A Wildcard Search That Works Poorly on the Fulton History Site

Cliff Lamere     30 Oct 2015

_________________________________________________________

The purpose of this webpage is to continue a discussion begun on a webpage which explains the use of boolean searches and wildcards to get better results on the Fulton History newspaper website.  The * wildcard does not work as it was intended to work on that site.  Here are some examples of how it should work.

   *          examples:  justi*     ca*ed     Em*eret*e

The * is supposed to represent any number of letters, including zero.  More than one * can be used in a word.  Words found by using justi* would include justice, justify, justification, Justin, etc.  In the second example, the words could be cared, cashed, catalogued, catapulted, etc.  If you are searching for a woman named Emmerette and want to be sure to get hits even if the name is misspelled with only a single m or t, you can do a search for Em*eret*e.  But, how well does the * wildcard actually work?  I describe my research next.

As seen on the boolean webpage from which you may have come, Gard?n?r got useful hits that had six different spellings.  However, there were more than the maximum of 5000 hits, so only 5000 showed.  To test the * wildcard, I had to reduce the number of hits so that I could see when the hits increased or decreased.  I used the w/ wildcard (followed by a number).

Gard?n?er w/9 married     (364 hits)

Adding a third ? got the following results:

G?rd?n?er w/9 married     (449 hits)

The second search should get more hits, and it did, because it found both Gard- and Gord- (the latter is a common spelling).  Those two searches worked as expected.  Then, I substituted * for the first ?.  That search should have given me even more hits because it would find Guard- (two characters) in addition to what the other searches found.  But...
 
G*rd?n?er w/9 married   (113 hits)

Replacing the first ? with a * returned 336 fewer hits, or a 75% decline.  This is a disastrous result, but could it be caused by using more than one kind of wildcard in a single word?  Let's find out by trying a different search with only the * wildcard used.

Henrietta Jones (an exact phrase search)     (1408 hits)

Henrieta Jones (an exact phrase search)     (2 hits)

Henrie*a Jones (an exact phrase search)     (321 hits, a 77% decrease)

So, we can conclude that the * wildcard is very unreliable.  It will return some hits, but you should not decide that a wedding announcement, obituary, or other article is absent from the Fulton History site based on a * wildcard search.  I will no longer use it.

_________________________________________________________

The webpage you are on is part of a series of webpages about the Fulton History newspaper site.  It is written to help you make more productive searches.

Using the Fulton History Newspaper Site

Boolean Searches and Wildcards on the Fulton History Newspaper Site

A Wildcard Search That Works Poorly on the Fulton History Site    (the page you are on)

Visitors since 31 Oct 2015