By and large, collection evaluation studies look to how well highly-variegated and specialized materials support academic research and teaching. Universities engage bibliographers or subject liaison librarians to support the specific requirements of users for comprehensive and pertinent information through an ongoing program of selection of materials. While this is clearly defined in the higher education context (where staff often have formal qualifications in their area of specialization and constant dialog with faculty over selection is assumed) the selection role in public libraries is less-easily defined. As collection specialization does not exist in most public libraries it would seem reasonable to assume that the materials selector requires uncommon insight into the variety of subject areas that the library's users may require in order to provide a comprehensive collection. This paper aims to help better define what a specifically materials-centered focus for collection development in non-fiction public libraries might look like and how it might be better understood, specifically with reference to other approaches to collections which are self-referentially userfocused, and with the concept of subject, so crucial to the crux of the debate, at the forefront of consideration.
The public library context
Research on the nature of materials-centered collection evaluation within public libraries is not widespread and, subsequently, what is extant requires supplementation from generalist evaluation research. A starting point can be found in an early study of evaluation in public libraries conducted by Denny (1992, p. 9) who identifies a lack of research on evaluative methods and that 'practical applications of the theories and methods of collection assessment' had not found traction in public library environments. Denny (1992, p. 56). also points to how collection assessment and development activities in public libraries are often curtailed to meet operational requirements.
Public libraries are institutions that express both a visceral and a highly rational character. They are at once repositories for the accumulation and sanctification of types of knowledge that drive civic progress, while offering a value-free, encyclopedic approach to knowledge that does not explicitly privilege science, humanism or any particular epistemological creed. Creating a framework for understanding how such knowledge is sifted and filtered prior to its authorization on a library shelf is important because it allows the assumptions that govern that activity to be interrogated for reliability, truth, representative validity or verisimilitude. Do library users and library sponsors notice though? With reference to the Australian experience, Bundy (2010, p. 329) points to how public libraries tend to be well used regardless of their quality: 'People often do not know what constitutes a good public library, even if they have a sense that the library on which they depend is deficient. They thus do not know when they are being denied one'. This signals the important, albeit somewhat misunderstood role, that the institution plays in the civic and intellectual life of its users. Bundy (2010, p. 321-322) also highlights how the funding models for public libraries are inadequate to meet the demands of an increasingly information-reliant society Poor funding leads inexorably to poor collections. In this context of underfunded public library collections it is crucial that the best use is made of these limited resources for building collections. I argue that only a materials-centered focus can reveal how best this can be done.
Why inquire into materials-centered evaluation?
Where should the locus for evaluation of a collection be sited in a public library context--with the users or with the collection? Davis (1998, p. 54) provides a clear definition of these two currents through delineation of effectiveness as the primary parameter for use and content characteristics for a collection-centered approach. Denny (1992, p. 4) highlights where the horizon between the two might be located: a qualitatively acceptable and representative collection can be formed with a bibliography but risks the charge of sterility if the user's needs are not considered. Separate to easily characterized considerations of use and materials, selection theory, more broadly considered, across most of the twentieth century, generally coalesced into debates about needs versus wants, or quality versus quantity (Evans, 1995; Ameen, 2006). For materials-centered considerations in the public library we must assay what quality of materials are needed and what ranges and depths of subject treatment can satisfy a broad cross-section of users.
Comparison of collections against their own potential to meet their user's needs and to provide comprehensive and substantive subject coverage in line with their particular mandate has rarely been at the center of public libraries' collection concerns. Measures of public library effectiveness have historically been positioned as unrelated to collections except in the most tangential of ways. While library effectiveness might have commonly been benchmarked through measures such as collection expenditure or materials'use, such approaches cannot really facilitate collection evaluation. A distinction should be made because while the evaluation of the collection can provide information for assessing broader program goals, the existence of other measures of service quality should not detract from the responsibility to look to collections on their own terms; volumes per capita and circulation per volume are examples of an overt reliance on positive data at the expense of looking at why materials are not used, or, why the collection fails to meet the needs of the user due to less than adequate selection choices. It is worth differentiating library assessment from collection evaluation so that we might understand where they overlap, and where they differ, and how they approach measuring what they measure. From an assessment point of view, Lyons (2008) points to the varied ways that so-called input and output criteria interrelate and how only limited significance can be drawn from these criteria, especially when attempting to make comparisons between libraries. His analysis is somewhat more sanguine on the role of assessment:
Public libraries should be proactive in insisting that library assessment measures be developed in methodologically sound ways. They should also be vigilant in requiring that assessment findings be used to draw only conclusions clearly supported by the methodologies and data utilized. As a matter of policy, libraries should expect that assessment tools (are) accompanied by statements disclosing their limitations and potential biases, as well as guidelines for responsible interpretation and use of assessment findings. (Lyons 2008, p. 93)
Attempts have been made to quantitatively analyse the performance of public libraries in the United States using the Hennan Annual Public Library Ratings system and a more recent competitor, with differing priorities, the LJ Index of Public Library Service. While the approaches differed (with the former focusing on, among other things, collections and the latter being almost entirely user-centric) they both agreed that Ohio's public libraries were generally the best in the United States (it is unlikely to be a coincidence that Ohio's public libraries have exceptionally well-funded collections, see Klentzin 2010). While large collections invariably will have an advantage in meeting user needs and covering subject fields, not all libraries can collect or store a wide range of materials. What can they do to select well?
It is not practical to either keep collecting or, to collect everything, when budgets and facilities are limited. Whilst it is acknowledged in academic libraries that quality research collections take decades to build (Martin 2009; Shachaf and Shaw 2008; Wood 1996) public libraries require a strategy that is geared towards targeting the best available materials as soon as possible: resource allocation constraints (in terms of physical space, collection budgets and staff time devoted to acquisitions) and the changing paradigm that prioritizes digital and user-centered approaches to the library, all militate against the presumption that an incremental approach to collection development is the best strategy for modern public libraries (Pymm 2006; Vergueiro 1997). Howard and Davis (2011, p. 16) describe the complexity of collections as examples of wicked problems with 'interrelationships and interdependencies between objects, spaces, environments, and people'--within which 'we start to see each book or resource as part of a much more complex and macro system.' Usercentered approaches have a tendency to miss the complexity inherent in dealing with collections as they tend to largely ignore the epistemic issues that are inherent with the study of a particular knowledge domain. Examples can be trite, but the historical context of the scientific revolution to the Renaissance, World War 2's relationship to World War 1, the Cold War's relationship to liberal democracy all show how certain popular topics are contextually orphaned without proper epistemic reference to other areas of focus. Collections that gave us an oversupply of the Renaissance, World War 2 and the Cold War at the expense of the important linkages to other themes would be significantly diminished, whatever the quality of the works covering these popular themes, without the effort to make the connections necessary to promote a global contextuality of meaning.
Facilitating access through a browsable collection
Recent decades have seen growth in both online information sources and user's ability to exploit information sources that lie beyond their local collection. Despite this, there remains a...