|Sažetak rada na drugom jeziku (engleski)|| |
Web 2.0 is an umbrella term for the second generation of web applications that support different types of interaction among users and enable them to create, share, organize, and integrate various artefacts. Despite the fact that Web 2.0 applications are widely used for personal and business purposes, a comprehensive framework for their evaluation is still not available. With the aim to identify all the relevant pragmatic and hedonic attributes and their relative contribution to the quality in use composite index, as the main outcome of this doctoral thesis a methodology for evaluating the quality in use of Web 2.0 applications was developed. Based on the literature review, a set of 263 manifest variables was generated. By applying the closed card sorting procedure, domain experts assigned the set forth pool of manifest variables to the predetermined quality in use attributes. In the next step, the domain experts employed the open card sorting technique and thus categorized the set of 123 content-valid manifest variables. Drawing on the data collected from the domain experts, hierarchical cluster analysis revealed a finite set of 38 quality in use attributes that were afterwards used as the foundation for the design of metrics, measuring instruments, and the conceptual model. Two empirical studies constitute the essential part of this doctoral thesis. In the first study, the research subjects were observed during their interactions with two Web 2.0 applications meant for collaborative writing, and in the second, during their interactions with two Web 2.0 applicationsaimed for mind mapping. During each study, the participants performed representative scenarios of interaction with Web 2.0 applications which had been designed by the author of the doctoral thesis. The objective data were gathered during the interaction with the Web 2.0 applications, while the subjective data were collected afterwards. Users' efficiency (time on task) and workload (number of mouse clicks, mouse movements, mouse scrolls, and keystrokes) were measured by means of the Mousotron tool. The number of scenario steps that were successfully completedwith a particular Web 2.0 application was interpreted as an indicator of users' effectiveness. The subjective assessment of Web 2.0 applications was conducted with the use of an online questionnaire which was composed of 117 items related to all the relevant facets of the quality in use. With the objective to enable the evaluation of higher-order (multidimensional) aggregate constructs, a novel hybrid modelling technique was proposed. The psychometric characteristics of the conceptual model were exploredwith PLS-SEM method. An analysis of measurement and structural models parameter estimates in the context of the four evaluated Web 2.0 applications uncovered that the conceptual model met all validity and reliability criteria. Considering results of fourstructural models estimation, values of 0.15, 0.34, and 0.46 for endogenous latent variables can, as a rule of thumb in similar future empirical studies, be interpreted as weak, moderate, or substantial, respectively. On the basis of the conceptual model structure, a quality inuse requirements tree was designed in which each manifest variable was specified as a performance variable. Drawing on the pilot study results, elementary criteria were defined for all performance variables in the context of both types of Web 2.0 applications. The relative importance of each performance variable within the underlying subsystem of preferences was determined with the use of the group weighting method. By employing the logic scoring of preferences method, elementary preferences for each subsystem of performance variables as well as global preferences for all the four Web 2.0 applications were estimated from the quality in use perspective. Furthermore, values of elementary and global preferences were used for the analysis of strengths and weaknesses of the evaluated Web 2.0 applications as well as for their comparison within a group. Finally, based on empirical findings a set of guidelines and recommendations that can be applied in the form of heuristics for the design and evaluation of Web 2.0 applications at various stages of their life cycle was proposed.