A hybrid approach to analysing large scale surveys: individual values, opinions and perceptions

Abstract High-dimensional large scale surveys enable broad research capabilities and potential insight. However, when dealing with the intrinsic complexity of social science, the underlying knowledge engineering process may play a critical role and require to consider the characteristics and peculiarities of a given problem in context. This study proposes an analysis framework based on clustering techniques, which have been applied to discover patterns among a number of abstracted features resulting from selected attributes of the World Values Survey (WVS). As an assumption, such features have been softly classified as values, opinions and perceptions, based on their theoretical likelihood to change along the time. From a more philosophical perspective, this work assumes hybrid practices as there is no pre-formulated theory but rather an attempt to discover patterns and new knowledge from data. Given the relatively manageable dimensionality of the input dataset, the feature selection has been performed according to an application-oriented approach (rather than driven by statistical analysis) to establish a more comprehensive and consistent research framework. Among the main findings, a symbiotic relationship between the perception of satisfaction and of financial stability, as well as between the perception of security and of happiness, in addition to more complex patterns involving traditional values (e.g. family and religion), politics and believes with an impact on society. Last but not least, despite its holistic focus, the study has allowed the identification of few research gaps and, therefore, potential further research direction in the broad domain of Social Sciences.