How are we doing on BISAC YA implementation?

Recently, the question, "Are the new Young Adult BISAC subject codes being used?" came up 'round the BNC campfire. 

Luckily, BiblioShare offers the possibility of deep diving into metadata implementation so the answer is easily found and, yes, indeed they are!

As of Oct. 5, 17,258 ISBNs (limited to 2016 publication dates) in BiblioShare were carrying one of the new YA BISAC codes as a Main Subject. But how many records should be carrying them but aren't? How are we doing, really?

But first, a brief preface: As I'm sure everyone knows, "good" metadata within North American best practices should meet these general qualifications:

  • Every record should have a single, retailer-appropriate main subject assigned to it.
    • BISAC subject codes remain dominant in Canada and Canadian retailers continue to rely on them.
    • BookNet Canada recommends that Canadian publishers also supply Thema subject codes and qualifiers, but that's outside the scope of this quick study.
  • Every trade book record should include a single, ONIX code list 28, Audience Code value between "01" and "03."
    • Trade YA books should use "03."
    • Trade Juvenile books should use "02."
    • Trade Adult books should use "01."
  • Every children's book record (YA and JUV) should carry Audience Range information.
    • Typically, this means an Interest Age in years supported by a Grade Range where appropriate.
    • Range values should be as specific as possible—a spread of more than three years is of little value.
    • There is no clear prohibition against providing Audience Range information on adult books, but if it's provided it should mean something. For a typical trade adult book, range levels are neither needed nor helpful. However, a book with content for adults where the reading level is intentionally low is an example of a book that would benefit from Audience Range information. 

The aim of the above qualifications is to provide retailers (and librarians and other stakeholders) a reliable means of selecting books that meet broad criteria.

Unfortunately, it's a sad truth when deep diving in BiblioShare metadata that one of the hardest criteria to match is for children's books. Audience Codes are notoriously inaccurate, and while Main Subject is generally a better choice, subject classifications are not exactly pristine data either.  

So for this study, I decided that looking at Interest Age would be best. Experienced publishers of all sizes provide it — and that's what we really want to know: How many of those publishers who you'd expect to have implemented the new YA codes actually have done it? It should be possible to isolate books appropriate to young adults using this range, and then determine how many of those books carry the new YA codes.

My initial data select (2016 only) of Interest Age with ages greater than 12 netted 35,000 records. After winnowing out some chaff (yes, I'm talking about you, 0-99) and educational books with non-trade codes, I was left with 30,111 as my Interest Age–appropriate universe for testing. Of those:

  • 10,662 or 35% carried a YA Main Subject
  • 10,152 or 34% carried a JUV Main Subject
  • 9,297 or 31% carried an Adult Main Subject

Let's add a level of security with a second criteria: 19,302 or 64% of the above carried an Audience Code of "03" for Young Adult. Looking at that group:

  • 9,228 or 48% carried a YA Main Subject
  • 6,708 or 35% carried a JUV Main Subject
  • 3,366 or 17% carried an Adult Main Subject

Only 2/3 of the books carrying YA codes met the Audience Range criteria I created (10,662 out of 17,258 or 62%).

So there you have it. This is something like spitting into the wind from the top of a hill o' beans as far as research goes, but I think I can say that somewhere between 1/3 and 1/2 of the possible universe of YA books have had the new BISAC codes applied.

Maybe this post will be a reminder that helps spur on YA code adoption, but let's try to improve our use of Audience Code while we're at it. As it stands now, librarians don't believe publisher data can be accurate once they try to either select children's books from publisher data or compare a known list of children's books against publisher data. We have a code value for this simple select. Let's make it mean something!