Tag Archives: Request a Copy

Who is requesting what through Cambridge’s Request a Copy service?

In October last year we reported on the first four months of our Request a Copy service. Now, 15 months in, we have had over 3000 requests and this provides us with a rich source of information to mine about the users of our repository.  The dataset underpinning the findings described here is available in the repository.

What are people requesting?

We have had 3240 requests through the system since its inception in June 2016. Of those the vast majority have been for articles 1878 (58%) and theses 1276 (39%). The remaining requests are for book chapters, conference objects, datasets, images and manuscripts. It should be noted that most datasets are available open access which means there is little need for them to be requested.

Of the 23 requests for book chapters, it is perhaps not surprising that the greatest number  – 9 (39%) came for chapters held in the collections from the School of Humanities and Social Sciences. It is however possibly interesting that the second highest number – 7 (30%) came for chapters held in the School of Technology.

The School of Technology is home to the Department of Engineering which is the University’s largest department. To that end it is perhaps not surprising that the greatest number of articles requested were from Engineering with 311 of the 1878 requests (17%) from here. The areas with next most requested number of articles were, in order, the Department for Public Health and Primary Care, the Department of Psychiatry, the Faculty of Law and the Judge Business School.

What’s hot?

Over this period we have seen a proportional increase in the number of requests for theses compared to articles. When the service started the requests for articles were 71% versus 29% for theses. However more recently, theses have overtaken request for articles to a ratio of 54% to 46%.

The most requested thesis, by a considerable amount, over this period was for Professor Stephen Hawking’s thesis with double the number of requests of the following ten most requested theses. The remaining top 10 requested theses are heavily engineering focused, with a nod to history and social research. These theses were:

The top 10 requested articles have a distinctly health and behavioural focus, with the exception of one legal paper authored by Cambridge University’s Pro Vice Chancellor for Education, Professor Graham Virgo.

When are people requesting?

Looking at the day of the week people are requesting items, there is a distinct preference for early in the week. This reflects the observations we have made about the use of our helpdesk and deposits to our service – both of which are heaviest on Tuesdays.

When in the publication cycle are the requests happening?

In our October 2016 blog we noted that of the articles requested in the four months from when the service started in June 2016 to the end of September 2016, 45% were yet to be published, and 55% were published but not yet available to those without a subscription to the journal.  The method we used for working this out involved identifying those articles which had been requested and determining if the publication date was after the request.

Now, 15 months after the service began it is slightly more difficult to establish this number. We can identify items that were deposited on acceptance because we place these items on a very long embargo (until 2100) until we can establish the publication date and set the embargo period. So in theory we could compare the number of articles with this embargo period against those that have a different date.

However articles that would provide a false positive (that appear to have been requested before publication) would be ones which had been published but we had not yet identified this – to give an indication of how big an issue this is for us, as of the end of last week there were 1768 articles in our ‘to be checked’ pile. We would also have articles that would provide a false negative (that appear to have been requested after publication) because they had been published between the request and the time of the report and the embargo had been changed as a result. That said, after some analysis of the requests for articles and conference proceedings, 19% are before publication. This is a slightly fuzzy number but does give an indication. 

How many requests are fulfilled?

The vast majority of the decisions recorded (35% of the total requests for articles, but 92% of the instances where we had a decision) indicate that the requestor shared their article with the requestor. The small number (3%) of  ‘no’ recordings we have indicate the request was actively rejected.

We do not have a decision recorded from the author in 62% of the requests. We suspect that in the majority of these the request simply expires from the author not doing anything. In some cases the author may have been in direct correspondence with the requestor. We note that the email that is sent to authors does look like spam. In our review of this service we need to address this issue.

Next steps

As we explained in October, the process for managing the requests is still manual. As the volume of requests is increasing the time taken is becoming problematic. We estimate it is the equivalent of 1 person day per week. We are scoping the technical requirements for automating these processes. A new requirement at Cambridge for the deposit of digital theses means there will be three different processes because requests for these theses will be sent to the author for their decision. These authors will, in most cases, no longer be affiliated with Cambridge. Requests for digitised theses where we do not have the author’s permission are processed within the Library and requests for articles are sent to the Cambridge authors.

Given the challenges with identifying when in the publication process the request has been made, we need to look at automating the system in a manner that allows us to clearly extract this information. The percentage of requests that occur before publication is a telling number because it indicates the value or otherwise of having a policy of collecting articles at the acceptance point rather than at publication.

Published 12 September 2017
Written by Dr Danny Kingsley
Creative Commons License

Request a copy: process and implementation

This blog post looks at a recent feature implemented in our repository called ‘Request a copy’ and discusses the process and management of the service. There is a related blog post which discusses the uptake and reaction to the facility.

As part of our recent upgrade to the University’s institutional repository (now renamed ‘Apollo‘), we implemented a new feature called ‘Request a copy’. ‘Request a copy’ operates on the principle of peer-to-peer sharing – if an item in Apollo is not yet available to the public, a repository user can ask the author for a copy of the item. Authors sharing copies of their work on an individual basis falls outside the publisher’s copyright restrictions; here, the repository is acting as a facilitator to a process which happens anyway – peer to peer sharing.

The main advantage of the ‘Request a copy’ feature is to open up the University’s most current research to a wider audience. Many of our users do not necessarily come from an academic background, or may be based within another discipline, or an institution where journal subscriptions are more limited. The repository is often their first port of call to find new research as it ranks highly in Google search results. We hope that these users will benefit from ‘Request a copy’ by being able to access new outputs early, at researchers’ discretion. Additionally, this may provide an added benefit to researchers by introducing new contacts and potential collaborations.

How it works

Screen Shot 2016-10-06 at 13.53.30Items in Apollo that are not yet accessible to the wider public are indicated by a padlock symbol that appears on the thumbnail image and filename link which users can usually click to download the file.

Reasons why the file may not yet be publicly available include:

  • Some publishers require that articles in repositories cannot be made available until they are published, or until a specified time after publication
  • We hold a number of digitised theses in the repository, and for some we have been unable to contact the author to secure permission to make their thesis available
  • Authors may choose to make their dataset available only once the related article is published

When a user clicks on a thumbnail or filename link containing a padlock, they are directed to the ‘Request a copy’ form. Here, they provide their name, email address and a message to the author. On clicking ‘Request copy’, an email is sent to the person who submitted the article, containing the user’s details. The recipient of this email then has the option to approve or deny the user’s request, to contact the user for more information, or (if they are not the author) to forward the request to the author.

How it really works

In practice, the process is slightly more complicated. For most of the content in the repository, the person who submitted an item will be a member of repository staff, rather than the item’s author. This means that for the most part, emails generated by the ‘Request a copy’ form were initially sent to members of the Office of Scholarly Communication team. In some cases, these requests were sent to people who have left the University, and we have had to query the system to retrieve these emails. As an interim measure, we have now directed all emails to support@repository.cam.ac.uk. These still need manual processing.

Theses

For theses where we have not received permission from the author to make them available, we forward requests to the University Library’s Digital Content Unit, who have traditionally provided digitised copies of theses at a charge of £65. We have  found however, that once information about this charge is communicated to the requester, very few (approximately 1%) actually complete the process of ordering a thesis copy.

We have been working with the Digital Content Unit on a trial where thesis copies were offered at £30, then £15. However, even at these cheaper prices, uptake remained low (it increased to 10%, but due to the small size of the sample, this only equated to two and three requests at each price point, and therefore may not be statistically significant). This indicates that the objection was to being charged at all, rather than to the particular amount. Work in this area remains ongoing to try and offer thesis copies as cheaply as possible to requesters, while allowing the Digital Content Unit to cover their costs.

Articles

If the request is for an article, we first need to check whether the article has actually been published and is already available Open Access. Although we endeavour to keep all our repository records up to date, unless we are informed that an article has been published, repository staff need to check each article for which publication is pending. This is a time-consuming manual process, and when we have a large backlog, sometimes it can take a while before an article is updated following publication.

If we found that the article has indeed been published and can be made Open Access, we amend the record, make the article available and email the requester to let them know they can now download the file directly from the repository.

On the other hand, if the article is still not published, or if it is under an embargo, we need to forward the request to the corresponding author(s). Sometimes their name(s) and email address(es) will be included within the article itself, and sometimes we have a record of who submitted the article via the Open Access upload form. However, if it is not clear from the article who the corresponding author is, or if their contact details are not included, and if the article was submitted by an administrator rather than one of the authors, we then need to search via the University’s Lookup service for the email addresses of any Cambridge authors, and search the internet for email addresses of any non-Cambridge authors, before we can forward on the request.

As a result, it can take repository staff up to 30 minutes to process an individual request. This is quicker if the article has been requested previously and the author’s contact details are already stored, but can take longer when we need to search. Sometimes, there is also repeat correspondence if the author has any queries, which adds to the total time in processing each request.

Amending our processes

Since introducing ‘Request a copy’, we have started collecting the email addresses of corresponding authors when an article is submitted, and we have commissioned a repository development company to ensure that ‘Request a copy’ emails can be sent directly to those authors for whom we have an email address – a feature that we are hoping to implement in the next few weeks.

However, if the author moves institution, their university email address will no longer be valid, and any requests for their work will again need to come via repository staff. One way to solve this would be to ask for an external (non-university) email address for the corresponding author at the point where they upload the article to the repository. However, this would introduce an extra step to an already onerous process and may act as a further barrier to authors submitting articles in the first place.

Generally, ‘Request a copy’ is a great idea and provides many benefits to the research community and beyond. But the implementation of this service has been challenging. The amount of time taken by each request has meant that some staff members have been redeployed from their usual jobs to facilitate these requests, which also has an impact on the backlog of articles in the repository that need to be checked in case they have since been published. If an article is published but still in the backlog (and therefore not publicly available in the repository), unnecessary requests for it could result in a reputational issue for the Office of Scholarly Communication and the University.

We will continue to look at our processes over the coming academic year, to see how we can improve our current workflows, and identify and resolve any issues, as well as determining where best to focus any further development work. In the related blog post on ‘Request a copy’, I’ll be talking about usage statistics for the service so far, some more unexpected use cases we have encountered, and feedback from our users that will help us to shape the service into the future.

Published 7 October 2016
Written by Sarah Middle
Creative Commons License

Request a copy: uptake and user experience

This post looks at the University of Cambridge repository  ‘Request a copy’ service from the user’s perspective in terms of uptake so far, feedback we have received, and reasons why people might request a copy of a document in our repository. You may be interested in the related blog post on our ‘Request a copy’ service, which discusses the concept behind ‘Request a copy’, the process by which files are requested, and how this has been implemented at Cambridge

Usage Statistics

The Request a Copy button has been much more successful than we anticipated, particularly because there is no actual ‘button’. By the end of September 2016 (four months after the introduction of ‘Request a copy’), we had received 1120 requests (approximately 280 requests per month), the vast majority of which were for articles (68%) and theses (28%). The remaining 4% of requests were for datasets or other types of resource. We are aware that this is a particularly quiet time in the UK academic year, and expect that the number of requests will increase now term has started again.

Of the requests for articles during this period, 38% were fulfilled by the author sending a copy via the repository, and 4% were rejected by clicking the ‘Don’t send a copy’ button. However, these figures could be misleading as a number of authors have also advised us that they have entered into correspondence with the requester to ask them for further information about who they are and why they are interested in this research. Eventually, this correspondence may result in the author emailing a copy of the paper to the requester, but as this happens outside the repository, it does not appear in our fulfilment statistics. Therefore, we suspect the figure for accepted requests is in actual fact slightly higher.

Of the articles requested during this period, 45% were yet to be published, and 55% were published but not yet available to those without a subscription to the journal. The large number of requests made prior to publication indicates the value of having a policy where articles are submitted to the repository on acceptance rather than publication – there is clearly interest in accessing this research among the wider public, and if they are able to make use of it rather than waiting during the sometimes lengthy period between acceptance and publication, this can make the research process more efficient.

Author Survey

To find out why authors might not be fulfilling requests through the repository links, Dr Lauren Cadwallader, one of our Open Access Research Advisors, sent a survey on 6 July 2016 to the 113 authors who had received requests but had not clicked on the repository link or been in touch with repository staff to advise of an alternative course of action. This survey had a 13% response rate, with 15 participants, as well as eight email responses from users who provided feedback but did not complete the survey.

The relatively low response rate is indicative of either a lack of engagement with or awareness of the process – it is possible that the request emails and survey email were dismissed as spam, or that researchers were unable to respond due to an already heavy workload. One way of addressing this could be to include some information about ‘Request a copy’ in our existing training sessions, in particular to emphasise how quick the process can be in cases where the author is happy to approve the request without needing any further information from the user. We have also been developing the wording of the email sent to the author, to explain the purpose of the service more clearly, and to make it sound like a legitimate message that is less likely to be dismissed as spam.

Of the 15 people who participated in the survey, the majority were aware that they had received an email, which shows that lack of response is not always due to emails being lost in spam filters. When asked for the reason why they did not fulfil the request via the repository link, 35% of authors replied that they had emailed the requester directly, either to send the file, to request more information, or to explain why it was not possible for them to share the file at this time. This finding is quite positive, as it indicates that over a third of these requests are indeed being followed up. Although it would be helpful to us to be able to keep track of approvals through the system, at least this means that the service is fulfilling its purpose in providing a way for authors to interact with other interested researchers, and to share their work if appropriate. In fact, one of the aspects that participants liked best about the ‘Request a copy’ service was the ability to communicate directly with the requestor.

Two authors did not respond to the request because the article was available elsewhere on the internet, such as their personal / departmental website, or a preprint server (where the restrictions relating to repositories do not apply), although they did not communicate this to the requestor. In these cases, it is definitely positive that the authors are happy to share their work; however, it does show that there is often an assumption among researchers that people interested in reading their articles will be restricted to those already in their specific disciplinary communities.

Requests from people who are unaware of sites where the research might also be made available demonstrates that there is indeed an appetite among those outside of academia, or from different subject areas. This is generally a really positive thing, as it facilitates the University’s research outputs to educate and inspire a new audience beyond the more traditional communities, and could potentially lead to new collaboration opportunities. To ensure that requestors are able to access the material, and that researchers are not bombarded with requests for documents that are already freely available, authors can provide links to any external websites that are hosting a preprint version of the article, and we will add them to the repository record.

Other responses indicated that we were not necessarily emailing the right person, as participants said that they had not approved the request because they were not the corresponding author, or because they thought a co-author had already responded. At the outset of the service, we felt that emailing as many authors as possible would increase the likelihood of receiving a response; however, the survey results show that it would be better to send requests to the corresponding author(s) only, at least in cases where it is clear who they are.

An issue we have encountered on a semi-regular basis since HEFCE’s Open Access policy came into force is that of making an article’s metadata available prior to its publication. Although HEFCE and funder policies state that an article’s repository record should be discoverable, even if the article itself must be placed under embargo based on publisher restrictions, there is concern among some authors that metadata release breaches the publisher’s press embargo. You can read about this issue in some detail here.

Receiving requests for an article via the ‘Request a copy’ service can be unsettling for authors as it demonstrates how easily the repository record can be accessed, and rather than respond to the request, they contact the Open Access team to ask for the metadata record to be withdrawn until the article is published. This demonstrates a need to communicate more clearly, both on our website and within the ‘Request a copy’ pages in the repository, what is required of authors as part of HEFCE and funder Open Access policies. We will also be more explicit in the ‘Request a copy’ emails sent to authors in stating that sharing their articles via this service will not be seen as a breach of the publisher’s embargo. In cases where the author does not wish to disseminate their article before it is published, they have the option to deny any requests they receive.

Facilitating requests

There have been several instances where press interest around an article at the point of publication has generated a large number of requests, each of which must be responded to individually by the author. This has resulted in several authors asking that we automatically approve every request rather than forwarding them on. Unfortunately this is not possible for us to do, due to the legal issues surrounding ‘Request a copy’.

It is perfectly acceptable for an author to send a copy of their article to an individual, but if a repository makes that article available to everyone who requests it before the embargo has been lifted, this would be a breach of copyright because it would be ‘systematic distribution’. While responding to multiple requests is likely to be seen as an annoyance by an already overstretched researcher, we hope that a large volume of requests will also be viewed in a positive light, as it demonstrates the interest people have in their work.

Use cases

An interesting example of a request we received was actually from one of the authors of the article, as they did not have access to a copy themselves. This raises some questions about communication between the researchers in this case, if the ‘Request a copy’ service was seen to be a better way of gaining access to the author’s own research, rather than contacting one of their co-authors.

A more surprising use case is that of a plaintiff who had lost a legal case. The plaintiff was requesting an as-yet unpublished article that had been written about the case, because the article appears to argue in favour of the plaintiff and could potentially inform a future appeal. This is a good example of how the ‘Request a copy’ service could be of direct benefit in the world outside academia.

Although the vast majority of requests have been for research outputs such as articles, theses and datasets, we also occasionally receive requests for images that belong to collections held in different parts of the University, where high-quality versions are stored in the repository under restricted access conditions. With these requests, it can be more difficult to find who the copyright-holder is, which sometimes requires detective work by the repository team. In one case, permission had to be sought from a photographer who only has a postal address, and therefore required more explanation about the repository more generally, as well as the specific request.

Looking to the future

We will use this research and any further feedback we receive to improve the experience of our ‘Request a copy’ service for both authors and requestors, including implementing the ideas suggested above. Usage statistics will continue to be monitored, and we may run a user survey again to determine how far the service has improved, as well as to identify any new issues.

In the meantime, if you have any comments or questions about our ‘Request a copy’ service, either as an author or a requester (or both), please send us an email to support@repository.cam.ac.uk .

Published 7 October 2016
Written by Sarah Middle
Creative Commons License