●질적내용분석(Philipp Mayring)

질적내용분석

Qualitative Content Analysis

Philipp Mayring

Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, Vol 1, No 2 (2000)

1. Introduction

여기서 다루고자 하는 질적내용분석은 우리가 약 20년전 실직자들의 정신사회적 영향에 대한 종단연구를 개발하면서 활용한 여러가지 총합적 텍스트분석 기술을 다루고 있다. 약 600개의 개방형 인터뷰를 수행하면서, 2만페이지 이상의 기록물을 가지고 질적 분석을 했다.

The qualitative content analysis (MAYRING 1983; 7th edition 2000), as it is presented here, consists in a bundle of techniques for systematic text analysis which we developed ca. 20 years ago in a longitudinal study about psycho-social consequences of unemployment (ULICH, HAUSSER, MAYRING et al. 1985). Conducting about 600 open-ended interviews we received more than 20.000 pages of transcripts which had to be analyzed in a qualitative oriented way. [1]

이 분석과정의 핵심 아이디어는 의사소통과학 분야에서 개발된 양적내용분석의 장점을 보존하면서, 질적-해석적 분석단계로 끌고가는 것이다.

The main idea of the procedure of analysis is thereby, to preserve the advantages of quantitative content analysis as developed within communication science and to transfer and further develop them to qualitative-interpretative steps of analysis. [2]

양석내용분석에 대한 더 많은 정보는 아래의 사이트를 참조하라

Further information to quantitative content analysis are available via the Internet at http://www.gsu.edu/~wwwcom/content.html [Broken link, FQS, December 2004], http://www.zuma-mannheim.de/research/en/methods/textanalysis/ [broken link, September 2002, FQS] or http://www.aber.ac.uk/media/Sections/textan01.html. [3]

모든 기록된 의사소통의 내용물은 질적내용분석의 대상이 될 수 있다. 내용분석은 드러난 내용만을 분석하는 것이 아니다. BECKER & LISSMANN은 내용을 다양한 수준으로 구분했는데, 주제와 핵심 아이디어를 주요내용(primary content)로, 컨택스트 정보를 잠재내용(latent content)로 구분하였다. 형식에 대한 분석도 내용분석의 목적이 될 수 있다. 아래에서 설명하는 것과 같이, 내용분석은 텍스트를 의사소통모델에 대입시키게 된다. KRIPPENDORFF의 표현을 빌리자면 내용분석은 "텍스트로부터 구체적인 추론을 하기 위한 재현가능하고 타당한 방법"이다.

The object of (qualitative) content analysis can be all sort of recorded communication (transcripts of interviews, discourses, protocols of observations, video tapes, documents ...). Content analysis analyzes not only the manifest content of the material—as its name may suggest. BECKER & LISSMANN (1973) have differentiated levels of content: themes and main ideas of the text as primary content; context information as latent content. The analysis of formal aspects of the material belongs to its aims as well. As outlined below content analysis embeds the text into a model of communication within which it defines the aims of analysis. This is expressed by KRIPPENDORFF, who defines "content analysis as the use of replicable and valid method for making specific inferences from text to other states or properties of its source" (KRIPPENDORFF 1969, p.103). [4]

이러한 점에서 질적내용분석은 무분별한 정량화를 하지 않고 내용분석의 규칙과 단계별 모델을 따라가는 잘 통제된 실질적/방법론적 텍스트 분석 접근법이다.

Qualitative content analysis defines itself within this framework as an approach of empirical, methodological controlled analysis of texts within their context of communication, following content analytical rules and step by step models, without rash quantification. [5]

2. History of Content Analysis

역사적 배경은 다음과 같다.

We can distinguish different phases in the historical background of content analysis (cf. MERTEN 1983; KRIPPENDORFF 1980; MAYRING 1994a): 

• Precursors: We find different approaches to analysis and comparison of texts in hermeneutic contexts (e.g. bible interpretations) early newspaper analysis, graphological procedures, up to the dream analysis by Sigmund FREUD.
• Communication theoretical foundation: The basis of quantitative content analysis had been laid by Paul F. LAZARSFELD and Harold D. LASSWELL in USA during he 20ies and 30ies of 20th century. The first textbook about this method had been published (BERELSON 1952).
• Interdisciplinary broadening and differentiation: In the sixties of 20th century the methodological approach found its way into linguistics, psychology (cf. RUST 1983), sociology, history, arts etc. The procedures had been refined (fitting into different models of communication; analysis of non-verbal aspects, contingency analysis, computer applications) (cf. POOL 1959; GERBNER,HOLSTI, KRIPPENDORFF, PAISLEY & STONE 1969).
• Phase of qualitative critics: Since the middle of 20th century objections were raised against a superficial analysis without respecting latent contents and contexts, working with simplifying and distorting quantification (KRACAUER 1952). In the following qualitative approaches to content analysis had been developed (RITSERT 1972; MOSTYN 1985; WITTKOWSKI 1994; ALTHEIDE 1996). [6]

3. Basic Ideas of Content Analysis

질적내용분석이 양적내용분석의 장점을 보존하면서 좀 더 질적인 텍스트에 대한 분석을 하고자 하는 것이라면, 양적내용분석의 장점은 무엇일까?

If we say, qualitative content analysis wants to preserve the advantages of quantitative content analysis for a more qualitative text interpretation, so what are those advantages? I want to emphasize four points:

• Fitting the material into a model of communication: It should be determined on what part of the communication inferences shall be made, to aspects of the communicator (his experiences, opinions feelings), to the situation of text production, to the socio-cultural background, to the text itself or to the effect of the message.
• Rules of analysis: The material is to be analyzed step by step, following rules of procedure, devising the material into content analytical units.
• Categories in the center of analysis: The aspects of text interpretation, following the research questions, are putted into categories, which were carefully founded and revised within the process of analysis (feedback loops).
• Criteria of reliability and validity: The procedure has the pretension to be intersubjectively comprehensible, to compare he results with other studies in the sense of triangulation and to carry out checks for reliability. For estimating the inter-coder reliability we use in qualitative content analysis (in contrary to quantitative content analysis) only trained members of the project team and we reduce the standard of coder agreement (COHENS Kappa over .7 would be sufficient). [7]

4. Procedures of Qualitative Content Analysis

전술된 양적내용분석의 요소들은 질적내용분석의 근본에 그대로 보존될 것이다. 우리는 여러가지의 질적내용분석 절차를 개발했는데, 아래의 두 가지가 핵심이라고 할 수 있다.

The above listed components of quantitative content analysis will be preserved to be the fundament for a qualitative oriented procedure of text interpretation. We developed a number of procedures of qualitative content analysis (cf. MAYRING 2000) amongst which two approaches are central: inductive category development and deductive category application. [8]

4.1 Inductive category development

전통적인 양적내용분석은, 카테고리가 어디서부터 만들어졌고 어떻게 카테고리 시스템이 만들어졌는지에 대한 답을 거의 주지 못했다. "어떻게 카테고리가 정의되는지는 사실상 '예술'의 영역이다. 이에 대해 알려진 바는 거의 없다"

Classical quantitative content analysis has few answers to the question from where the categories come, how the system of categories is developed: "How categories are defined ... is an art. Little is written about it" (KRIPPENDORF 1980, p.76). [9] 

그러나 질적내용분석에서는 이 것이 핵심이며, 카테고리가 되는 가장 원자료에 근접한 분석의 측면을 개발하고, 원자료를 기반으로 이것을 만드는 것이다. 이러한 관점에서 질적내용분석은 '귀납적카테고리개발'단계를 만들게 되었는데, 텍스트가공 절차를 통한 환원적 과정이라고 할 수 있다.

But within the framework of qualitative approaches it would be of central interest, to develop the aspects of interpretation, the categories, as near as possible to the material, to formulate them in terms of the material. For that scope qualitative content analysis has developed procedures of inductive category development, which are oriented to the reductive processes formulated within the psychology of text processing (cf. BALLSTAEDT, MANDL, SCHNOTZ & TERGAN 1981; van DIJK 1980). [10]

구체적인 단계들이 이 짧은 그림에 다 담길 수는 없다. 핵심 아이디어는 기존의 이론과 연구질문을 바탕으로 특정한 정의에 대한 준거를 만들고, 이를 통해서 원자료의 여러 측면을 고려하는 것이다. 카테고리는 잠정적인 것이며, 각 단계를 밟으면서 추론을 하게 된다. 피드백과정을 통해서 카테고리를 검토하고, 궁극적으로 주 카테고리의 신뢰도를 확인하게 된다. 만약 연구질문이 양적 측면을 다루고자 한다면(빈도) 이를 분석할 수도 있다.

The specific steps cannot be explained largely within this short overview. The main idea of the procedure is, to formulate a criterion of definition, derived from theoretical background and research question, which determines the aspects of the textual material taken into account. Following this criterion the material is worked through and categories are tentative and step by step deduced. Within a feedback loop those categories are revised, eventually reduced to main categories and checked in respect to their reliability. If the research question suggests quantitative aspects (e.g. frequencies of coded categories) can be analyzed. [12]

4.2 Deductive category application

연역적 카테고리 적용은 기존에 만들어진 이론을 가지고 하게 된다. 이 분석 방법은 각 카테고리를 텍스트의 어떤 구절에 대입시키는 방법론적으로 통제된 절차를 밟게 된다. 

Deductive category application works with prior formulated, theoretical derived aspects of analysis, bringing them in connection with the text. The qualitative step of analysis consists in a methodological controlled assignment of the category to a passage of text. Even if several procedures of text analysis are processing that step, it is poorly described. Here the step model within qualitative content analysis: [13]

여기서의 핵심은 각각의 카테고리에 대해서 분명한 정의, 예시, 코딩규칙을 정하는 것이다. 이를 통해서 정확히 어떤 텍스트가 그 카테고리로 분류될 수 있는지를 결정할 수 있다. 이렇게 정의된 카테고리가 coding agenda에 같이 기록된다.

Then main idea here is to give explicit definitions, examples and coding rules for each deductive category, determining exactly under what circumstances a text passage can be coded with a category. Those category definitions are putted together within a coding agenda. [15]

카테고리 정의, 전형적 문구, 각 카테고리를 구분하는 법 등이 이론과 자료에 따라 만들어져야 하며, 분석 과정에서 계속 교정되게 된다.

Category definitions, prototypical text passages, and rules for distinguishing different categories were formulated in respect to theory and material, are completed step by step, and are revised with the process of analysis. [17]

5. Computer Programs for Support of Qualitative Content Analysis

몇 가지 컴퓨터 프로그램들이 개발되었다.

Especially within the last years several computer programs had been developed within the framework of qualitative analysis to support (not to replace) steps of text interpretation (cf. HUBER 1992; WEITZMAN & MILES 1995; MAYRING 1996; FIELDING & LEE 1998). The computer plays here a triple role:

• He works as assistant, supporting and making easier the steps of text analysis on screen (working through the material, underlining, writing marginal notes, defining category definitions and coding rules, recording comments on the material ...). He offers helpful tools handling the text (searching, jumping to different passages, collecting and editing passages ...).
• He works as documentation center, recording all steps of analysis of all interpreters, making the analysis comprehensible and replicable (e.g. to trace back in the material causes of non-reliabilities between two coders).
• He offers links to quantitative analysis (often already implemented within the program), e.g. to compare frequencies of categories, without the dangers of errors in data transfer by hand to another computer program. [18]

Working with qualitative content analysis two computer programs had especially proved it's worth, ATLAS/ti and winMAX, which both are available in free demo-versions (http://www.atlasti.de and http://www.winmax.de). [19]

6. Examples of Projects Working with Qualitative Content Analysis

To demonstrate the possibilities of qualitative content analysis we want to give some short examples of research projects working with the above explicated procedures:

7. Discussion

Communication science의 장점을 보존하면서 질적인 절차를 개발했다.

With the qualitative content analysis we wanted to describe procedures of systematic text analysis, which try to preserve the strengths of content analysis in communication science (theory reference, step models, model of communication, category leaded, criteria of validity and reliability) to develop qualitative procedures (inductive category development, summarizing, context analysis, deductive category application) which are methodological controlled. Those procedures allow a connection to quantitative steps of analysis if it seems meaningful for the analyst. [26]

질적내용분석은 다음과 같은 경우에 적절하지 않을 수 있다.

The procedures of qualitative content analysis seem less appropriate,

• 연구질문이 지나치게 개방적이거나, 탐색적이거나, 다양하거나, 카테고리를 두고 작업하는 것이 제한이 될 때
• 계획에 따른 단계별 분석절차보다는 holistic한 분석을 하고자 할 때
• if the research question is highly open-ended, explorative, variable and working with categories would be a restriction, or
• if a more holistic, not step-by-step ongoing of analysis is planned. [27]

질적내용분석은 다른 질적 연구방법과 함께 이용될 수도 있다. 연구질문이 어떤 연구방법을 주로 사용할 것인가의 판단근거가 될 수 있으며, 구체적인 내용 측면에 관해서 질문에 대한 논의가 필요하다.

On the other hand qualitative content analysis can be combined with other qualitative procedures. The research question and the characteristics of the material should have the priority in the decision about adapted methods. So it would be in my opinion better to discuss questions about methods in respect to specific content areas (cf. coping of illness MAYRING 1994b; emotion research SCHMITT & MAYRING 2000) and then to compare different methodological approaches (quantitative approaches as well). [28]

---

Qualitative Content Analysis

Philipp Mayring

Abstract: The article describes an approach of systematic, rule guided qualitative text analysis, which tries to preserve some methodological strengths of quantitative content analysis and widen them to a concept of qualitative procedure.

First the development of content analysis is delineated and the basic principles are explained (units of analysis, step models, working with categories, validity and reliability). Then the central procedures of qualitative content analysis, inductive development of categories and deductive application of categories, are worked out. The possibilities of computer programs in supporting those qualitative steps of analysis are shown and the possibilities and limits of the approach are discussed.

Key words: qualitative content analysis, content analysis, category, induction, intercoder-reliability