rm-lec-12

QUALITATIVE ANALYSIS

INTRODUCTION

The characteristic of qualitative analysis is that it deals with data presenter in words; that it contains a minimum of quantitative measurement, standardization and statistical techniques, and that it aims to transform and interpret qualitative data in a precisely and, scholarly manner. Beyond this there is simply no consensus as to how qualitative analysis should proceed, or what makes an acceptable analysis. Writers on the subject more that the process of analysis is deeper, more focused and more detailed than in quantitative research and takes one of two options. These are analytic induction and grounded theory. But the list of options for qualitative analysis does not stop here.

Other writers describe further discursive approaches, based mainly on discourse analysis, deconstructionism, conversation analysis, and narrative analysis. The common elements of these analytic methods are their affiliation with interpretive paradigms, their understanding of the world as being socially constructed through language.

1          THE TIMING OF QUALITATIVE ANALYSIS

In some cases it follows the same path as in quantitative research, but in most cases it is conducted during data collection. A combination of both models is also possible.

1.1       Analysis during data collection

Analysis during data collection is the most common practice, and the one that is most consistent with the principles of qualitative research. In this case, data are collected, coded, conceptually organized, interrelated, analyzed, evaluated and then used as a spring-board for further sampling, data collection, processing and analysis, until saturation is achieved. Data collection is thus merged with data analysis.

For instance, Field researchers scrutinize the collected data thoroughly, and may even meet with colleagues to discuss their findings, compare notes, check consistency, conduct data analysis and draw conclusions before continuing with further data collection.

The process of analysis In qualitative analysis researchers: ·         Focus on the gathered data, reading them within the context of the setting and the research purpose, and usually employing a basic, descriptive content analysis. ·         Identify chunks of data that demonstrate some commonalities and assign them relevant codes. ·         Note thoughts and initial reflections on parts of data in memos. ·         Work through the materials, to identify patterns, themes, sequences, differences and so on. ·         Construct matrices, network maps, flow charts, diagrams and other relevant presentation material. ·         Gather new data, and work through it as in previous steps. ·         Establish commonalities and eliminate negative cases, leading to some degree of consistency and to small-scale generalizations. ·         Link generalizations together, testing and re-testing, contrasting, and comparing constantly, leading to typologies and eventually to theories,

1.2       Analysis after data collection

There are even cases where qualitative analysis is wholly conducted after data collection. This is, for instance, the case when the recording of data is facilitated electronically, for example on video and audio tapes. In this case, the main analysis is conducted after data collection, when the videos are viewed. The type of analysis varies according to the nature of the study,

1.3       Analysis during and after data collection

While collecting data, researchers conduct some basic analysis, record the data and then intensify their analysis when the study is completed by focusing on more specific aspects of the research question as shown in the transcripts. In a number of cases, analysis during data collection serves to guide research in the right direction, and to facilitate a more effective treatment and coverage of the research topic. The actual analysis is conducted after data gathering is completed.

2          VARIETIES OF QUALITATIVE ANALYSIS

The process of analysis can be categorized in three distinct groups. These are the fixed, the iterative and the subjectivist models of analysis. Briefly, the fixed model is employed when data analysis occurs after collection, and the iterative model is employed when analysis is conducted while collecting data. The subjectivist model follows few - if any - rules, and hence one can say little about it (see Box 15.2)

Main types of qualitative analysis ·         Iterative qualitative analysis: This analytic method employs two major strategies. These are grounded theory and analytic induction. They are characterized by the fact that their analytic process involves repeated use of data collection and analysis. ·         Fixed qualitative analysis: This method of analysis is employed mainly when research followed a fixed research design. In this case, data analysis is conducted after data collection, and chiefly entails a method of content analysis or text analysis. The type of method depends on the nature of the data, the purpose of analysis and the nature of the underlying paradigm. ·         Subjectivist qualitative analysis: This type of qualitative analysis in effect covers whatever cannot fit into the other two types of analysis. All these types share their relative aversion to structure research and analysis, and a mistrust of strict techniques and methods. In a number of cases, the disbelief in objective truth, as well as in the ability of standard methods to extract knowledge, leads to a subjective choice of methods that do not rely on general rules and principles.

3          GROUNDED THEORY

Aspects of grounded analysis ·         Identifying indicators of a concept. ·         Studying indicators and comparing them with each other. ·         Coding indicators, looking for answers and formulating hypotheses. ·         Categorizing similar indicators as a class. ·         Naming the class and perceiving it as a coded category, which reflects the indicators' similarities, and the smallest common denominator. This is the conceptual code, the concept. ·         Comparing indicators with concepts and with other indicators; this helps to refine them and relate them optimally to the data. ·         Working through more attributes of the categories, refining them and getting additional information until the codes are tested and saturated. The more similar the indicators regarding the concept in question, the higher the degree of saturation of the attributes of the concept. ·         Developing and saturating more categories through the process of constant comparisons, ·         Including in the theory concepts and their attributes developed as described. ·         Further testing, contrasting and comparing of theories, and perhaps refining and changing them.

Instruments of analysis

Concepts and indicators

Indicators are concrete data, such as behaviour patterns and social events, which are observed or described in documents or interview texts. They indicate the presence of a concept which the researcher develops, at first provisionally and then with more confidence. Such concepts occupy a central position in the context of grounded theory.

Categories

A category is a unit of information that involves events, happenings, and instances. Categories that are adequately processed lead to the identification of key categories which assist with theoretical sampling and data collection. A key category is one that possesses the power required to explain the largest degree of variance in behaviour; it also helps to integrate, tighten and saturate a theory. A key category must meet the following criteria. It

·         must be neutral

·         must appear frequently in the data

·         should allow easy reference to other categories

·         should possess clear implications for the formal theory

Theoretical sampling

In grounded theory the subjects, settings and events are selected predominantly during the process of data collection and analysis. Here sampling begins with a guided selection of the first unit. The information collected from the first respondent guides the researcher to select the next respondent. The information that emerges through the study of the two respondents will lead to the selection of the third respondent and so on. More specifically, the emerging theory controls the choice of the respondents; this is why it is called theoretical sampling.

Theoretical saturation

Saturation indicates the Stage in the research process at which no new or relevant data emerge, the category is well developed, and the relationships among categories are well established and validated.

Coding

Within grounded theory, coding is the central pathway to theory construction; hence it is important that it is performed in an accurate and disciplined manner.

Be cautious with coding ·         Do not paraphrase sentences but discover and name genuine categories. ·         Set categories as directed by the coding paradigm (i.e. in reference to conditions, strategies, consequences etc,). ·         Set categories in relationship with subcategories, one by one. ·         Relate categories to data and refer to them adequately. ·         Stress the factors that make sorting easier. ·         Interrelate categories and subcategories. ·         Relate or eliminate unrelated categories.

Within these parameters, coding appears in three forms: open, axial and selective coding.

Open coding

The purpose of open coding is to identify first-order concepts and substantive codes. In this process, the researcher begins by opening up the data: breaking it down and looking for empirical indicators of concepts. This implies taking notes constantly, sorting them out, looking for meanings, and comparing notes. Following this, the researcher labels the data, and while working through them, changes the labels as required.

Axial coding

Axial coding (also known as theoretical coding) is a more advanced level of coding and aims to interconnect substantive codes and first-order concepts to construct higher-order concepts. In this form of coding, the researcher puts in 'axis' through the data to connect the concepts. While open coding 'opens' data to theoretical possibilities, axial coding ‘puts together’ concepts, and interrelates them to reach higher level of abstraction.

Selective coding

Analysis enters higher levels of abstraction in this step. The researcher works towards identifying 'the higher-order core category'. This means that from the categories identified so far, one has to be chosen which meets the demands of the tests. Here the analysis is directed towards a central focus. The process of analysis goes through paths that are similar to those employed in the previous steps. This means working through notes, diagrams and categories, searching for the central phenomenon and the central category.

4          ANALYTIC INDUCTION

In simple terms, analytic induction aims to produce complete and universal statements about social phenomena. Induction means proceeding from the specific to the general that is, constructing abstract concepts from a study of specific data. This is the opposite of deduction, which proceeds from the general to the specific, verifying abstract concepts by testing whether they are supported by data. In a way, analytic induction entails both induction and deduction. As an 'induction' it studies data obtained through qualitative research and attempts to establish regularities and commonalities that could support the formulation of hypotheses. As a 'deduction' it tests the validity' of these hypotheses by contrasting them to research data, and examines whether they are still applicable and valid. This is commonly referred to as the analytic spiral.

Using analytic induction ·         Define and describe the topic. ·         Examine raw data produced through the study. ·         Formulate a working hypothesis as to the nature of the characteristics which the researcher initially assumes are the most important. ·         Examine a specific case to establish whether the hypothesis regarding the assumed characteristics applies. ·         If the case confirms the hypothesis, continue the comparison by considering new data, until saturation is achieved. The hypothesis is then accepted, and its validity confirmed. ·         If the case does not confirm the hypothesis, either exclude the case and redefine the phenomenon, or reformulate the hypothesis, changing the originally hypothesized characteristics so that the case may become part of the phenomenon understudy. ·         Continue contrasting with new cases until analysis leads to a theory that would explain the phenomenon within a particular context

5          CONSTRUCTION AND DECONSTRUCTION

Construction and deconstruction are parts of an analytical mechanism that is most closely associated with interpretivism and the notion that the world is constructed.

·         Deconstruction: In order to achieve a deep understanding of phenomena, it is necessary to separate oneself from everyday occurrences, explanations and typicalities. One must go beyond the 'known'. In this sense, the capacity of interpretation must be widened, hasty identifications of phenomena avoided and new structuration of knowledge achieved. For this to happen, materials must be cut into the smallest pieces and de-contextualized, to the extent that their original context can no longer be recognized. In this way, texts are converted to small units of meanings, free from previous meaningful connections, to other units and to contexts, and free from overarching, general assumptions.

·         Comparative differentiation: The resulting de-contextualized parts are then integrated into new constructs, which are then compared to other possible alternative options. This is done so that the researcher can widen the understanding of the whole process of the object under investigation.

·         Contextual construction of meaning: This entails an analysis of the reconstructed materials within a material and a social context.

·         Extensive interpretation of meaning: The researcher speculates here as to various possible meanings, and analyses such possibilities as if they were factually effective. This is intended to allow the researcher to achieve possible information about the whole context of the phenomenon.

·         Testing of results: Testing of the outcomes of this process is continued further, until the validity of the resulting constructs is established.

This brief discussion demonstrates that analysis and interpretation in qualitative research are neither simple nor uniform. The diversity in this area is so important and so well founded that it is almost impossible to integrate it into one model, pattern or template. Unlike quantitative research, in qualitative research analysis and interpretation are pluralistic, as pluralistic as the methods used and (even more so) as pluralistic as the underlying paradigms.

6          DATA ANALYSIS IN NARRATIVE INTERVIEWS

In narrative interviews, analysis begins when narration, debate and transcription have been completed. In this sense, analysis means analysis of transcripts, content analysis or textual analysis. Focus of analysis is not on the interview itself but on the text that is produced while documenting the story presented during the interview.

For instance, attention is given here to the analysis of conversation (it is therefore referred to as conversation analysis, see, for instance, Bergmann. 1991), with particular emphasis on the structure of conversation, such as the fluctuations in story-telling, patterns of presentation of the story, the ranking of experiences, hesitation in the flow of narration, fluctuations in the degree of detail in story-telling, inconsistencies or contradictions in information content and so on. The presence of such gaps, inconsistencies and contradictions in the narrative are taken into consideration when drawing conclusions.

The process of analysis of the transcripts of narrative interviews can be broadly summarized in six points.:

·         Formal textual analysis: The first step of analysis includes cleaning the text of non-narrative material and preparing the text for analysis. This includes the identification of sequences in the text (that is, whether the interviewee sets some events as more important or of a higher priority, and others as less important or of lower priority). Analysis then follows within the context of the sequenced text.

·         Structural description of the content: Here emphasis is placed on the overall structure of the text and on its composition. To achieve this, the text is searched for indicators of connectors between individual presentations of events (such as 'because', 'then', 'after', 'already' etc.) as well as of deficient plausibility shown in the tone of the voice, structure of language, self-corrections and the like (Heinze, 1995: 72). The purpose of this is to demonstrate which parts of the statements have a limited and which a more general significance.

·         Analytic abstraction: At this point of analysis, the results allow a perception of the situation which is less bound to single statements or descriptions of single events and more to general and abstract expressions. Such abstract statements are contrasted with statements relating to specific life sequences, and their validity is tested. This is expected to lead to an identification of the basic and dominant experiential frequency that best describes the life events.

·         Knowledge analysis: Analytical abstraction prepares the scene for a more realistic and convincing interpretation of the life processes of the informant extending to levels of knowledge, and an analysis of the ways in which knowledge is employed to respond to social demands.

·         Comparisons: At this stage comparison between more text parts are undertaken to allow relative generalizations. This eventually leads to the construction of elementary categories.

·         Construction of a theoretical model: The final aim of this analysis is to construct a theoretical model whose elements are sequentially contrasted to statements presented in the text. Those consistent with the text are maintained; others are eliminated. This contrast of the whole with the parts and vice versa is critically important for the construction of the model.

The researcher is here interested not simply in detailed un-interpreted descriptions of social conditions, but in the story. Descriptions are reports of happenings as recalled by the teller. Stories are characterized by the fact that they not only bring the listener to the scene of the happening, including space and time, but entail also a sequential presentation of the event, and most of all a retrospective interpretation and a concluding summary of the story. Basically the NI focuses on a reconstruction of the orientation of patterns of action. The two most important characteristics (and advantages) of NIs are the retrospective interpretation and the closeness of detailed stories to reality.

The theoretical basis of the analytical interpretation employed in the NI is the homology of narrative and experience, namely that narratives and experiences are identical (a point that has been criticized by other researchers). This means that stories about the personal life of the interviewee truthfully reflect life experiences which contain information about basic structures and mechanisms of social life that are pertinent to other people. Following this, the NI is thought to have the capacity to focus not only on life stories of the interviewee but also on stories about collective experiences with social structures and processes, historical events community experiences and reactions and other community responses to social events of the time.

The process of analysis of transcripts of narrative interviews bears a strong resemblance to the general pattern of qualitative analysis introduced earlier in this chapter. The imminent contrast of abstractions to raw data and the constant validation, as presented in the analytic induction, is more than evident also in this analysis.

7          DATA PRESENTATION IN QUALITATIVE RESEARCH

7.1       Introduction

Qualitative researchers employ various methods to present their data visually. While some employ methods of presentation used by quantitative researchers, in their original or in a modified form, in other cases different forms of presentation are used. Certainly, tables and graphs are useful tools of presentation in qualitative research, but the structure of presentation does not seem to adhere to any strict rules and procedures, given that graphs and tables in qualitative research are always tailored to serve the need of the particular study.

The methods seem to be developed by researchers for a particular study to meet their personal styles, but: those dial: prove useful are taken up by other researchers and after sonic time they become an element of qualitative research. Miles and Huberman, for instance, presented some or the methods used in their studies in a widely-read publication (1994); many of these are very interesting and are used widely by other qualitative researchers. In the following sections we shall report on some of their techniques, as presented and Justified theoretically and methodologically by these writers, giving a few examples of the types of displays that they report.

7.2       Matrices in qualitative research

A matrix is a form of data presentation that, to a large extent, resembles and is equivalent to a table. It has a title, a heading, cells, and other forms of information similar to those of the tables typical of quantitative research, but it differs in its purpose and nature.

Matrices are a form of summary table. They contain verbal information, quotes, summarized text, extracts from notes, memos, standardized responses and, in general, data integrated around a point ur research theme that makes sense. In the main, matrices contain information about and explain aspects of research, and allow the researcher to get a quick overview of data related to a certain point. In this sense they serve a similar purpose to that of tables employed in quantitative research.

Matrices can become very complex and also serve many diverse goals. In one form (checklist matrix} [hey present integrated data on a summative index or scale, organizing in that way several components of a single, coherent variable (Miles and Huberman, 1994). In another form they contain information ordered by time {time-ordered matrix} or according to roles {role-ordered matrix). The former can be thought of as a table in which the columns are arranged in a time sequence, demonstrating what happened and when as the research progressed from one point of time to the next. In a role-ordered matrix, the Cable rows contain verbal information about the views of role occupants on a specific issue of the project. A combination of die elements of matrices is also possible (e.g. role by time or role by group matrix).

Matrices can be ordered according to a central theme (conceptually clustered matrices}, or according to outcomes and dependent variables (effects matrices), or present: forces that are at work in particular contexts showing processes and outcomes (site dynamics matrices); they can present series of events displayed in any possible order (event listing), or in the form of a causal network. In the Litter case, the matrix presents a field of interrelationships between dependent and independent variables describing causal connections between them.

It must be noted that models such as those described above, although of a qualitative nature, have some strong quantitative overtones, which might not be accepted by traditionally orientated qualitative researchers. Terms such as variables, for instance, defined as dependent and independent, especially when a notion of causation is attached to them, are elements that many researchers may find incompatible with qualitative analysis. This illustrates how diverse the field of qualitative research is.

Building matrices

Constructing a matrix is a process that relates more to the personal ingenuity competence and creativity of the researcher than to rules and principles Miles and Huberman (1994: 240-1) admit that there are no fixed canons for construction a matrix, and that: there are no 'correct' matrices, but rather functional ones. They nevertheless recommend that researchers follow some 'rules or thumb' when working with matrices. Such 'rules' suggest: that matrices should be kept to one-page displays; should include between 15 and 20 variables in rows and columns; be constructed while keeping in mind that they can and will be changed, regrouped and modified by adding new rows and columns; that rows and columns be kept fine-grained so that adequate differentiation is possible' and that new matrices may evolve out of other matrices as the research unfolds.

7.3       Figures in qualitative research

Figures are as useful in qualitative research as matrices. They combine lines and curves with verbal comments and indicators. However, there is no set format or organization and construction; how a figure will be constructed and what format it will take depend on the complexity of data and on the ability and imagination of the field worker (Miles and Huberman, 1994).

7.4       Charts in qualitative research

Miles and Huberman (1994) describe a form of chart that they call a context chart. This presents in graphic form the in re relations hip that exists between elements of the environment, for example roles and groups that make up the context in which behaviour develops. Context is a significant factor in die understanding of behaviour. Context charts offer a visual presentation of behaviour in context.

7.5       Displays in qualitative research

Displays are employed in qualitative research for many and diverse purposes. To a certain extent, displays substitute for the work accomplished by means of statistics. In this sense, chares, matrices and figures provide the information that mathematical figures and coefficients provide in quantitative research. They present visual information which allows the researcher to make sense of the collected information and so to draw relevant conclusions. There are several types of displays, each of which differ in type and complexity and serve a different purpose. Miles and Huberman (1994) offer a very detailed discussion of these types of displays; a brief list is given below.

Within case displays

The first group of displays comprises the Within-Case displays, which have the purpose of exploring and describing conclusions related to a single case study. The types of displays used in this context are: partially ordered displays including context charts, check-list matrices and transcript-as-poem displays); time ordered displays (such as event listing, critical incident charts, event-state networks and activity records); role-ordered displays (e.g. role-ordered matrices and role-by-time matrices); and conceptually ordered displays (such as conceptually clustered matrices, thematic conceptual matrices, fork taxonomies, cognitive maps, and effects matrices).

In the same group belong also types of displays which are intended not only to ascribe but also to provide explanations and predictions. Such displays include the case dynamic matrix, which displays 'a set of forces for change and traces the conceptual processes and outcomes' (Miles and Huberman, 1994: 148) and the causal network which displays the various variables and their interconnections.

Cross-case displays

The second group of displays involves more man one case. Some of these displays aim to explore and describe conclusions, others to order and explain findings. Some are partially ordered displays; others are case-ordered and time-ordered displays. Many of these displays are matrices, graphs and tables. But: other types of displays are also used; the scatter plot is an example. Here a relative affinity to quantitative research is evident. This is more so when we explore die second part of this group, which relates to ordering and explaining. Here a strong emphasis is placed on predicting and variable analysis as well as on causal analysis (causal chains and causal networks), which for many qualitative researchers do not belong to the qualitative methodology.

8          COMPUTER-AIDED DATA ANALYSIS (CADA)

Computers can be, and are being, used in qualitative research in die context of both pure qualitative research, where analysis is done the traditional way, and the so-called 'enriched' qualitative research (Conutrad and Reinharz, 1984; Fielding and Lee, 1998; Fisher, 1997; Richards, 1986; Richards and Richards, 1987, 1994, Weitzman and Miles, 1994). Computer-aided data analysis (CADA) is used in many forms and allows die qualitative researcher to process data in a way parallel to that in quantitative research (Huber, 1991; Madron, Tate and Brookshire, 1987; Ragin, 1987).

There are now more than 24 programs currently in used in qualitative research (Fielding and Lee. 1998; Weitzman and Miles, 1994), Apart from basic word-processing programs, which are used in low-level analysis of qualitative data, there are programs devised specifically for research purposes. Each program is specialized in a particular area of research, for example text retrievers that code and retrieve information, text managers that manage text, conceptual network builders that draw diagrams, or code-based theory building programs that test hypotheses, and develop theories,

Some of the most common programs are (see Flick et al., 1991; Kelle and Erzberger, 1999; Richards and Richards, 1987, 1994; Weitzman and Miles, 1993, 1994):

AQUAD AskSam ATLAS, ti HyperQua! HyperRESEARCH

HyperSoft Info Select: Kwalitan Martin MAX

MECA Metamorphi Nud*ist NVivo Qualpro

TextBase Alpha The Ethnograph The TextCollector The WordCruncher ZyINDEX

The extent to which CADA is used by qualitative researchers depends on the nature of the study, type of analysis and underlying paradigm. Briefly the types of functions offered by computer programs are shown in Box 15.6.

Uses of computers in data analysis Areas in which computers assist are: Recording/storing: Computers assist researchers with the recording and storing of data. Word-processing programs and specific programs for data analysis are employed for this purpose. Coding: Computers allow researchers to isolate segments of text and code them for future retrieval and linking with other segments of the text In addition, such programs allow researchers to make side notations on the margins. Retrieving/linking: Coded segments of text can be retrieved and set in separate files and/or linked with other parts of the text, facilitating within-part and between-part searches and linkages. Programs can also perform content analysis, count frequencies of codes and produce quantitative results. Displaying: Results of search and retrieval can be displayed by the computer in a variety of ways. Highlighting relevant sections one at a time is one way; gathering and displaying related segments of text together in 3 file Is another. Finally, computer programs can display results graphically. Integrating: More advanced computer programs have the capacity to develop classifications, categories, propositions and semantic networks and establish links between them to the extent that they can build and/or test theories. Analysis: Computers can analyze differences, similarities and relationships between text segments. Developing typologies and theories: Testing: Computers can text theoretical assumptions on the basis of qualitative data and the integration of qualitative and quantitative methods.

Although most computer programs employed for qualitative analysis contain modules for recording, coding and retrieving data, higher-level functions such as linking, displaying and integrating are available in a few programs only. ATLAS.ti, HyperRESEARCH. MECA, MetaDesign, Nud*ist and QCA are a few examples of programs with advanced capabilities. Overall, the strengths and weaknesses of CADA those listed below (Ragin and Becker, 1989; Tesch, 1989.1990),

Strengths of CADA

Saves time: Computers are faster than researchers in almost all analytic tasks. Even when Liking into account the time required to program the computer to complete tasks, computers are faster. Given also that most instructions are systematized and packaged, readymade for the user, computer advantage is more than obvious.

Save work: Many tasks are not integrated and a lot of work that was once necessary is no longer required. Relieving researchers from a large part of the menial work allows them more time to concentrate on more important aspects of analysis.

Produces quick results: Operations are completed in a fraction of the time required for assistants to complete them.

Is convenient: Researchers can use computers at any time and place (e.g. late at night, during holidays, in an airplane).

Reduces the need for personnel: Computers perform the work of several people, reducing the number of assistants required.

Offers a more efficient analysis: Computers make no mistakes in recognizing or counting codes.

Offers easier access to texts and codes: This is particularly the case when data is stored electronically.

Offers easier text reproduction and sharing: Copying parts of the text, identifying similar codes and integrating them into a common file, sending files to other researchers and similar tasks are conducted more easily with the aid of computers than manually.

Is accurate: It is more accurate than manual processing.

Is reliable: It is more reliable than manual processing.

Is flexible: It allows flexibility in the analysis.

Is powerful: It allows the processing of larger amounts of data than manual processing.

Weaknesses of CADA

Costs: CADA requires an infrastructure that adds to the costs of the project budget, for computers and programs.

Artificial treatment: It is often argued that the essence of data is not accessible to 'machines', regardless of flow intelligent the programs may be,

Theoretical inconsistency: Qualitative methods offer in essence a path away from structured thinking and operations, such as quantification and researcher distance from the researched. The use of computers in qualitative research works against this principle, and makes it no different from the models it is intended to overcome and avoid.

Inadequacy of programs: There is often no consistency between computer methodological principles and die assumptions of die theory that underlies the research. This is clearly shown by in the fact that one computer program can serve so many and diverse qualitative theories. The increase in the number of computer programs and their growing diversity may rectify this weakness.

Emphasis: The use of computers often displaces the weight of the analysis from theory development to coding, and to technical aspects of analysis.

On balance, CADA brings more advantages to researchers than manual processing. It is a matter of time and training to correct weaknesses, to make researchers aware of the dangers of computerized analysis, and to adjust computer programs so that they accurately meet the needs of the researchers and the academic community in general. Computers are an indispensable tool of research, and as technology advances will become even more useful, even indispensable, in the future.

MAIN POINTS

·         In most cases, qualitative analysis is conducted either during data collection, or during and after data collection.

·         Overall, qualitative analysis is not as fixed or as uniform as quantitative analysis.

·         There are also cases where qualitative analysis follows the straight line of a fixed design, where it begins after recording, transcription, checking and editing have been completed, followed by interpretation, and generalizations.

·         Most qualitative researchers do not employ mathematical/statistical methods in data analysis.

·         Qualitative researchers use computers extensively in data analysis. Several computer programs have been developed to aid with qualitative data analysis. They are used for recording and storing, coding, retrieving and linking data, and displaying and integrating data.

·         Overall, data analysis in qualitative research is not as clear or as uniform as it is in quantitative research.

·         In qualitative interviews one possible way of data analysis entails transcription, checking and editing, analysis and interpretation, generalizations and verifications.

·         Qualitative analysis is more advanced and more complex when it entails analytic induction, construction and deconstruction, and grounded theory.

·         In case-study research, data analysis can be accomplished by one or more of the following methods: pattern-matching, explanation-building techniques, time-series analysis, making repeated observations and secondary analysis.

·         Computers are used in qualitative research for recording and storing, coding, retrieving and linking data, and displaying and integrating data.

·         Graphs are common in qualitative research. Examples are matrices, figures and charts.