An assessment of the performance of digital libraries based on an information systems impact model / Mohd Razilan Abdul Kadir

Digital Libraries (DL) have existed since the late 1990s; however, research on assessing the success of such Information Systems is still scarce. In the Information Systems (IS) literature, there are substantial discussions on measuring the IS success, but a firm consensus has yet to be achieved on the best criteria to assess the IS success. This is so because they are designed and developed for different uses and with diverse objectives. Some prominent success models are, however, too complex to realise in practice and to develop with a sound theory. Moreover, discussions on issues related to the Information Systems usage continue to gain significant impetus among Information Systems researchers. In addition, the use of formative indicators (as opposed to reflective indicators) in defining latent constructs has been understudied until recent years, with interest building up especially in the Information Systems domain. All of these concerns create a focus towards the formulation of the research objective to assess the performance of the digital libraries based on an IS Impact model. The IS Impact model, recently gaining growing attention among Information Systems researchers, is said to be the most comprehensive and validated model to date. It consists of four formative constructs namely individual impact, institutional impact, information quality and systems quality. This research, however, introduces an extended version of IS Impact (called DL Success model) by reconciling the individual impact dimension with the properties of the Information Systems usage theory (named as Digital Library Usage for Information Provisioning, DLUIP). Using a user-centred approach, a survey was conducted on 959 students, academic staff and library personnel at four research universities in Malaysia. The DL Success is modelled using the posited constructs in the literature by using a multi-dimensional formative partial least squares (PLS) path model to reveal the relationships between the variables of interest. Results from this study suggest that DLUIP is capable of explaining 52.5% of the variability in individual usage. However after incorporating awareness measures as mediating factors, the variability increased to 70.9%. In the full model, 73.9% of the success can be explained by the DL Success model, which shows a substantial degree of R2. All the weights for the formative measures and path coefficients for the inner and outer models show positive relationships, as hypothesised. Findings also suggest that the system and information dimensions are highly influential as compared to individual and institutional dimensions. The findings are significant in encouraging the use of formative measures in assessing the DL Success and as a formal means for library authorities to evaluate the digital libraries. The research has four main contributions. 1) the DL Success model can be added as a benchmark for the overall evaluation of academic DLs. 2) the conceptual understanding of the proposed success measures would better assist librarians towards improving the quality of the digital library services and functions. 3) extensive knowledge acquisition among academic users would serve as a basis in understanding the DL and the reasons for its use. 4) all the indicators posited in the study are capable of measuring the success of the academic DLs. Future research may focus on testing different orders of the formative model, and validating and generalising formative tests using guidelines discussed in the thesis. In addition, research could be extended to test the generalisability of the IS Impact model across different universities (public and private) in search for the best measures and dimensions to be the best surrogate of DL Success