Are You Using the Right Search to Find Your Ancestors?

A search box is a search box, right? Actually, not so much. Not all searches act the same way. Understanding the differences in them can help you find more ancestors as well as discover more resources to use in your genealogy research. Let's take a look.

Are You Using the Right Search to Find Your Ancestors

More...

The Search Depends on the Type of Record

I've seen people on social media lately talking about searching in newspapers. Digitized newspapers where you can do a full-text search are wonderful! I've made so many discoveries using them. But the search strategies we use for a database (like a census index) usually don't work for a full-text search. 

Databases use what's called "fielded data." Specific types of information are put into specific fields. In the case of a census index, the surnames are in one field, first names are in a different field, age is in another field, etc. The data is nice and organized.

In this search form on FamilySearch to search the 1891 England census, I can search for John Ramsey and it will look specifically in the first name field for John and the surname field for Ramsey. It won't look for anyone with the surname of John or the first name of Ramsey. It also won't search in the location field for either of those terms.

Census of England and Wales, 1891, Search

This is an example of fielded data. Specific words are searched for in specific fields. (Image courtesy FamilySearch.)

Fielded data makes it easier to manipulate the data. We can do things like wildcard searches or give a range for the birth year. 

A Big Difference Searching Digital Texts

However, when we're using a digitized text, like a digitized newspaper or book, we're not dealing with different types of information in their own fields. We're using "full-text search." Essentially, we're looking for strings of characters, regardless of where they are found. That has two major implications for our searches:

  • The text has to match what we searched for
  • The text could be used in any meaning

Take a look at two marriage announcements from the Shepherdstown (West Virginia) Register, available on Chronicling America:

Newspapers full-text search

Shepherdstown (West Virginia) Register, 2 February 1878, page 2. Image courtesy Chronicling America.

The newspaper is searchable, but it is a full-text search. If I had searched for Margaret Kline, it would not have found this entry. Full-text searches look for matches of the character strings. Margaret is not the same as Maggie in a full-text search. Similarly, if I had searched for James Gregory, it wouldn't have found the second announcement because James does not equal Jas.

Searching in digitized texts, like newspapers, isn't the same as searching in a database. 

Click to Tweet

Newspapers.com is a notable exception. You can use wildcards there to get a wider range of results. A search for br*n will give results for both brown and brethren (because they both begin with br and end with n). However, this is the exception. The vast majority of full-text searches do NOT handle wildcards. (Remember, too, that even use a wildcard like Mar*t to get variant spellings of Margaret wouldn't have given us this record with Maggie.) 

Full-text search also has the challenge of words being in any context. A search in a newspaper for John Ramsey could bring back references to people with the last name of John, references to Ramsey County, etc. 

Understanding the Database

When you're searching anything, not only do you need to consider if it's fielded data (like a census index) or digitized text (like a newspaper or yearbook); you also need to consider what it is that you're searching. What is included?

A common mistake is to treat a library catalog like you would search on a genealogy website. An online catalog is not the best place to search for a specific name. Catalogs don't include all of the names within the book; catalogs list what the book is about.

I can search for John Peter Kingery in a library catalog, but unless a book has been written about him, I'm not going to find anything. (I will get a few results if the catalog looks for John and Peter and Kingery across the entire catalog entry, but that's not the most meaningful set of results.)

A search in a library catalog or WorldCat is much more efficient if you search for what the book is about, rather than what you hope to find in it. One of my strategies for finding published family histories is to use "surname family" without the quotes. If I want to find books about the Kingerys, I would search for kingery family in the catalog. (If you want more strategies on finding more genealogy materials, check out my article, "How to Find Hidden Genealogy Resources in Libraries and Archives.")

Match the Search

Online resources open up countless opportunities for our genealogy research. We can get the most out of them when we match our search to what it is we're using. Consider if it's fielded data (like a database) or digitized text (like a newspaper). Also change your strategy when you're using library catalogs. You'll end up making more discoveries. 

Using the Right Search to Find Your Ancsestors
Posted: March 29, 2018.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

  • Another great blog post! Imagine my “hits” when i search for my ancestor with the last name “Loving.” LOL. Luckily, this family has been well-documented by others, once I got past a few generations!

  • Thank you for these tips. I have discovered this difference myself but need to remember it better because I tend to slip back into the standard “fielded” search method most often, even in OCR records

  • Regarding your comment:

    “Full-text search also has the challenge of words being in any context. A search in a newspaper for John Ramsey could bring back references to people with the last name of John, references to Ramsey County, etc. ”

    This is true but often not a problem as most search engines will perform page-ranking calculations and display pages with “John Ramsey” higher in the search results than one that references someone with the last name of John and “Ramsey County” somewhere else on the page. So, yes, it will be in the search results, but probably on page 10 rather than a much better hit on page 1.

    i.e. Searching for “John Ramsey” on Google it reports that there are “About 82,800,000 results (0.69 seconds) “. But because of page-ranking the better results float to the top and we normally don’t care about search results past page 1 or 2.

    • I’m referring to full-text searches (like newspapers and books that have been OCR-ed). While some sites and databases that use full-text searches will bring back the more relevant results first, many will not.