Tagging and Subject Headings

June 6, 2006 at 12:33 pm 2 comments

Jane suggests that LC Subject Headings are Dead
Are we afraid that a simple tagging structure might be more effective and make more sense to our users than headings that look like this “United States — History — Civil War, 1861-1865 — Participation, African American”? Are we just afraid to learn something new? What are we afraid of?

I will be the first to allow that certain LoC headings are outdated and LoC needs a more efficient and effective means of updating the headings. I will even say that some of the ways that the headings are formulated (reverse listing such as Participation, African American") are outdated and could be much more user friendly now that we have the ability to link the terms without needing to flip terms around in order to collocate them.

However, as I stated on her blog:

I’ve participated in this conversation many times within the taxonomy community! I don’t know that “afraid” is the operative term )

I assume you are discussing allowing user tags in a specific resource (perhaps a library OPAC or intranet). I find that internal and focused collections really need the structured terminology more than the enormous collection we call the World Wide Web. When we are talking about Subject Headings, we are really talking about controlled vocabularies (CVs). Taxonomies and thesauri are often used to inform CVs, hence the cross over into the taxonomy forums.

The short answer is that loose tagging is a great way to provide *additional* access points and gather potential updates to an internal controlled vocabulary. However tagging alone can muddy the results in a specialized and/or relatively small resource.
When you are searching the web, you are statistically more likely to surface relevant results using algorithms that include free tagging because of the sheer number of possibilities. When you have a finite group of resources being searched, such approaches tend to pull too many results to be useful to the user. Free tagging can become as problematic as only providing free text searches in a smaller resource. There are a number of organizations that benefit from free text, usually users that do not require complete resource listings nor the most precise result. These users have a lower satisfaction level and may benefit more from user contributed tagging than they do from LoC headings (and other controlled vocabularies)

But for those who require more differentiation and specified results benefit more from having a controlled vocabulary. Not only do CVs enabled a more focused subject term search, the user can be better assured that it is a complete list of relevant results. Our search engine index does cross reference internal acronyms and jargon with the preferred terms. However, if a user runs a subject term search (contains x) they are more assured of retrieving useful items. For example, our free text search will return clutter such as newsletters that draw their information from the very primary resources are staff are seeking! These items are NOT cataloged the same way as the primary, official source of subject information. This differentiation prevents the free text clutter. I have found that some users are inclined to over catalog — which renders the search useless as the results include items that are only tangentially related.

For an example, I will draw from the finance glossary and taxonomy , which was listed in the taxonomy  warehouse.

We will have our users tag 2 items, one a compliation of banking laws (which includes the Community Reinvestment Act, the other a description of the Community Reinvestment Act Statement.

In my experience, users will apply tags  law  and Community Reinvestment Act"  to both of these items. Thus when a user runs a search for "Community Reinvestment Act" they are going to also retrieve the compilation of banking laws — which is unlikely to be helpful to this particular user. This is why I advise against only relying upon free text search or free tagging for information retrieval in these circumstances.

I have heard of an interesting approach that includes both controlled vocabulary and tagging. If I understood the posters correctly, if a user searches the "subject" field, their results will be unaffected by the tags. In addition, the tags are used to further inform the taxonomy (or CV) — helping to keep it up-to-date as new terms roll out.


Entry filed under: Information Retrieval, User Interface.

Leadership styles and Communities of Practice Creating a culture that embraces change

2 Comments Add your own

  • 1. Jane  |  June 6, 2006 at 1:53 pm

    I was not advocating for a complete destruction of LC headings. I agree that they *can* be useful. However, it would be nice to have the ability to search tags as well.

    I want to have my cake and eat it too.

  • 2. librarygeek  |  June 6, 2006 at 3:11 pm

    Jane ~
    I think you can have your cake and eat it too 😀
    If a system will permit user tagging — to be searched with the free text search as well as subject delimeted search — I think we get the both of best world. Of course — the trick is to convince the vendors to make these systems!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

Trackback this post  |  Subscribe to the comments via RSS Feed

Currently Reading or Using

%d bloggers like this: