Grounded Theory

Outline

Grounded Theory, despite its name, as described here is essentially a method for eliciting representing knowledge. The term 'grounded theory' derives from the fact that the data and representation is taken to be based in a theory that is 'grounded', ie, is faithful to the local circumstances and the local actors interpretations of those circumstances. Thus, the important distinction to make is that the theory does not come from outside, but arises from the data itself.

This type of inquiry is common to ethnographic and qualitative research methods in the social sciences, but is also a tested method for information systems analysis.

Qualitative Inquiry

1. Phenomena are examined in a natural setting.
2. Data are collected by multiple means.
3. The research is directed at the exploration, classification and hypothesis development stages of the knowledge building process.
4. No experimental controls or manipulation are involved.
5. The investigator does not specify the set of independent and dependent variables in advance.
6. The results derived depend heavily on the integrative powers of the investigator.
7. Changes in site selection and data collection methods could take place as the investigator develops new hypotheses.
8. The research is addressing "why" and "how" questions rather than frequency or incidence.

Grounded Theory

Grounded Theory emerged from the work of Glaser and Strauss who challenged:

1. the arbitrary division of theory and research;
2. the prevailing view of qualitative research as primarily a precursor to more 'rigorous' quantitative methods by claiming the legitimacy of qualitative work in its own right;
3. the belief that qualitative methods were impressionistic and unsystematic;
4. the separation of data collection and analysis phases of research; and
5. the assumption that qualitative research only produced descriptive case-studies rather than theory development.

Grounded Theory

• A social science methodology for the systematic generation of conceptual models from qualitative data sources such as ethnographs.
• Essentially 'bottom-up' approach to the conceptual analysis of unstructured or semi-structured qualitative data.
• Places great stress on the detailed examination and cataloguing of qualitative data as the first stage of developing rich conceptual models that accurately describe, and hence are 'groundedŐ in data. By emphasizing that theoretical models should emerge from and be firmly grounded in data, the approach stresses the critical importance of the model building phase of scientific method.
• A particular advantage of the grounded theory approach is in uncovering potential conflicts between different frames of reference in use by actors. The product, ie, the grounded theory, comprises the set of conceptual categories and their properties, presented with the relationships among them. Conceptual descriptions of the data are systematically worked out throughout the research process producing three levels of generated conceptualizations:
  1. Properties of the data, usually organized hierarchically into various categories and subcategories;
  2. Substantive theory made up of more abstract descriptions of the data as it relates to specific domains;
  3. Formal theory which integrates the theoretical descriptions across fields.
• There are basically 3 complementary conditions that indicate completion:
  1. The analytic framework forms a systematic substantive or formal theory;
  2. The written account is a reasonably accurate statement of the matter;
  3. The findings are presented in a form that the target audiences find useful.

Grounded Theory Methods

• a logically consistent set of data collection and analytic procedures aimed to develop theory
• consist of a set of inductive strategies for analysing data
• start with individual cases, incidents or experiences and develop progressively more abstract conceptual categories to synthesize, explain and understand data and identify patterned relationships within it
• distinctive characteristics of the method include:
  1. simultaneous involvement in data collection and analysis phases of research;
  2. creation of analytic codes and categories developed from data, not from preconceived hypotheses;
  3. the development of middle-range theories to explain behaviour and processes;
  4. memo-making, ie, writing analytic notes to explicate and fill out categories, the crucial intermediate step between coding data and writing descriptions;
  5. theoretical sampling, ie, sampling for theory construction to check and refine the analyst's emerging conceptual categories; and
  6. delay of literature survey or use of (imported) theoretical models and analysis.
• Fundamentally, grounded theory methods explicitly unite the research process with theoretical development.

Collecting data

Generating data

Analysts' prevailing understanding of field should be a point of departure only, and not dominate the inquiry.

Aim for full, thick written descriptions from:

1. participant observers' field notes
2. interviewers' transcriptions
3. autobiographies, journals, diaries, etc
4. records and log books

Aim to generate data by investigating aspects that the person takes for granted;
Get as much material down on paper as possible.

With rich data, grounded theorists can more readily discern what participants mean and how they define their experiences.

Making meanings explicit

Emphasis on what people are doing leads to understanding multiple layers of meanings of actions including the personŐs:

• stated explanation of his or her action;
• unstated assumptions about it;
• intentions for engaging in it, as well as
• its effects on others; and
• consequences for further individual action and interpersonal relations.

Study the emerging data.

Coding the data

The first major analytic phase of the research - means creating the codes as you study the data. This is the pivotal step between collecting data and developing an emergent theory to explain the data.

To begin, this is usually done line-by-line and defining the actions or events that are represented.

Ask questions of the data:

1. What is going on?

   What process is at issue here?

2. What are people doing?

   Under which conditions does this process develop?

3. What is the person saying?

   How does this person act and think while involved in this process?

4. What actions and statements can be taken for granted?

   When, why and how does the process change?

5. How do structure and context serve to support, maintain, impede or change these actions and statements?

   What are the consequences of the process?

Line-by-line coding helps define what kinds of data to pursue next.

Focused coding refers to taking earlier codes that continually reappear in initial coding and using those codes to sift through large amounts of data. Focused codes should still be kept close to the data.

Memo-writing

The intermediate step between coding and the first draft of completed analysis.

Begin as soon as there are interesting ideas and categories to pursue.

Memo-writing consists of taking categories apart by breaking them into components. Define categories as carefully as possible by identifying properties or characteristics, looking for underlying assumptions and showing how and when the category develops and changes.

Much memo-writing should be directed to making comparisons (constant comparative methods). Hence, compare one respondent's beliefs, stance, experience and actions with another's.

Through memo-writing you will clarify which categories are major and which are minor.

Theoretical sampling

Theoretical sampling is collecting more data to clarify ideas and plan how to fit them together.

This helps to fill out categories, discover variation within them and define gaps between them.

Using comparison define the properties of categories and specify the conditions under which they are llinked to other categories. In this way, the categories are raised to concepts in the emerging theory.

When this conceptual stage is complete, then comparisons with the prevailing literature and other sources can be made.

The Grounded Theory

Having sifted the mass of qualitative data, and found some patterns through a process of inductive reasoning we can start to produce some theories to explain the patterns we are finding. These probably will not be fundamental theories of human behaviour, but little local theories that serve to explain the behaviour of our data. This local theory is what Glasser and Strauss can 'grounded theory' and will "fit the situation being researched, and work when put to use. By 'fit' we mean that the categories must be readily (but not forcibly) applicable to and indicated by the data under study; by 'work' we mean that they must be relevant to and able to explain the behaviour under study." (Glasser and Strauss 1967).

Reason (1981) quotes Diesing's book 'Patterns of discovery in the social sciences': "The information that is discovered in the field situation is used by the holist to build a model which serves both to describe and to explain the system. The model is built by "connecting themes in a network or pattern"; the connections may be of various kinds but they are 'dicovered empirically rather than inferred logically'; the result of this is an empirical account of the whole system. This account explains the system because it describes the relations the various parts have for each other, so that 'the relations between that part and other parts serve to explain or interpret the meaning of that part'. This type of explanation is called a pattern model of explanation......"

Grounded theory requires that data collection and analysis be constantly compared and contrasted thoughout the data collection and analysis process. Evolving theory directs attention to previously established important dimensions while the actual data simultaneously focus attention on the theory's suitability as a frame for the most recent data being collected. The result of this fluid movement between theory and data is a reconceptualization, often based on a creative leap, that should account for and encompass all the nuances of the data.

Strengths and Weaknesses

Strengths

• identifies the situated nature of some expert knowledge, as well as the contingent nature of practice, with the need to deploy qualitative and ethnographic approaches.
• Produces a 'rich' or 'thick' description that properly acknowledges areas of conflict and contradiction.
• More likely to determine what actually happens rather than some official or folk version of practice.

Weaknesses

• The large documentary overhead.
• Difficult to manage.
• Pragmatic issues about action research, such as access, confidentiality etc.
• High level of social skills required from the investigator.
• Investigors need to be skilled in the method.
• The product may not be as well structured as with other knowledge elicitation methods.
• Not based in science (may be seen as a merit by some).

Tutorial

Rob Stephens