Temporal Based Thematic Discovery and Characterization in the Domain of Human Computer Interaction and Information Behavior

Abstract

In this dissertation we proposed a combination bibliometric, graph theory and card sorting methods to discover and characterize the research themes in the domains of Human Computer Interaction/User Experience (HCI/UX) and Information Behavior (IB). For the first case, 519 papers, during the period of 1990-2016 were retrieved from Web of Science, published in the area HCI/UX using the search strategy (Human Computer Interaction and User Experience). The time-frame of the first research was petitioned into three time intervals (1990-1999, 2000-2009, and 2010-2016) to show Temporal based pattern discovery. The behavior related papers were found dominant in the case of HCI/UX analysis. Therefore, we focused on Information Behavior related aspects in our second research in this dissertation by selecting the representative journal of the clusters of citation network of journals related HCI/UX i.e. Computers in Human Behavior for Analysis. The aim is to make the in-depth exploration of the research themes Information Behavior within the general context of HCI/UX. 4771 papers published in journal of computers in human behavior starting 1990-2017 were included. The time span for the second research was partitioned into three, namely

1990-2003, 2004-2010 and 2011-2017. In both cases ADKs network was constructed and clustered for the three time periods using simple center algorithm. Clusters were considered as themes of research. Cluster networks were used to highly associated ADKs through their co-occurrence that formed a theme to help extract different research themes. The central ADK in a cluster network is used as a name of a theme name based on simple clustering algorithm, which gives more weight to the ADK with higher degree centrality in the cluster as representative of a cluster. The themes discovered through these process were grouped into different high level concepts perhaps subject matters addressed using card sorting methods by experts and color coded. Those color codes were used across the rest of the analysis i.e. evolution pattern discovery and strategic diagram based classification based on centrality and density into different roles and level internal maturity of themes. Evolution pattern discovery was used to show the evolution linkages of themes in different periods. This in turn gives insights to the level of paradigm shift (thematic dynamism) in the field. To show the conceptual periodic overlap, we used the overlapping map (stability diagram). It showed the level of newly emerged, obsolete, and overlapped ADKs in different periods. In both cases the number of thematic areas and conceptual (ADKs) stability increased while thematic dynamism increased over the time intervals. For example, in the case of HCI/UX domain, the stability of ADKs increased from 15% between in 2000-2009 to 52 % in during 2010-2016 while thematic dynamisms were 100% and 83% for similar periods respectively. The conceptual stability in Information Behavior has increased from 39% for the period 2004-2010 to 74% for the period 2011-2017 while thematic dynamism is 100% and 88% for those periods respectively. One, eight and twelve themes were discovered for the time intervals 1990-1999, 2000-20009, 2010-2016 respectively in case of HCI/UX. Three, eleven and thirty-four themes were discovered for the periods 1990-2003, 2004-2010, and 2011-2017 respectively in the case of Information Behavior. The variety and dynamics is huge for in the thematic areas of Information Behavior. In the case of high level concepts, in concepts six themes were related to measurements of HCI/UX, six themes were related to technology/systems, five themes were related to methods/approaches in the case of HCI/UX over the entire time span covered in the research. A total of 17 unique thematic areas were discovered over the entire time span. In the case of IB, seventeen themes belong to human factor/behavioral issues, eleven themes related to theories/concepts, ten themes belong to technology/systems, and seven themes are related learning environments. A total of 45 unique themes were detected in the IB domain for the entire time period. Overall, the proposed methods are effective to discover and characterize the thematic areas of research in both cases as we answered our research questions successfully. Therefore, these methods are promising in discovering and characterizing research themes in similar interdisciplinary fields of studies as are test successful on HCI/UX and IB domain, which are highly interdisciplinary domains.