2.4. Evaluation Methods for Sustainable Creative Communities

In previous study, scholars typically adopted a qualitative methodology to examine the issue of non-sustainability in the development of creative communities. The qualitative data of these studies were mainly collected through methodologies such as field investigations, interviews and observations, and the data were analyzed with inductive reasoning [3,4,7,8]. When focusing on smaller and more concentrated samples, using a qualitative approach helps to achieve a more thorough understanding of the specific cases. This allows for a deeper understanding of causal connections in the study issue as well as a more nuanced and comprehensive discussion that includes diverse viewpoints. Moreover, compared to quantitative research, the results of qualitative research are more susceptible to influence of the researchers’ understanding and proficiency of the methodology [50]. Apart from that, the previous literature shows that, from the perspective of sustainable development, scholars increasingly use qualitative analysis to understand and assess the current developmental situation of creative communities. But they seldom achieve a systematic assessment and are, moreover, unable to generate a concrete improvement strategy based on the assessment results.

To date, sustainability science still seeks to integrate quantitative sustainability ideas into a new definition of how sustainability can be achieved [19]. As with most assessment research using quantitative analysis, what first needs to be done in the course of research on sustainability, is to analyze the assessment tools. In order to understand more clearly the priority of all criteria in the assessment framework, the weight of each criterion must be determined. The common method used for solving such problems in previous research was the analytic hierarchy process (AHP) [51,52], however, this method involves the problematic assumption that the criteria are all independent. But in practice there is always a certain degree of correlation or influential interrelationship among different factors, and even a certain degree of conflict between multiple attributes [53]. Therefore, scholars began to use the analytic network process (ANP) instead of AHP for assessment research [54,55,56]. Earlier, Saaty [57] proposed the ANP to solve dependence and feedback problems between dimensions (clusters), criteria (inner dimensions/clusters) in diagonal matrixes until they are assumed to be independent (zero matrix) or assumed to be self-related (identity matrix, I), and weighted supermatrixes obtained using equal weights (assumptions). As such, using ANP to determine the assessment of choice in practice is likely to result in relationships between the dimensions and criteria where their mutual effects are unclarified. This will affect the accuracy and applicability of the results of the analysis of decision-making. However, real-life problems are often quite complex because there are many mutual influential relations among the items. The D-DANP-mV (DEMATEL-based ANP with modified VIKOR) model is a hybrid method in the Multiple Criteria Decision-Making (MCDM) field. Based on the systematic concept, the main contribution of this model is to provide an improvement strategy for decision makers (DMs) and thereby improve performance of the alternative. The model breaks down the assumption that the relationships among variables are independent. It also emphasizes the influential relationships among the variables that replace the correlative relations in conventional regression analysis. From this perspective, the D-DANP-mV model is especially suitable for solving complex and practical problems.