Obtaining Copyright Permission to Digitize Published Works Remains a Significant Barrier

A review of:
 
 George, Carole A. “Testing the Barriers to Digital Libraries: A Study Seeking Copyright Permission to Digitize Published Works.” New Library World 106.1214/1215 (2005): 332-42.
 
 Objective – To assess the effectiveness and efficiency of the copyright permission-seeking process and to suggest improvements in order to improve outcomes. 
 
 Design – Workflow study.
 
 Setting – Carnegie Mellon University Libraries, Pittsburgh, PA, USA.
 
 Sample – A random sample of titles published 1999-2001 was selected from the library’s circulating collection. After eliminating duplicates, technical reports, theses, dissertations, and missing items, the sample comprised 337 titles. Of these titles, 70% were books, and 56% were from commercial publishers. From this a working sample of 273 titles was derived, comprising those titles protected by copyright and with the rights owner clearly indicated. About 73% of this working sample appeared to be out-of-print; their median publication year was 1981. 
 
 Method – In this two year study (1999-2001), a random sample of books was selected, and pertinent bibliographic and copyright holder information researched and recorded. Permission letters were sent and, six weeks later, follow-up letters were sent to non-respondents. The letter allowed respondents four options:
 
 Grant full permission to digitize the work and provide unrestricted Web access;
 Grant permission to digitize the work and provide read-only Web access, limited to Carnegie Mellon University users;
 Declare that they do not hold the rights, and hopefully provide information to identify and locate the actual rights holder;
 Deny permission for digitization. 
 
 
 Results were then recorded and analyzed.
 
 Main results – Of the 273 letters mailed, a clear ‘yes’ or ‘no’ reply was obtained for just over half (52%) of the documents. Sixteen percent of the rights holders could not be found (the letter was returned, or a referral proved impossible to locate and contact). Another 25% of the copyright holders simply did not reply, and 7% were otherwise problematic. Of the 143 ‘yes’ or ‘no’ responses, 54% denied permission, while 46% granted permission. (Note: these percentage figures appear to be erroneously reversed in Table 1 of George’s article.) Therefore, of the overall working sample of 273 titles, permission to digitize was obtained for only 24% of the titles. 
 
 A substantial portion of the permissions (41 of 66, or 62%) carried some restriction. This represents 15% of the total working sample. Only a few restriction requests were deemed too great to make use of the permission. 
 
 Commercial publishers who made up 58% of the working sample granted permission at the lowest rate (13%). 
 
 Response time averaged three months from the time the initial letter was sent until a ‘yes’ or ‘no’ response was received. Negative responses averaged a bit longer than positive responses (101 days to 124 days). However, some of this time was attributable to delays in issuing follow-up or redirected request letters (a step required in 60% of cases), owing to the limited staff resources at Carnegie Mellon. 
 
 The copyright ownership had changed in 23% of the sample, requiring more than one and up to three different addresses to be contacted before a response was received or the effort was terminated. 
 
 Conclusions – The study concluded that the permission rate would remain low unless additional efforts were made in the permission-seeking process (e.g., personal contacts in addition to letters and emails), or unless more selective approaches were employed (e.g., targeting non-commercial publishers). It also concluded that the process to seek copyright permissions was neither quick nor easy, suggesting the need for dedicated staff time and a readily accessible database of publisher contact information. As a result, subsequent projects have improved their permission-seeking process, focusing on more non-commercial publishers or older publication dates, and asking publishers for blanket consent for all of their out-of-print titles. 

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.