can be used within the first three letters of the name. (Searches of that magnitude would eat up bandwidth slowing down or blocking searches for other users significantly.) The asterisk replaces anywhere between zero to six characters, and the question mark replaces one character. Since the asterisk would clearly cover instances where one character was misinterpreted, that’s the most commonly used wildcard, but if you were getting too many hits returned with a search, you might try the question mark to narrow that down a bit.

As an example, let’s use Mary Tob*n, born 1772 and since Mary liked to lie about her age, let’s give her a wide berth and also add some flexibility to the date field by selecting “+/ – 10” years. This finds my fourth great-grandmother in the 1850 and 1870 censuses, and her passenger arrival record as well. The variant spellings it picked up included the original spelling of Tobin, as well as Toban, Toben, and Tobbin.

Next Week
Next week, we’ll continue our look at searching options at Ancestry by getting a little closer to the databases, which allows us to really unleash some of the power behind Ancestry searches.

Click here for a printer friendly version of this article.

Juliana Smith has been an editor of Ancestry newsletters for more than nine years and is author of “The Ancestry Family Historian’s Address Book.” She has written for “Ancestry” Magazine and wrote the Computers and Technology chapter in “The Source: A Guidebook to American Genealogy,” rev. 3rd edition. Juliana can be reached by e- mail at Juliana@Ancestry.com, but she regrets that her schedule does not allow her to assist with personal research.

 

3 thoughts on “Using Ancestry: Exact Searches, by Juliana Smith

  1. Good advice except my brick wall is finding the parentage of my Great Grandmother Mary Story born in So Boston in 1823. I even have the street address except that there are over forty hits when I use my “Narrow Streets of Boston” disc.

  2. As I am continuing to learn the “Art of Ancestry Searches” I am also beginning to find out that those old family rumors that have been passed down through the generations have a lot of substance and fact to them. I remember some stories about the 1800′s and 1900′s that I heard when I was a child – not much on detail, but memorable. For instance – my ancestors on both sides came from Wales – True! We had an Indian Ancestor – almost sure this is true. Some notable figures in public offices from early America were our ancestors – some were coal miners – some were Quakers – some were slave holders – most were not – some freed their slaves long before the Civil War – Some were veterans of Union Army – true – some of Confederate Army – right. In delving through the past I have learned that we had heroes and zeroes – good and bad – what they did – where they lived – where they came from – what they owned! Many sad facts of early deaths – Indian wars and casualties – And how MANY children they had, and how FEW survived to adulthood. All in all, I have a lot already to pass down to my grandchildren and great grandchildren so when they ask me – “Granma, where did I come from?” I can tell them about their heritage and leave the answers to that specific question up to their parents. Many, many thanks! I hope to finish this history, right down to my great-grandchildren, while I can still research and write.

  3. I appreciate your effort to explain the Ancestry programmers’ rationales for how the search system is set up. However, I think you may give some users the wrong impression when you imply that exact search can always be deselected for a specific field in advanced search. I often find only the one Exact Search check box in the upper left-hand corner.

    Many times, I KNOW that a detail is exactly true, yet must accept between 800 and 50000 hits because setting exact search covers ALL the fields. In your explanation, you suggest that if an item in the search definition is not indexed, and search is exact then the answer will be no hit. Isn’t this a poor design decision. If a field is not in one of the files searched, or that field isn’t indexed, then cast out the field in the search form, not the records from the result. (I hope you find this intelligible, it is hard to use popular terminology and be precise.)

    We need more complex advanced search forms, with the possibility of making each field exact or not, and the confidence that if a datafile cannot handle the requirements of a field, the search process will ignore the request, not fail to give me records – you could even have the notice on the search result, for example – “This data set does not contain nationality data. If you are searching on nationality, this search result does not take account of your request.”

    Good luck in improving the search system – I love the results so far but I find it awfully frustrating sometimes.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>