Some people eat, sleep and chew gum, I do genealogy and write...

Friday, October 11, 2013

Searching the Web: Specific to General or General to Specific?

the normal reaction of a genealogist confronted with an advanced search form in is to start filling in all of the blanks with information about the ancestor. My initial question is always, if you know all that much about the ancestor, why are you searching for information in But moving along, contrary to the average impulse the best search strategy is to put in as little information as is necessary to find the ancestor. I usually start with name, residence and gender. Then if that doesn't work, I begin adding one piece of information at a time, such as the spouse's first name and maybe a date. Each time you add one piece of information, you do the search over and see the results before editing the search parameters again by adding one more piece of information.

Why is this better than putting in all the information you know? The answer is not really very simple and depends on the way the search engine is designed. Usually, the search engine (i.e.'s search or or whatever) will try to match as many fields as possible with records in the database. That will usually result in either false positives (responses that look similar to the one you are searching for but are not the right person) or no responses at all. Let me illustrate this with a couple of searches on

The first search will be for my Great-Grandmother, Eliza Ellen Parkinson Tanner. In the first pass, I will put in only her first name, no surname, the location as Arizona, her gender, and her spouse's first name Henry.

That's it. That's all I put into search on Here is a screen shot of the search fields filled in with this information:

Now here are the results of the search:

You can see that my Great-Grandmother and her family came up on the first page of results in the 1930 U.S. Census. Could I force the entries to the top? Yes, probably by adding the surname. But why do that when you have found the family?

Another example of searching from the general to the specific is when using the Research Wiki. When searching in the Research Wiki, you are not looking for a specific person, you are looking for types of record sources that might help you find an ancestor. The Research Wiki is an analog of the real-world way of recording and storing records. The key to searching on the Research Wiki is to work from the general to the specific. If you wanted to find a marriage record for your ancestor who lived in the United States, you would start with an article about United States Marriage Records. Then move to the State, the County and ultimately a smaller jurisdiction. In this way, you would learn about marriage records and be knowledgeable enough to find any actual records or substitute records that might exist in the jurisdiction where your ancestor got married.

Learning to search the Web is a skill, just like any other learned skill, such as playing piano or swimming or riding a bicycle. There are some genealogists who have a talent for searching, just as there are those who have a talent for music or other skills, but nearly everyone can learn to search more skillfully.

1 comment:

  1. We agree about starting with a general search and going more specific. I have a practical reason for that - suppose you do a tight search and get a single answer. The temptation is to believe that is THE one. But if the search were slightly more generic, then you might have seen that there's another 2 records satisfying your criteria just up the road. Doing the more generic search first, you're starting your analysis with the question - which is it? Doing the more specific first, you might never look further.

    Just one point - I would ALWAYS set "Match all terms exactly". Maybe Ancestry's attempts to provide inexact-but-relevant matches work better on US datasets but I find that for the UK records the inexact-but-relevant matches often put poor matches ahead of decent marches.

    Setting exact matching also means it pays to use the more complex settings for the name items