Tuesday, January 15, 2013

Analyzing changes to journal prices over time

I've finally gotten around to carefully reading the recent article put out by the AAP that suggests that prices of journals have not increased that terribly much over the last twenty years.  The first post I read that referenced this was the Scholarly Kitchen.  Kent Anderson supports the idea that we have, indeed, been looking at the wrong measure (price of print subscriptions when the vast majority of subscriptions are now online or print+online), which is a fair argument.  However, it is clear that, due to proprietary interests, publishers are not willing to share the deals that it makes with libraries, so it's really been the only data that is available to look at.

The author of the original article, Paula Gantz, starts with the fact that Library Journal's Annual Periodical Price Survey uses "list prices of print subscriptions to calculate the real increase in serials expenditures", which she notes is a "misleading and inaccurate method for tracking how libraries are spending their serials budgets and fails to recognize the increased value they are receiving from the print-to-digital transition."  Gantz presents a number of issues that I thought I'd address, from a library collections development perspective.

Measures of journal expenses
The primary issue Gantz addresses is the use of list prices of print subscriptions over time, which is not an accurate method of tracking costs to libraries.  I agree that list prices of print subscriptions are no longer relevant to libraries, particularly academic libraries, which are primarily licensing online journal content.  The difference between purchasing and licensing content is not addressed in this article, something which I will address later.  Focusing on the measure, it may be accurate (it is what the publishers provide), but it is indeed not a valid measure.  It is, however, the only one we've got.  Publishers are not willing to release what libraries actually pay for their subscriptions, and data on prices for online subscriptions is confounded by the "Big Deals" and "Small Deals" or subject-specific packages.

It is surprising that Gantz did not mention that for the last two years, the LJ survey included prices for online subscriptions (print+, online only and first-tier) for the titles indexed by ISI products.  Comparing the prices for the subject areas provided in both 2011 and 2012 reports, there was only a 1% increase in the price.  Of course, this is a very limited set of titles, and only 20 of 33 fields have data for both years.

Field 2012 2011 $ Diff % Diff
Chemistry $3890 $3676 $214 6%
Physics 3185 3492 -307 -10%
Biology 2070 1913 157 8%
Engineering 2009 2691 -682 -34%
Zoology 1842 1140 702 38%
Botany 1760 1620 140 8%
Health Sciences 1593 1406 187 12%
Geology 1485 1518 -33 -2%
Technology 1341 1306 35 3%
Math & Computer Science 1328 1133 195 15%
General Science 1261 892 369 29%
Agriculture 1075 1016 59 5%
Geography 973 1038 -65 -7%
Business & Economics 733 754 -21 -3%
Sociology 713 638 75 11%
Education 708 830 -122 -17%
Arts & Architecture 481 536 -55 -11%
Philosophy & Religion 405 490 -85 -21%
History 383 363 20 5%
Language & Literature 351 368 -17 -5%

Rather, Paula Gantz looked at what libraries actually spent on serials.  It is interesting that she used the Association of Research Libraries' (ARL) data, which is a very small subset of academic institutions, usually much larger than average.  But are expenditures a valid measure of journal prices?  If every title was purchased separately, then it is.  However, the effect of package deals distorts this relationship.  True, the per-title price of packages is less than title-by-title, the titles in the package do not reflect the true collection development decisions of the librarians.

Difficulties of Calculating Cost per Journal
Using library expenditures as a measure of cost per journal is confounded by the title packages, "Big" or little deals.  The value is based on the entire package, and usually is not distributed evenly nor the same for every library.  Libraries may provide access to more titles, but there are at least some titles that would the library would not pay for, if given the opportunity.  Gantz does mention how libraries have started rejecting the package deals are many are returning to title-by-title subscriptions.

Increased Access and Usage
This may be a point that publishers may believe is not well appreciated by librarians when it comes to pricing of electronic journals.  The tremendous increase in access has had a major impact on our patrons, as indicated by the equally impressive growth in usage.  I'd be interested to learn what the average usage per print title was just prior to ejournals overtaking print.  But I imagine that the usage was much more limited, given limitations in access and costs of copying.  Indeed, increased spending on ejournals have been associated with better research outcomes.  However, the increased proportion of the budget spent on journals, particularly by academic libraries, is evidence of the increased value librarians do see in providing extended access to them.

Similarly, with many (not all, though) online subscriptions, access is made available to an extensive archive, some all the way to the first issue.  The flip side of this is that this instant access can instantly disappear upon cancellation of the subscription.  Furthermore, unlike print, libraries cannot fill gaps in coverage through donations or one-time purchases of limited ranges.  This makes the collection much more ephemeral.  Librarians appear to be ambivalent regarding the impact of this on the quality of our collections.

Growth in Content
Paula Gantz notes that not only are there more journals being published out there, but that journals are putting out more articles.  The implication is that libraries are getting more content for their money.  The increase in journals, however, has no bearing on this argument because each journal requires a subscription. However, the increase of content provided by each subscription is important to consider.  Indeed, in STM journals, the annual article rate per journal has increased from 117 in 1988 to 139 in 2005 (if somebody has it, I'd be interested to see more recent rates).  However, content was already increasing in the print-years - from 83 in 1975 to 154 in 2005 in science journals.  These numbers come from studies of different sets of journals, but they are both referenced in this 2009 report from STM.  So it is clear that increased content is not the primary reason for recent (post-print-age) increases in costs.

Gantz also mentioned the growing amount of supplemental data or features that add to the growth in content.  Unfortunately, usage of these features is difficult to measure.  Most journals that do provide this additional content do not provide usage data.

Decreased Cost of Maintenance
In the summary, Paula Gantz provides recommendations for both publisher and librarians.  To the former, she essentially advises to be more open about pricing and work more with libraries.  To the latter, she recommends, essentially, to be prepared to fight for your budget and be careful when consider author-paid OA titles.  She also suggests keeping in mind the cost savings of going digital, in terms of space, staff and processing.  I'll concede savings in space, but that is all.  A good summary of differences in maintaining print versus electronic is here.  While a bit dated, most of the issues are addressed.  The differences in processes have essentially resulted in higher total costs (excluding subscriptions) due to the greater complexity and greater need for technology.  Instead of having low-cost clerks and paraprofessionals handle 90% of the tasks of managing print journals, we require higher-paid electronic resource librarians and librarians specializing in contracts and negotiation, as well as high-tech staff to handle data management to make the resources available and the data associated with evaluating usage.  Instead of 3-5 copiers, we require dozens of computer workstations, with concomitant color printers.

Conclusions? More questions...
So instead of being assured that prices of ejournals have, indeed not risen as steeply as we originally were led to believe, I am even more confused and have many more questions, including:

  • What is the best way to measure the costs of journals?  Per subscription? Per article? Per usage?
  • Where can we get this data?
  • How do library expenditures compare with listed prices?
  • How do we measure costs per title of journal packages?
  • What has been the trends in content growth per subscription?
  • What is the usage of the supplemental data?  How many journals provide it?  How many require a subscription to access it?  
  • Are we buying too much?  Is the growth in content truly worth our money?
  • What is the extent of content that is provided with each subscription?  How far back are the archives with the subscription?  How much are the archives used?
  • What are the true costs of maintaining ejournals?


  1. There's a thread over on FriendFeed that I started when I offered a link to your post. You might find it interesting.


  2. Thanks for the link...I had not noticed this activity (I've had my head buried in accreditation reviews). Walt Crawford has added some interesting data looking at NCES data (that's all libraries, not just ARL). His big takeaway is that while print subscription expenditures increased a mere 4.7% from 2008-2010, expenditures on electronic serials increased 24.3%! He admits this is "quick and dirty", but it's a clue that things may, indeed, be as bad as we've said it is (and worse than the AAP will admit).

    But now the pressure is on...Walt has subscribed to my blog.