Free Text Searching

Free text searching, aka natural language searching, is the most commonly used method of searching databases by general users. Novice users may only have experience with free text searching, and therefore expect that all search experiences work similarly.

Free text searching is the only method available to search Google / internet. While searching Google or Google Scholar using free text can be effective at finding relevant results, there are problems associated with it. For example, Google search biases include Google customizing its search for you as a user based on your search history. It also varies its search results based on your location/region and time of day. Marketing interests, website metadata, and AI skew Google search results in favor of links that have been promoted by their owners and more. Also, in Google certain results are more likely to be closer to the top of your search based on popularity of results, number of comments, and more.

Example Novice search used in Google: Don't do this!

Treatment options for conduct disorder and oppositional defiant disorder beyond therapy or medication like risperidol and psychostimulants

Example Expert search using free text: Do this instead!

("conduct disorder" OR "oppositional defiant disorder" OR "behavioral disorder") AND (treatment OR therapy OR medication OR risperidal OR psychostimulant*) 

Most databases allow for an "All fields" free text search. Depending on the database, this includes the resource title, abstract, author supplied keywords, resource full-text and/or controlled vocabulary. Therefore, this is a broad method of searching. For greater control of your search accuracy, as is required for evidence synthesis, you may specify which subfields you want to search as free text that you'd like to search depending on the database you search.

Example: Expert search using free text search and field tags (in PubMed)

("conduct disorder"[tw] OR "oppositional defiant disorder"[tw] OR "behavioral disorder"[tw]) AND (treatment[ti] OR therapy[ti] OR medicat*[tiab] OR risperidal[tw] OR psychostimulant*[tw]) 

More information on PubMed field tags

To find the results that you need for an in-depth evidence synthesis, like a systematic review, an expert searcher must use a combination of free text and controlled vocabulary searching. This maximizes your search sensitivity and specificity for optimal precision.

Controlled Vocabulary Searching

Controlled vocabulary, Subject Terms and Thesaurus are interchangeable terms. We will use the term Controlled Vocabulary throughout these modules.

Controlled vocabulary are specifically index terms within databases developed by metadata or cataloging librarians / information professionals. Controlled vocabulary are pre-defined terms or phrases developed and indexed within a database to identify resources within that same database that are highly related, or primarly focused on that topic. The resources are also assigned that controlled vocabulary term.

Controlled vocabulary strengthens your database search specificity (remember specificity means getting fewer results that are right on your desired concept target).

Real World Example

When you take a photo of yourself in front of University of Illinois' mammoth and post it on your social media account, you might include #woolymammoth #uiuc when you post it. Why? This allows anyone who sees the picture to know that the statue behind you is of the UIUC woolly mammoth. Also, it makes it more likely that a person who searches for photos of the UIUC woolly mammoth will find your photo. In this example #UIUC and #woolly mammoth are your controlled vocabulary.

New Build: Woolly Mammoth Sculpture. UIUC Facilities and Services. Accessed September 24, 2024 from https://fs.illinois.edu/Projects/natural-history-building-woolly-mammoth/

PubMed Controlled Vocabulary (MeSH) Example

In PubMed, controlled vocabulary is called Medical Subject Headings aka MeSH. Say that I would like to find articles that are very specifically about conduct disorder in PubMed. A free text search yields over 16,000 results. However, a Controlled Vocabulary search yields less than 4,000 results. AND each of the 4,000 results in my controlled vocabulary search will be highly relevant to the concept of Conduct Disorder. See below.

What are Author Supplied Keywords?

Authors submitting research products to journals for publication supply a few keywords that they identify as highly relevant to their research product as descriptors. 

Example

The authors of the following article provide a few keywords to help people recognize what the article is about.

• Article title: Experiences of Palliative and End-of-Life Care among Older LGBTQ Women: A Review of Current Literature
• Author supplied keywords: chronic illness; elders; end-of-life; LGBTQ; older adults; terminal illness

Author supplied keywords operate similarly to controlled vocabulary; however, they are not specifically pre-determined and assigned by the database. Instead, the authors control these terms

Google Search Tips

Sometimes it's useful to search Google depending on your information / resource need. For example, sometimes incorporating a search of Google will yield grey literature resources that you would not identify using databases. NOTE: incorporation of grey literature is very important to reduce publication bias in evidence synthesis (more on this in a later module). So how can you do so effectively given its constraints? Below are some search tools that Google understands:

Google understands

• Boolean Operators, AND, OR
  • NOTE: you can also use + or AND as well as | for OR
• Parentheses
• Quote marks

Search within domains

To search within a specific domain, search = site:domain name AND keyword.

EXAMPLE 1

A Google search, site:.apa.org AND "conduct disorder" = gets you results from American Psychological Association website about conduct disorder

EXAMPLE 2

A Google search, site:illinois.gov AND "public health" AND water = gets you information provided by the state of Illinois government having to do with public health and water, water quality, etc.

Proximity Operators

Proximity operators are search tools that are focused on character string analysis when conducting phrase searching or full text searching in databases. Proximity looks at the relative position of  words for a given two terms. Some databases offer NEAR or variations that specify closeness of the terms, with or without specifying the order of the terms. There are frequently variations of NEAR that specify the order that the words should appear in. 

Proximity operators allow searchers to avoid the restrictiveness of exact phrase searching and still retain the specificity of the search results. Proximity operators additionally allow more flexibility and finetuning than using solely Boolean Operators when building complex searches.

The symbol or operator for proximity searches varies from database to database. Notice as you review the table below that some databases allow you to specify a specific order of words presentation. Other databases only allow you to search certain fields using proximity operators. To find the proximity operator for a given database, check the Help file for the database you are interested in using. Some sample proximity operators include:

• PubMed Field tags
• Keywords and Controlled Vocabulary