Workpackage 2: Mediated Deposit Bureau (MDB)
Introduction
At the start of the WRN-EP it was recognised that of the 12 partner institutions, only twoAberystwyth University and Bangor University, had a dedicated member of staff working on the
repository system. These institutions were also the two that had the largest number of deposits,
Bangor being able to boast 7487 records in July 2009. This was in stark contrast to other partner
institutions which only had one item within their repository. It was thought that the number of items
deposited could be seen as a direct effect of staffing levels.
Within this workpackage an experimental, centralised, national mediated deposit service- the
Mediated Deposit Bureau (MDB), was created and utilised. The aim of the MDB was to work with a
selection of pilot institutions to outsource the deposit of items to their repositories to the WRN
team. The objectives of the MDB were to:
 Directly increase the number of deposited items within each repository.
 Identify per unit costs, in both time and money, for repository item deposit.
 Engage institution staff with the repository deposit process in an attempt to embed the
processes involved within their working cycle and further increase deposit.
 Encourage further deposit by academic staff within each repository through the presence of
new repository deposits.
Methodology
The first step in setting up the MDB was the identification of a small number of pilot institutions to
work with. Four partners expressed their interest in the service: Glyndŵr University; Swansea
Metropolitan University (SMU); University of Wales Institute Cardiff (UWIC); University of Wales
Newport. Shortly after the launch of the service Glyndŵr recruited a member of staff to take
responsibility for their repository so did not pursue their involvement with the MDB.
Institutions would be given the option to outsource the whole of an item’s deposit process to the
WRN including the checking of deposit permissions. As this would involve the interpretation of
Publisher and Intellectual Property Rights and could therefore, carry a certain level of risk for all
involved, a Memorandum of Understanding was drafted and signed by each institution as part of a
risk management strategy (Appendix A).
The next stage of the process was the identification of items within each institution to be outsourced
and deposited. It was thought easier for the process if each institution could identify discrete
collections of items for deposit rather than engaging the WRN team in a drip-feed of deposits. A
scoping document was drafted for institution’s to complete for each collection to be deposited
(Appendix B). Within this document an institution identified the collection to be submitted to the
MDB. They were asked what format the collection of items for deposit currently took, what level of
involvement they required from the WRN i.e. metadata enhancement of records; deposit permission
checks etc., and how and in what format they would like the items to be finally deposited.
Identified items
SMU were looking for the items submitted to their 2008 Research Assessment Exercise (RAE) return
to be deposited. The WRN team were provided with a PDF copy of the return which included full
metadata records for each item. The WRN team were asked to:
 enhance the metadata for repository deposit;
 check the deposit permission for each item;
 provide template e-mails for SMU staff use to obtain full items and/or further permissions
from authors or copyright holders;
 provide further information as needed to SMU staff so that the resultant repository records
could be made as complete as possible.
UWIC identified two collections for deposit from their School of Art and Design. The first collection
consisted of the original computer files of a collated group of research outputs held in UWIC’s
SharePoint site. The WRN team were tasked only with the creation of repository records for the
items held. UWIC staff would carry out any copyright checking necessary and upload the full-item
files where possible at a later date. The second collection was an Access database of records for the
School’s Ceramic Research Archive. Again the request for the WRN team was only to enhance the
metadata for repository deposit.
Staff at University of Wales Newport had collated citation lists detailing publications produced by
academics from each of the University’s academic schools. As with the SMU RAE records, the WRN
were asked to:
 enhance the metadata for repository deposit;
 check the deposit permission for each item;
 provide template e-mails for SMU staff use to obtain full items and/or further permissions
from authors or copyright holders;
 provide further information as needed to SMU staff so that the resultant repository records
could be made as complete as possible.
Full case studies describing the specific approach taken with each collection within the MDB are
attached: UWIC SharePoint items (Appendix C); Newport Publication Lists (Appendix D); SMU RAE
(Appendix E); UWIC Ceramic Archive (Appendix F).
Results
Specific timings for the processing of individual items within each collection are available within the
case studies. However, it was thought useful to highlight certain results for the purposes of the MDB.
The table below shows the total amount of time taken with each MDB collection, the cost of that
time, the average time taken per item and the average cost for processing each item in a given
collection.
Institution
Collection for
deposit
UWIC
SharePoint items
Newport
SMU
Nature of deposit
Item identification and
duplicate checks
Records created directly
on repository system
Ceramics database Database data
transferred, tidied and
bulk uploaded
Sub-total
No. of 7.5 hour days
Citation data
Record enhancement
and bulk upload
Art citations Duplicate checks
Records created directly
on repository system
Sub-total
No. of 7.5 hour days
RAE data
Record enhancement
and bulk upload
Sub-total
No. of 7.5 hour days
TOTAL
Total no. of 7.5 hour days
Total time taken
(hours: minutes: seconds)
No. of
items
Average time
taken per item
(minutes: seconds)
Cost in £
per item
3: 00: 00
Cost in £ at
£14.03 per
hour
42.09
1: 45: 46
24.79
9
11: 45
2.75
73: 48: 00
1035.41
1505
02: 57
0.69
4
11: 53
2.78
130 (timed)
15: 37
3.65
78: 33: 46
1102.29
10 days 3 hours 33 minutes 46 seconds
44: 50: 00
629.01
01: 00: 00
00: 47: 32
11.11
46: 37: 32
640.12
6 days I hour 37 minutes 32 seconds
33: 50: 00
474.68
33: 50: 00
4 days 3 hours 50 minutes
159: 01: 18
2217.09
21 days 1 hour 31 minutes 18 seconds
There were errors with the recorded processing times for the bulk upload items within the Newport collection. Therefore, it is not possible to provide any
average data for this part of the collection. A greater amount of time was spent processing these items therefore, the cost should also be considered
incorrect. The figure provided merely reflects what is discernable from the data recorded.
The times given above reflect the total amount of time taken for MDB work. However, it should be noted that this work was in fact conducted over a period
of 9 months along with other WRN project work. The first deposit took place on 26 March 2010, with the last deposit taking place on 9 December 2010.
As can be seen from the data above, the average time taken per item was very similar between the
UWIC SharePoint items and the Newport Art citation items (records for both sets were added
directly via the repository interface). There was a difference of 8 seconds between the two average
times with Newport records taking the longer length of time to complete. Deposit checks were
performed on these items, where no such check was performed on the UWIC items. Also, it was
possible to locate and upload a full-text copy of an article processed within the Newport collection
which added time to the processing of this item and would have affected the average time
calculated. However, it is difficult to imply that the deposit checks were the cause of the difference
in time. The original metadata sources for the two collections were very different; data for the UWIC
items was taken directly from the original files; data from the Newport items came from citations.
The types of items processed within each set were slightly different also. The UWIC collection
consisted of Articles, Presentations, a Book chapter and an image. The Newport collection consisted
of three Presentations and one Article.
In contrast, the average time taken per item within the SMU collection is almost 4 minutes longer
than the Newport or UWIC averages, at 15 minutes 37 seconds. This value included the length of
time taken for bulk upload preparation and deposit. This is a valid addition as it is important to
consider the length of time taken to take the original metadata source to a submitted repository
record. It can be said then that within the MDB it took a greater length of time to process and
prepare items for bulk upload than to process and deposit items directly via the repository.
Item type processing comparison by collection
Through the data collected within the UWIC SharePoint items, Newport and SMU MDB collections it
is also possible to compare the average times taken for processing particular item types. This was
thought useful as it provides some data on the differences in times where the originating data for a
record is significantly different, and where deposit permissions have and have not been conducted.
Institution Deposit type
Deposit
check?
UWIC
Newport
SMU
Records created directly
on repository system
Records created directly
on repository system
Record enhancement
and bulk upload
No
Average time taken
per Article
(minutes: seconds)
09: 26
Average time taken
per Presentation
(minutes: seconds)
11: 57
Yes
17: 06 (only 1 item)
10: 08
Yes
07: 26
Unfortunately, no real conclusions can be drawn from this data. It can be seen that it took almost
two minutes longer to process Presentations within the UWIC collection than within the Newport
collection. However, it took considerably longer to process one article within the Newport collection
than the average time taken to process an Article within the UWIC collection. It therefore cannot be
said that the time difference is due to copyright checking. In fact, as it took even less time to process
Articles within the SMU collection, where copyright checking was conducted, this acts to disprove
the argument. The differences in time may just be due to the nature of the items processed. It must
also be remembered that the sample sizes within both the Newport and UWIC collections were very
small.
Item type processing times
Through the data recorded within the SMU MDB collection it has been possible to identify average
times for enhancing the metadata records and checking the deposit permissions of particular item
types. These can be seen in the table below.
Item type
No. of items Total time taken
for processing
(minutes)
Artefact
9
76
Article
51
379
Book
9
73
Book chapter
11
121
conferenceObject 18
271
Exhibition
22
217
Video
9
136
Average time taken per item
(minutes: seconds)
08: 27
07: 26
08: 07
11: 00
15: 03
09: 52
15: 07
It must be noted that these timings relate to items where a full metadata record was already present
and which were processed outside of a repository system for later upload, the timings for which
have not been included.
UWIC Ceramics Archive
This collection was significantly different to the other collections involved, dealing with a unique
data source, and therefore no real comparisons can be drawn from the data.
Evaluation
After all the deposits had been made the three pilot institutions were sent a short evaluation
questionnaire of 10 questions regarding their involvement with the MDB. In their responses to the
questionnaire all partners indicated that:
 They felt their institution’s participation within the MDB was worthwhile
 The outcomes of the MDB process were as they expected
 They were happy with the overall approach for deposit taken in each case
 They were happy with the level of involvement they had with any decision making processes
 They felt they were sufficiently informed of progress during the MDB process
 They were happy with the quality of the records they received
 They were satisfied with the level of support given post-deposit for full-item requests/
record completion
UWIC commented that the length of time taken to make the deposits was longer than they
expected. However, as the service was being provided free as part of the rest of the work of the
WRN-EP team, they took no objection to this.
Partners were asked whether circumstances had changed within their institution which would have
meant the records put forward for the MDB could have been processed in house. Both UWIC and
Newport reported that during the 9 months that the MDB was actively depositing items, both
institutions actually lost staff that may have been able to process the MDB items. SMU reported that
the item submissions ‘just wouldn’t have happened without the MDB.’
Only UWIC could report real evidence of increased repository interaction by academic staff as a
result of the deposit of their MDB items. It had been recognised that there had been an increase in
the number of staff from the School of Art and Design who had a) registered to use the repository,
or b) asked to deposit their material, although no-one in this group had actually done so at the time
of the evaluation. Although no evidence could be reported, both Newport and SMU realised the
potential of the new deposit of items in engaging academics with the repository.
Conclusions
The objectives of the MDB were to:
 Directly increase the number of deposited items within each repository.
 Identify per unit costs, in both time and money, for repository item deposit.
 Engage institution staff with the repository deposit process in an attempt to embed the
processes involved within their working cycle and further increase deposit.
 Encourage further deposit by academic staff within each repository through the presence of
new repository deposits.
It can be seen that each of these objectives have been met and that the pilot institutions were
extremely positive about their involvement with the MDB.
Some useful results have been drawn from the processes of the MDB in regards to per unit costs in
both time and money. The most useful data can be drawn from the traditional research outputs
processed within the UWIC SharePoint, Newport and SMU collections.
The UWIC Ceramic Archive was significantly different to these other collections so no comparative
data could be drawn. However, it is hoped that the case study information may help others if
presented with the task of transferring bespoke database records to a repository.