Over the next several months the WestlawNext Canada (WLNC) platform will be rolled out to existing subscribers, and made available to new subscribers. Law students on academic accounts will not have access until September 2014, but faculty will have access before then. This is the first of several posts that will explore selected features of WLNC.
Plain language and Boolean searching
The “plain language” search engine is the default search method in WLNC, when using the main search box at the top of the page. However, it is easy to run a Boolean search in that box if you prefer.
- If you type a query containing a proximity connector or a truncation command the program assumes you want a Boolean search and runs that type of search instead.
- If your query doesn’t contain those items, but contains commands such as AND, OR or quotation marks, then you are asked if you want to run a Boolean search.
- You can also change your preferences so that a query containing any of those commands will automatically be run as a Boolean search.
The Boolean commands have not changed from the old Westlaw Canada platform. You can tell whether your search is being processed as plain language or Boolean, because Boolean searches will have adv: inserted in front of the query.
The plain language search engine looks for the best match for your search terms, using frequency and proximity algorithms. It automatically looks for variant word-forms of the terms in your query. It also uses phrase recognition, and searches equivalencies for commonly used acronyms.
The plain language search engine works well, but it will work even better if the following adjustments are made to your query.
First, although phrase recognition is incorporated, you can force phrases to be recognized by putting terms in quotation marks.
Second, the plain language search engine may retrieve documents that do not contain all of your search terms, because it looks for the best match.
- To force a term to be included put + immediately before the term. To force a term to be excluded put – immediately before the term.
Third, the plain language search engine does not incorporate concept-based searching, in which synonyms are automatically searched. To capture synonyms use the following method, which also works in the Natural Language search on the old platform.
- Put synonyms in parentheses immediately after the search term, and separate the synonyms by commas. For example, to find synonyms for timber, your query might contain timber (lumber, wood, log)
A search from the main search box runs a federated search through all WLNC content except for the finding tools (eg., Abridgment Digests, Words and Phrases, ICLL). Alternatively, you can search particular content, including one of the finding tools, by clicking on that content link.
Search results for federated searches initially appear in an Overview format, showing the top documents in each content type. The results for each content type can be expanded from a menu on the left.
A drawback of the federated search is that your search results will contain documents that are outside of subscription. They will be marked as outside subscription in your search results.
- One way to avoid seeing these documents in your results is to restrict your search to a particular product or content type that you subscribe to, such as LawSource.
- If your search is restricted in this way, the search box will show that restriction.
Find by citation or name
Citations are searched on WLNC by typing them into the federated search box. If the citation is valid, all is good. You will easily retrieve the document without the search being assigned any “suggested charge-back” value. However, if the citation is invalid then a federated search will automatically be run.
- To avoid this, you can insert fi: or find: before the citation.
- Multiple documents can be retrieved by separating the citations with a semi-colon.
- If a KeyCite result is desired for a citation, then type kc: before the citation(s).
To find or note-up a document by name, it is necessary to use a separate tab, labelled “Find and KeyCite by Name”.
- Unlike CanLII, the federated search doesn’t seem to rank first those documents that contain the search terms in the name. Also, using the federated search will result in the search being assigned a “suggested charge-back” value.
- It is therefore better to use this separate tab to find or KeyCite documents by name. This method searches only in the name field.
- Unlike Quicklaw, there is no look-up list generated to help the user get to the correct document.
The Quicklaw interface is superior for finding by name and by citation, because of the look-up lists that are generated as soon as information is typed into the Find a Document box on the Start Page. CanLII does not include look-up lists for finding by name or citation. However, it does use this feature for its Noteup search box. Also, CanLII’s federated search engine is more effective at searching by name, because it puts those documents whose name contains the search terms at the top of the result list.
WLNC does generate look-up lists below the federated search box for some items:
- the name of a jurisdiction, such as British Columbia
- the name of a publication, such as a law journal
- the name of a database or topical collection
- the name of a statute (not all statutes are included in the look-up lists).
The subtle ADVANCED link to the right of the search box is the entry point to search templates for Boolean searches.
The Advanced Search option is best used for a content category, and provides a template designed for searching that particular content. Select a category from the Home Page, such as Cases and Decisions, and then click on ADVANCED.
- You will see a template similar to the current search templates in Westlaw Canada. ADVANCED queries are based on Boolean searching. Results will be relevancy ranked.
- If you type a query that contains search commands into the template, that will override the default connector in the template. The actual query that will be run appears in the search box at the top of the page, as you type into the template.
I have long held the view that researchers who have not learned Boolean search syntax will achieve better results on Westlaw using plain language searching.
- The main advantage to the templates is if you want to use fields to narrow your search, such as specifying a particular judge or counsel. However, many of the fields in the templates are not necessary for narrowing your search in WLNC, because you are able to filter your search results by the same parameters.
- WLNC does allow use of multiple search boxes from the template to create a query, which provides more flexibility than the old platform. In addition, you can boost words in your query by requiring them to appear at least a certain number of times in retrieved documents. However, there are no proximity options built into the template.
- You can overcome this by adding proximity commands to the query you type into the template. Pay attention to how the query appears at the top of the screen, to ensure that the actual query is what you intended.
All results in WLNC are initially ranked by relevance. This is a major improvement over the default ranking method in the old platform.
The relevancy ranking takes several factors into consideration, including citation relationship data. Some of the factors that influence ranking in the US product, such as digest classifications and user print/downloads, are not employed in WLNC.
Results for case law can be re-sorted by date, citation frequency, and court level. This type of functionality has been available for some time on Quicklaw and CanLII.
Results can also be filtered in many ways, depending on the content type.
Initial impressions of searching on WLNC
WLNC puts a “Google-like” search experience at the front and centre, but ensures that Boolean searches can be easily run. I think this is the right choice for a product designed for lawyers. However, the search engine must be very good for experienced legal researchers to rely on behind-the-scenes algorithms rather than crafting Boolean queries. My experience so far with the WLNC plain language search engine is that it works very well.
I suggest that users try both Boolean and plain language searches and compare the results. I would be interested to hear your views. One drawback to plain language searching as presented in WLNC is that there is no obvious way to search synonyms. However, this can be resolved using the method described above. It would be helpful if WLNC referred to this method in its Search Tips page.
The ability in WLNC to re-sort search results, and the easy access to filtering options, greatly improve on the old platform.
The tools for finding by name and citation in WLNC do not live up to the rest of the product. This basic function needs some attention.
Future posts will explore other features in WLNC, such as viewing and working with search results, using folders, and the new “suggested charge-back” structure.