Is open access (OA) publishing a vanity press industry?

At one end of the OA publishing spectrum are highly-selective journals like PLoS Biology; at the other, journals that may accept any manuscript — even a piece of computer-generated nonsense — as long as the author was willing to pay.  In the middle of these extremes is some vague territory that has not been explored outside of anecdotes and speculation.

Last week on his blog “The Occasional Pamphlet,” Stuart Shieber attempted to answer this question empirically.  In his post, Shieber plots an OA journal’s Article Influence Score (a measure of article quality) by the journal’s article processing fee.  The direction of the relationship is his chief interest.

Shieber argues that a negative relationship between quality and price is indicative of a vanity press market.  His argument is derived from the trade book publishing market, where vanity presses publish low-quality books at high prices to authors, while at the other end, very selective publishers generate high-quality books at zero cost (or often negative cost) to the authors.  Shieber summarizes this inverse relationship between quality and price:

In summary, the higher the publisher’s quality standards, the lower the fee the author must pay for services.  Conversely, if an author is willing to pay enough, some publisher will be willing to take on the publication.  This is the genesis of vanity publishing, and its hallmark is the inverse correlation between publishers’ quality standards and the fees they charge.

Shieber then applies this logic to the OA journal market.  Not surprisingly, he finds a strong, positive relationship between quality and price and concludes that OA publishing — while not immune to vanity press publishers — is no more prone to this kind of business model than subscription-access publishing.

Does Shieber’s argument make sense?  Yes, but his starting assumptions are incorrect, leading Shieber to misinterpret his data and derive the wrong conclusions.

First, Shieber confuses fixed costs with marginal and total costs.  The fixed costs to produce a book include the editorial work, design and layout of a book.  These are often referred to “first-copy costs,” and they do not vary whether you print ten, one hundred, or a million copies.  Marginal costs are the additional expenses required to print and distribute one more book.  Total costs, are simply fixed costs plus total costs, as I’ve illustrated below.

In the print book market, each copy that is sold includes its marginal costs along with some fraction of fixed (or first-copy) costs.  When hundreds or thousands of copies of a book are sold, total revenue can cover the entire first copy costs involved in producing the book.  When even more are sold, the publisher can make a profit and even share some of these profits with the author in the form of royalties.

For a hypothetical vanity press that sells no books — and thus incurs no marginal costs — first copy costs are borne entirely by the author for the privilege of publishing.  Provide higher-quality editorial services, professional layout and design, and these costs must be transferred entirely to the author.  Hence, for a publishing model that incurs only first-copy costs and no marginal costs, one would expect that quality is directly related to price, and thus one would expect a positive correlation between the two.  Still with me?

Now an open-access publisher is like that hypothetical vanity press described above because it only deals with first-copy costs.  While there are certainly overhead costs incurred with setting up computers and purchasing bandwidth, the cost incurred to the publisher for sending one additional copy of an article to a potential reader is essentially zero.  And like that hypothetical vanity press described above, we should expect a positive correlation between author processing charges and article quality, which is exactly what Shieber reports in his study.

Second, Shieber is only plotting those journals which have been given an Article Influence Score by ISI — that’s just 103 journals out of the 4,370 OA journals listed in the DOAJ, or about 2%.  While he computes the correlation between price and quality for journals that do not levy an article processing fee, what’s missing from his analysis are those 98% of OA journals which levy a processing fee but have no Article Influence Score.  Had these journals been included in the analysis (either with imputed influence scores, or by given arbitrarily low scores), the graph may have looked very different.

Third, there are some other issues with the analysis which should be raised:

1. Prices charged to authors may not be the true costs incurred to the journal for publishing an article.  PLoS ONE authors are charged more to publish their article with revenues diverted to subsidize authors who publish in PLoS’s flagship journals, Biology and Medicine.
   
2. Journals have many sources of revenue and these sources of revenue may keep article processing fees lower than the true costs.
3. Publishers may keep article processing fees below cost because there is no author market at the true cost.  Remember how BMC introduced article processing costs at $525 just a few years ago, fees that are now more than three times that figure, with some titles charging substantially more?

In sum, prices charged to consumers may not reflect true costs, and true costs should not be equated with prices.

In conclusion, Shieber’s positive correlation between quality and price cannot be viewed as evidence against an OA vanity press industry.  If anything, they support the claim.

Moreover, by plotting only the 2% of open-access journals, we may only glimpse what is taking place in the most successful strata of open-access journals.  This is like reporting on the book publishing industry by focusing solely on the New York Times Bestseller list.

Open access publishers have flooded the market with thousands of new author-pays titles in recent years, advertising fast publishing at very competitive prices, some now as low as $99.95.  Shieber’s analysis entirely misses this emerging publishing market geared to service the needs of authors first and readers second.  This may be the dominant paradigm in author-pays, open access publishing, with publishers like PLoS bucking the trend.  We would not know this, unfortunately, from this study.