Advanced Research

Stop Using The Within Connector

Look at this query:

           “income tax” w/10 fraud

In law school a Westlaw Rep taught you that the foregoing query returns any opinions where the phrase ‘income tax’ appears within 10 words of the word ‘fraud.’

This is a true statement. But it’s not the end of the inquiry. The query does something else. Something arbitrary.

What you are really doing with this query is asking the search engine to exclude any judicial opinions where the term ‘income tax’ appears within 11, 12 or 13 words of the term ‘fraud.’ You are saying: “Eleven words apart? I have no interest in such opinions. How could such a case possibly be relevant?”

Subscribers to WestlawNext and TheLaw.net who use the Within Connector are actually overriding the default algorithm. At TheLaw.net, our algorithm already considers proximity of search terms in relation to each other, and ranks them accordingly. So, that opinion you arbitrarily asked the search engine to exclude — that opinion with 11 words of separation — you’re going to see that opinion near the top of your results, by simply using the And Connector.

The correct query, therefore, is:

          “income tax” and fraud

In the results, each opinion receives a percentage ranking in relation to the opinion that appears before and after it. The first opinion receives a ranking of 100% and the percentages digress in descending order from there. In addition to “proximity,” TheLaw.net’s algorithm also considers density, diversity and numerosity of search terms. Hence, this is not an algorithm you want to override. It is an algorithm you want to leverage. And, unlike WestlawNext which rents some databases for up to $4,000 an hour, TheLaw.net is still less than $50 a month for unlimited access to all 315 jurisdictions and much, much more.

SHOULD I EVER USE THE WITHIN CONNECTOR?

Yes. Use the Within Connector when you have a specific reason to override the default algorithm, thereby imposing your will on the search engine.

For example:

          search w/2 seizure

In this example we know two things: (a) we know that search engines see the word ‘and’ as a connector, not as a word, per se. We also know that the term ‘search’ appears within two words of the term ‘seizure’. In other words, this is not an arbitrary query.

Here’s another example:

          18 w/5 1344 and “false loan application”

In this search we know that the title number of the Federal bank fraud statute appears within five words of the section number. This query asks the search engine to return that subset of opinions that also includes at least one instance of the phrase “false loan application” anywhere in the opinion. By default, TheLaw.net ranks highest those opinions where ‘false loan application’ appears in close proximity to the item of information driving our search; in this example, that item of information is the Federal bank fraud statute.

How To Override Google’s Default Algorithm and Find Just What You Want In A Click Or Two

Nearly 75% of all legal research begins with Google. I recently began researching a piece on military enlistment contracts. Using Google I entered:

enlistment contract marines

The fifth hit in my result set is embedded below. It’s a form at the website controlled by the Defense Technical Information Center. Look at the file structure of the URL. See the folder captioned eforms? I want to pop that folder open, generate a master list, then search my result set by keyword.

emlistment-contract-marines

Here’s my query:

site:dtic.mil and inurl:eforms

My query restricts my search to a specific folder within a specific domain. I receive 946 hits on separate forms contained in the folder that interests me.

site-url-command

The foregoing example provides a mass customized approach to searching any website. Using this technique, every web site looks the same. Even if I could replicate this result using the framework provided by the DTIC — which I can’t — it would too take long and involve too many clicks.

Now that I have my result set, I can refine my search. I can enter, for example:

site:dtic.mil and inurl:eforms and intitle:enlistment

The result of this search will reveal any forms containing enlistment in the title field of the page. Title searches are great, because presumably the page is all about the term that most interests you.

I receive one hit on a highly relevant form. To cover the landscape, I broaden my search:

site:dtic.mil and inurl:eforms and enlistment

This query returns any forms containing at least one reference to the term enlistment anywhere in the form. The upshot is that from my original list of 946 forms, 65 contain at least one reference to the term enlistment.

This is the list of forms I’m going to mine.