소비자 집단 인터뷰에서 의미 네트워크 응집 구조의 이해 = Understanding the Semantic Network Structure in the Consumer Group Interview with the Subnetwork Analysis

다국어 초록 (Multilingual Abstract)

Researchers have performed the qualitative studies utilizing consumer group interviews to reveal consumers`s responses in depth. In general, most of the qualitative studies have analyzed by researchers` examination every aspect of participants` opinions and interpretation of the distinctive responses. There is concern in this approach that the results may have validity issues, because it is difficult for researcher to keep objective viewpoint. Some studies suggest an alternative methodology like content analysis and LIWC, but these approaches still remain unsolved subjectivity issues. First, content analysis has some disadvantages. Coders encode the qualitative data so the researcher could analyze in the further analysis in a rather quantitative ways. But it is possible that the encoded data may contain the coders` subjectivity according to the coders` understanding of coding schemes. And sometimes the coding scheme implies the researcher`s hypothesis implicitly. Second, LIWC is a kind of text analysis software program which is designed to calculate the degree people use different categories of words across a wide array of texts. It categorizes about 3,000 words into 72 categories. LIWC helps researchers understand the qualitative data more objectively in some degree. Nevertheless it has limit to interpret the words beyond the pre-defined ones designated in the categories. This research proposes a semantic network analysis as an alternative method for the existing qualitative methodologies. The purpose of network analysis is to investigate the relation between linked nodes. If it is applied to the relationship between shared meanings in linguistic communications, then it represents semantic network. In this case, a word becomes a node and a cooccurrance of two words in a sentence becomes a link. Linked nodes are the basic element of semantic network. Semantic network analysis uses words which are included in a qualitative data, and does not exclude any of them by the subjective judgement basis. Thus it can overcome validity issues previously associated with content analysis and LIWC. Semantic network analysis is relatively less likely to be reflected the subjective viewpoint of the researcher and gives us a chance to find out inherent contents structure of qualitative data. Semantic network analysis produces several network indexes which reflect the characteristics of network or node of interest such as density and degree centrality. Network density is calculated the number of connected links in a network compared to the number of possible links. High density implies that people are very interested in the topic in the qualitative data, because very dense network is resulted from highly related topics and keywords in the context of the semantic network. Degree centrality is the number of nodes directly connected to an individual node. A word with a high degree centrality is able to activate many other words, thus it functions as a keyword within the semantic cluster. In this study, we try to make the theoretical background of semantic network analysis base on the existing theories of cognitive linguistics, cognitive psychology, and sociolinguistics. A word which conveys a concept in individual cognitive structure is represented as a node in a semantic network analysis, and researchers may interpret the relation between linked words. One semantic network is composed of several subnetworks which imply the cohesive group of similar concepts within cognitive structure. Moreover language has common meaning between individuals who are belonging to the same social groups through interaction of group members. Therefore we can interpret the subnetwork of semantic network in the group interview as the clique of words which are about one topic. As a result, we can understand the entire semantic network contents structure through interpreting subnetworks. Subnetwork also can be represented by the indices such as density and degree centrality. The process of semantic network analysis is as follows. First, meaningful keywords are extracted from an entire qualitative data. Second, cooccurrance frequency of keywords within one sentence is calculated. Finally, the network analysis is performed, then subnetwork and the network indexes of each subnetwork like density and degree centrality are extracted. The exemplary analysis is shown using consumer group interview data regarding the internet shopping values. The managerial implication and the future research directions are discussed.

참고문헌 (Reference)

• 1 한덕웅, "집단행동이론" 시그마프레스 2002
• 2 조의연, "의미란무엇인가? 인지의미론과해체주의" 2 : 117-128, 1996
• 3 이혜준, "의미 네트워크 분석을 통한 프랜차이즈 교육 프로그램 개발" 한국경영학회 14 (14): 105-128, 2010
• 4 한승희, "용어 클러스터링을 이용한 단일문서 키워드 추출에 관한 연구" 한국문헌정보학회 44 (44): 155-173, 2010
• 5 옥경영, "온라인커뮤니티 사이트에서 하위네트워크의 특성이 정보의 초기 확산에 미치는 영향에 관한 연구" 한국소비자학회 20 (20): 37-64, 2009
• 6 이창환, "언어적 특성을 이용한 '심리학적 한국어 글분석 프로그램(KLIWC)'의 개발과정에 대한 고찰" 한국인지과학회 16 (16): 93-121, 2005
• 7 이정화, "심리치료과정 분석 도구로서 K-LIWC과 통찰수준 및 감정표현 분류척도의 비교 연구" 한국임상심리학회 29 (29): 1009-1028, 2010
• 8 한상만, "사회적 네트워크에서의 고객무형가치에 대한 연구" 한국마케팅학회 10 (10): 99-121, 2009
• 9 김용학, "사회 네트워크 이론" 박영사 2007
• 10 한상만, "마케팅에서 Network 연구를 위한 탐색적 고찰" 한국소비자학회 17 (17): 61-88, 2006
• 11 이창환, "남북한 방송언어의 차이에 대한 기초 분석" 한국산학기술학회 11 (11): 3311-3317, 2010
• 12 김봉주, "개념학 : 의미론의 기초" 한신문화사 1988
• 13 이현근, "개념론적 및 인지론적 어의 연구" 6 (6): 159-177, 1999
• 14 Slatcher, Richard B., "Winning words: Individual differences in linguistic style among U.S. presidential and vice presidential candidates" 41 : 63-75, 2007
• 15 Bellotti, Elisa, "What are friends for? Elective communities of single people" 30 (30): 318-329, 2008
• 16 Odgen, C. K., "TheMeaning of Meaning" Routledge and Kegan Paul 1923
• 17 Granovetter, Mark S., "The Strength of Weak Ties" 78 (78): 1360-1380, 1973
• 18 Antia, Kersi D., "The Severity of Contract Enforcement in Interfirm Channel Relationships" 65 (65): 67-81, 2001
• 19 Radil, Steven M., "Spatializing Social Networks: UsingSocial Network Analysis to Investigate Geographiesof Gang Rivalry, Territoriality,and Violence in Los Angeles" 100 (100): 307-326, 2010
• 20 Wasserman, Stanley, "Social network analysis: methods and applications" Cambridge University Press 1998
• 21 Ryan, John Barry, "Social Networks as a Shortcut to Correct Voting" 55 (55): 753-766, 2011
• 22 Scott, John, "Social Network Analysis" SAGE 2011
• 23 Van Atteveldt, Wouter Hendrick, "Semantic Network Analysis: Techniques for Extracting, Representing, and Querying Media Content" BookSurge Publishing 2008
• 24 Lakoff, George, "Moral Politics: What Conservatives KnowThat Liberals Don't" University of Chicago Press 1996
• 25 Newman, M. E. J., "Modularity and community structure in networks" 103 (103): 8577-8582, 2006
• 26 Pennebaker, James W, "Linguistic Styles: Language Use as anIndividual Difference" 77 (77): 1296-1312, 1999
• 27 Pennebaker, James W., "Linguistic Inquiry and Word Count" Lawrence Erlbaum 2001
• 28 Taylor, John R., "Linguistic Categorization: Prototypes in Linguistic Theory, 2nd ed" Oxford University Press 1995
• 29 Forsyth, Donelson R., "Group Dynamics, 5th ed" Wadsworth 2010
• 30 Ruigrok, Nel, "Global Angling with a Local Angle: HowU.S., British, and Dutch Newspapers FrameGlobal and Local Terrorist Attacks" 12 (12): 68-90, 2007
• 31 Fillmore, Charles, "Frame Semantics, In Cognitive Linguistic: Basic Readings" Mouton de Gruyter 373-400, 2006
• 32 Newman, M. E. J., "Finding and Evaluating Community Structure in Networks" 69 (69): 69-83, 2004
• 33 Falzon, Lucia, "Determining groups from the clique structure in large social networks" 22 (22): 159-172, 2000
• 34 Newman, M. E. J., "Detecting community structure in networks" 38 (38): 321-330, 2004
• 35 Girvan, M, "Community Structure in Social and Biological Networks" 99 (99): 7821-7826, 2002
• 36 Radden, Günter, "Cognitive English Grammar, Amsterdam and Philadelphia" John Benjamins Publishing Company 2010
• 37 Freeman, Linton C, "Centrality in Social Networks Conceptual Clarification" 1 (1): 215-239, 1979
• 38 Fiske, Susan. T, "Category-Based Versus Piecemeal-BasedAffective Responses: Development in Schema-Triggered Affect, In Handbook of Motivationand Cognition: Foundations ofsocial Behavior" Guilford Press 167-203, 1986
• 39 Smith, Edward E, "Categories and Concepts" Harvard Business Press 1981
• 40 Henderson, Geraldine R, "Brand diagnostics: Mapping branding effects using consumer associative networks" 111 (111): 306-327, 1998
• 41 Grebitus, Carola, "Analyzing semantic networks of pork quality by means of concept mapping" 19 (19): 86-96, 2008
• 42 Tajfel, Henri, "An integrative theory of intergroup conflict, In The social psychology of intergroup relations" Brooks/Cole 33-47, 1979
• 43 Cohen, Joel B., "Alternative Models of Categorization: Toward a contigent Processing Framework" 13 (13): 455-473, 1987
• 44 Barsalou, Lawrence. W., "Ad hoc categories" 11 (11): 211-227, 1983
• 45 Collins, Allan M., "A spreading-activation theory of semanticprocessing" 82 (82): 407-428, 1975
• 46 Hsieh, Mo-Han, "A new method for finding hierarchical subgroupsfrom networks" 32 (32): 234-244, 2011
• 47 Anderson, John R., "A Spreading Activation Theory of Memory" 22 (22): 261-295, 1983
• 48 Tyler, Kimberly A., "A Qualitative Study of the Formation and Composition of Social Networks Among Homeless Youth" 21 (21): 802-17, 2011
• 49 Fiske, Susan T., "A Continuum of Impression Formation, From Category-Based to Individuating Processes: Influence of Information and Motivation on Attention and Interpretation" 23 : 1-74, 1990