Re: Institutional Journal Costs in an Open Access Environment

["William Walters" <William.Walters@millersville.edu>]

Good afternoon.

I have discovered and corrected a methodological error in my study that
has some bearing on the results.  I've posted the corrected paper -- the
version that will appear in JASIST -- at
<http://www.library.millersville.edu/public_html/walters/journal_costs.pdf>

The correction affects the total system-wide revenue estimates as 
well as the results for the Equal-Revenue Model -- the model in 
which total Open Access revenue (from all institutions combined) 
is held equal to total subscription revenue under the current 
system.

My original paper ensured that the Equal-Revenue assumption held 
true *for my sample* -- a stratified random sample that includes 
two or three institutions from each U.S. News & World Report 
category (Doctoral Universities, Master's Universities, etc.).

The problem with the old version of the paper is that the the 
sample does not reflect the true distribution of colleges and 
universities among the various institutional categories.  The 
sample includes three doctoral institutions and two master's 
institutions, for example, although there are actually more than 
twice as many Master's Universities as Doctoral Universities. 
The aggregate revenue estimates must therefore be adjusted to 
account for the actual number of institutions within each 
category.  (The new, adjusted, revenue estimates are presented in 
Table 4 and described on pages 11 and 12.)

In the old version of the paper I relied strictly on the sample 
data, underestimating the relative number of small colleges -- 
institutions with low publishing productivity.  I therefore 
overestimated the total (systemwide) PLoS-Model revenue and 
consequently underestimated the amount that would need to be paid 
by the research universities in order to "make up the 
difference."

My general conclusions are not affected by this correction, and 
the abstract for the paper remains the same as before.  There are 
substantial changes in the quantitative results, however. 
Specifically, I now estimate that within the four disciplines 
included in the study,

-- the PLoS Model brings in only 15% as much revenue as the 
Conventional Model -- not 28%, as reported earlier

-- a switch from the current pricing model to the Equal-Revenue 
Model would increase the University of Michigan's costs by 237% 
-- not 83%, as reported earlier

-- in general, institutions larger than Brandeis (about 1.1 
million volumes) can expect to pay more under the Equal-Revenue 
Model -- not institutions with more than 3.9 million volumes, as 
reported earlier.

As several people have mentioned, these results are valid only if 
we assume that all Open Access journals charge publication fees, 
and that academic institutions are responsible for paying those 
fees (with tuition money, grant money, government funding, or 
some other source of income).

These corrections do not alter the main conclusion of the paper 
-- that the wholesale implementation of either Open Access model 
would reduce costs for the vast majority of institutions while 
dramatically increasing the proportion of the total system-wide 
cost paid by the major research universities.

Bill

William H. Walters, PhD
Assistant Professor of Librarianship
Collection Development Librarian
Helen A. Ganser Library
Millersville University
Millersville, PA 17551-0302
(717) 871-2063