The Actual Importance of RDA Elements in Supporting Key User Tasks

The Resource Description & Access (RDA) cataloguing standard stipulates sets of core elements for cataloguers to include in their description of information resources and in the resources' associated content and agents. The elements RDA deems core were selected according to their relative utility in meeting key user tasks, such as finding, selecting and identifying items and works, and finding and identifying persons and bodies associated with items and works. However, this selection was not based on empirical evidence, but on expert opinion. A study by Hider and Tan (2008) which investigated the relative value of various elements in catalogue records suggests that the RDA core elements may require revision. This new study examines the validity of the RDA core sets in more depth, employing think-aloud usability testing techniques and sets of typical bibliographic tasks, to gauge which of all the RDA elements are the most critical in a university library context. The results provide guidance on the adoption of the RDA standard and its elements, and on the value of fully implementing the Functional Requirements for Bibliographic Records (FRBR) and Functional Requirements for Authority Data (FRAD) models on which RDA is based. RDA elements and user tasks The new cataloguing standard, Resource Description & Access (RDA), departs from the Anglo-American Cataloguing Rules (AACR) in a number of important ways. It is based on a more comprehensive set of principles, and has been developed in consultation with a much larger and heterogeneous metadata community. Although certain details of the new code are debated, the principles which underpin it are widely accepted, including those it adopted from the Functional Requirements for Bibliographic Records (FRBR) and Functional Requirements for Authority Data (FRAD) reports. The development of the FRBR and FRAD models has likewise been profoundly influenced by the broader metadata community, particularly by members of the ‘Semantic Web’ movement (Dunsire 2010). The result is an RDA which specifies elements of bibliographic data much more clearly and systematically than does AACR with its eight areas. Ultimately, it has led to many more elements – that is, ones that are spelt out as such. For example, the copyright date is now an element in its own right, not merely mentioned in the rules for publication date as it is in AACR. By breaking free from the shackles of the International Standard for Bibliographic Description (ISBD), RDA has also provided scope for some new elements that are not covered by AACR, not even implicitly. Provenance, for instance, has been introduced for archival resources. RDA’s scope has expanded not only across domains, but also beyond ‘bibliographic description’ in the AACR sense: authority records and subject vocabularies, for example, are now covered, albeit to greater or lesser extents. All parts of a description based on RDA, whether it’s for a book, an author, or a subject, are particular elements as defined by RDA. The descriptions of the relationships between these things (entities) are also elements, in RDA’s language. Not surprisingly, we end up with a lot of elements. According to Danskin’s analysis (2009), there are 367 of them, if one counts elements, sub-elements and element-types. The RDA/FRBR/FRAD model is based on the principle that the catalogue exists to meet a set of particular user tasks. Both the FRBR and FRAD user tasks serve as the basis for the elements covered in RDA. Each RDA element supports one or more user task. The FRBR tasks are: find, identify, select and obtain access to resources. The FRAD tasks are: find and identify persons, bodies and families associated with resources; clarify the relationship between persons, bodies and families associated with resources; and understand why a particular name has been chosen for a person, body or family. Although some discussion has taken place regarding the exact nature of these tasks in different contexts (e.g. in the archival context), and although it has been acknowledged that some other tasks (such as those performed by librarians) are also important, the FRBR/FRAD user tasks are generally accepted as given. What is less clear, however, is the extent to which each RDA element supports the tasks. Of course, in a world of unlimited resources this would not matter, as all applicable elements could be included in descriptions. However, cataloguing is expensive, and there is a need more than ever to determine the most important elements to include. Whereas AACR employs three levels of description to indicate different levels of importance, RDA restricts itself to two. Each RDA element, at least in particular contexts, is deemed to be either ‘core’ or not ‘core’. This classification is broadly aligned with the minimal standards currently used in the vast majority of AACR/MARC-based libraries. It is also based on analysis of the elements’ support of the four FRBR user tasks and the ‘find’ and ‘identify’ FRAD user tasks (Kiorgaard 2008). However, this analysis has been criticized by Hider and Tan (2008) as over-reliant on expert opinion and lacking empirical evidence. Since most, if not all, of the (relatively few) experts consulted were heavily involved in the implementation of existing standards, their judgments cannot not be considered entirely independent. Hider and Tan (2008) also criticize the FRBR analysis for the way in which the experts’ ratings of ‘none’, ‘low’, ‘medium’ and ‘high’ importance were converted to points (0, 1, 2 and 3) and then aggregated. Although the Joint Steering Committee of RDA has called for user studies to support its work, most RDA-related research has not been of an empirical nature. The main user study that is currently taking place involves not end-users, but cataloguers (Library of Congress 2010). The library community is naturally concerned with how usable RDA is for its staff; but the fundamental question is how useful is its output for library patrons. Of course, this is very difficult to quantify in absolute terms, but the relative usefulness of the various RDA elements can be estimated empirically, at least in the context of a particular catalogue and its users. Unfortunately, the only empirical research published on this topic appears to be the studies conducted by Hider and Tan (2008), and by Hider (2007, 2008). Although these studies did not set out to evaluate either the AACR levels of description or the new set of core RDA elements, they did produce data that pointed to the need for such an evaluation. In the earlier studies by Hider, users at the National Library of Australia (Hider 2007) and at the State Library of Victoria (Hider 2008) were asked to rate the usefulness of various elements in their library catalogues. In the later study, a large sample of users of the National Library Board (NLB) system in Singapore was surveyed to elicit their views on elements’ usefulness in more depth; some think-aloud sessions with users were also conducted to triangulate the survey’s findings (Hider & Tan 2008). The surveys in all three studies were based on the particular catalogue’s interface and functionality. Therefore, the bibliographic elements evaluated varied across the studies, and were not directly based on either AACR or RDA. The surveys also distinguished between the ‘identify’ and ‘select’ tasks, but without affording any means of weighting the results according to the relative importance of the two tasks. Although the users’ ratings of elements might reflect, approximately, the order of their importance, it is unlikely that they would accurately reflect the degree to which one element was more important than another, given, for instance, the problem of differing interpretations of the scale used (seldom useful, sometimes useful, etc.). It is also quite possible that users tended to over-estimate the usefulness of particular elements for their actual searching, considering in addition the elements’ potential usefulness. Nevertheless, a table of relative values was produced from these studies, providing the basis, in the case of the Singapore study, for a measure of record quality (Hider & Tan 2008). Although a fair degree of correlation with the Australian results was found, it was by no means perfect, pointing to the importance of local policies based on local use. The study reported here features another catalogue context, but one that is no less typical than the ones studied earlier. The Charles Sturt University (CSU) library catalogue serves a medium-sized academic user community; the vast majority of its MARC21 records are derived from large bibliographic databases; its library management system is supplied by Ex Libris. If the most useful elements for users of this catalogue do not strongly correlate with the RDA core elements, it is likely that they do not for many other catalogue users. The RDA core element set will thus be evaluated in one particular context: the study does not attempt to establish a revised set of core elements for all libraries. However, the project does attempt to inform particular policy and practice through an investigation of real-time catalogue use, as reported in think-aloud sessions with users, as well as use reported through a survey questionnaire. Survey questionnaire The first stage of the project involved a questionnaire to establish a set of bibliographic tasks for which the CSU library catalogue was being used. It also provided the researchers with an opportunity to replicate the earlier studies’ method of assessing usefulness of catalogue elements, collecting responses to general questions of usefulness. The survey questionnaire was administered online via the CSU library catalogue’s home page, with the aim of attracting participants about to use the OPAC. Both remote and on-site users access the same home page, on which was placed an invitation to participate in the survey. Participants who completed the survey were invited to enter a draw for the chance to win one of two Apple iPads. The survey was administered in August and September 2010. The questionnaire consisted of six questions. A total of 182 responses were collected before the survey was closed, of which 158 indicated an intention to use the catalogue immediately after completing the survey, and provided a description of the task for which the catalogue was to be used. These descriptions were used as the primary source of bibliographic tasks for the second stage of the research. Participants were also asked to describe three tasks that they had recently carried out using the CSU catalogue; these descriptions have not been used in the research to date, but may be used at a later stage. The frequency of use of the CSU library catalogue amongst the 158 respondents (we assume that each response represents a different respondent; the software blocked multiple responses from the same IP address) is shown in table 1 below. Table 1. Frequency of catalogue use Frequency Response Percent Response Count This is my first time 7.6% 12 About once a month 23.4% 37 About once a week 29.7% 47 Two or three times a week 24.7% 39 Most days 14.6% 23 158 Of the 158 respondents, therefore, 146 were not (or did not report to be) first time-users. Of these, 124 responded to the fifth question in the survey, namely, ‘When you use the CSU library catalogue, how often do you search for’ the following types of resource, as shown in table 2. Table 2. Type of search catalogue used for Search type never occasionally sometimes often always a specific item 3 16 26 65 14 resources on a particular topic 1 6 9 78 30 resources by a particular author or organisation 5 34 36 38 11