E-learning systems, or virtual learning environments (VLEs), are systems that use modern information and communication technology to support educational and training efforts. In this paper we present the evaluation of a newly developed Compendium Platform (CP) that can be used to create educational applications that support effective learning of statistics and associated analytical skills. (freestatistics.org) The novelty of the CP lies in the fact that students are empowered to easily archive, exchange, reproduce, and reuse statistical computations. The CP was implemented in three statistics courses and assessed (by students) in a survey. The goal of this article is two-fold: evaluating the success of the CP and proposing a predictive and re-usable success model (on which the survey was based).

The success and acceptance of information systems has been investigated in an overwhelming amount of studies. Recently, Davis’ widespread technology acceptance model (TAM) and the IS success model of Delone & Mclean have been fruitfully applied to study the success of VLEs in particular (see Sun et al., 2008; van Raai, 2008; Marins et al., 2004). Wixom & Todd(2005) have integrated both models to combine their strengths. They propose to use designoriented concepts (beliefs) as determinants of the TAM’s behavioural beliefs. In this way, a focus on design characteristics can be merged with IS usage prediction. Drawn from Wixom & Todd, we present a parsimonious model to evaluate the usability and acceptance of an elearning platform. System and information quality (design-based concepts) are proposed as determinants of perceived ease of use and relative advantage (which is an alternative to perceived usefulness and compares the VLE with conventional courses). In accordance with the TAM, ease of use and relative advantage are supposed to impact our major dependent variable: “intention to use CP in the future” (“intention”).

Because CP is a web-based solution, we turned to the website usability literature to measure system and information quality (both concepts include several dimensions). Multiple items were used per construct and scales from the literature were adapted or completed if required.

Based on a sample of 200 students, the scores on intention and relative advantage show that CP is well accepted. In particular, more than 150 students give a positive score for intention.

After performing reliability and validity checks, the “partial least squares” (PLS) method was used to test our model. We report that the majority of our hypotheses are confirmed. System quality has a considerable impact on ease of use and does directly influence intention. The effect of information quality on intention is fully mediated by relative advantage.

On a detailed level, “content sufficiency and effectiveness”, and “ease of finding information” are dimensions of information quality that correlate most with relative advantage and intention. The presence of sufficient functionalities and the attractiveness of the user interface are important dimensions of system quality, correlating strongly with ease of use and intention. In sum, (course) content, search options, and the user interface should receive special attention to improve the CP. In the future, new variables and new samples will be used to increase the explanatory power and practical use of the research model.