| 
  • If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • Stop wasting time looking for files and revisions. Connect your Gmail, DriveDropbox, and Slack accounts and in less than 2 minutes, Dokkio will automatically organize all your file attachments. Learn more and claim your free account.

View
 

Research Method - Phenomenography

Page history last edited by Roger Hadgraft 11 years, 5 months ago

CRITICAL FEATURES OF PHENOMENOGRAPHY

 

Llewellyn Mann, CQUniversity

l.mann@cqu.edu.au 

 

History

 

Phenomenography is the empirical study of the qualitatively different ways in which aspects of the world are experienced. That is, it involves mapping phenomena, or the relations between persons and aspects of their world (Marton 1994). Historically, phenomenography has been used to research the experience of learning, the experience of teaching, the different ways of experiencing the content learned, and describing aspects of the world around us (Bowden 2000). First used in the ground-breaking work of the Swedish researchers Marton (1981, 1981, 1976), Säljö (1981, 1988), Svensson (1983) and Dahlgren (1984) in the mid-70s, phenomenography was initially developed to investigate learning among students and lead to identifying the ‘surface’ and ‘deep’ approaches to learning (Marton and Säljö 1976). While it appeared in its own right as an empirically based research approach for describing people’s experiences during the early 1980’s (see for example Marton (1981, 1986)), it is important to note that only more recently has the underpinning theory been elaborated (Marton and Tsui 2004, Pang 2003). The research described in this paper used developmental phenomenography (Bowden and Green 2005), in that it examines sustainable design practitioners’ ways of experiencing sustainable design, in order to better inform current practice, as well as to help educate future engineering students about sustainable design.

 

 

The Object of Study

 

The object of study in phenomenography is the variations in the ways that an aspect of the world has been experienced by a group of people (Marton and Booth 1997). It is about describing the world as experienced, and revealing and describing the variation that exists (Bowden 2005). Figure 1 illustrates this focus of phenomenography, not on specific aspects of the world or the subjects in isolation, but on the relationships between them. Hence, phenomenography takes the position that experience is relational, not purely objective, independent of people, nor purely subjective, independent of the world. Knowledge is then created from the relations between persons and in relation to the world. As Marton & Booth explain, with reference to a learner (1997)(p13): “There is not a real world ‘out there’ and a subjective world ‘in here’. The world [as experienced] is not constructed by the learner, nor is it imposed upon her; it is constituted as an internal relation between them. There is only one world, but it is a world that we experience”.

 

 

Figure 1: Focus of Phenomenographic Research (Based on Bowden 2005)

 

The focus on the world as experienced gives phenomenography a non-dualist ontology. It takes neither a positivist/objective approach, independent of human interpretation, nor does it take a subjectivist approach, focusing on internal constructions by the subject (Marton and Booth 1997, Trigwell 2000). We are not interested in what people think per se, but instead what their experiences are and have been in situations where they have had to deal with aspects of the world. What people think may be clouded by rhetoric that they have been told or read, whereas their experiences reveal more about their understandings of the aspect of the world of interest. A non-dualist ontology also has implications for the relationship between the researcher and the aspect of the world under investigation, as depicted in Figure 1. This relationship is important as it allows the researcher to carry out the research, as some understanding of the research topic is needed to interpret the statements made, and to keep the research focused. However, any preconceptions or theories about the aspect of the world under consideration that the researcher has from their own experiences must be held at bay during the research (Sandberg 1997). This allows the researcher to be open to other ways of experiencing the particular aspect of the world under study, and able to present these other experiences as genuinely as possible.

 

One of the strengths of phenomenography is that it “provides a way of looking at collective human experience of phenomena holistically despite the fact that such phenomena may be perceived differently by different people and under different circumstances” (Åkerlind 2005)(p72). Trigwell (2000) provides an overview of how phenomenography is distinguished from other research approaches (p77) “The key aspects of a phenomenographic research approach … are that it takes a relational (or non-dualist) qualitative, second-order perspective, that it aims to describe the key aspects of the variation of the experience of a phenomenon rather than the richness of individual experiences, and that it yields a limited number of internally related, hierarchical categories of description of the variation.”

 

These aspects and their points of departure from other research approaches can be seen in

Figure 2. As shown from the series of bifurcations indicated by the numbers 1-5, phenomenography is: (1) non-dualist in that reality is seen as constituted from the relations between the individual subjects and an aspect of the world (Trigwell 2000); (2) qualitative in a grounded way as it tries to explore and describe a phenomenon from the data, rather than trying to fit the data to predetermined categories; (3) a second order approach, as it is the experiences of others rather than the researcher, that are the base of the investigation (Trigwell 2000); (4) focused on the variation in the ways an aspect of the world has been experienced; and (5) able to produce results in a set of categories that are internally related. The focus on qualitatively describing the variations and relationships between categories of description is one of the major differences between phenomenography and other research approaches (Bowden et al. 1992).

 

 

 

Figure 2: Points of Departure Between Phenomenography and Other Research Approaches (Trigwell 2000)

 

Outcomes of Phenomenography

 

The major outcomes of a phenomenographic study are the description and organization of the variations in ‘ways of experiencing’ an aspect of the world into ‘categories of description’, and the describing the logical relationships between them, often in the form of a hierarchy from less comprehensive to more comprehensive, referred to as an ‘outcome space’ (Åkerlind 2002). The categories of description are an attempt to clarify the different ways the same aspect of the world has been experienced by a group of people who are all highly confident that their interpretation is the most reasonable (Åkerlind 2005). The hierarchical relationships between the categories are not value judgments from ‘better’ to ‘worse’ (Åkerlind et al. 2005). However some categories of description are inclusive of other categories and, as such, the structural relationships in a phenomenographic outcome space are those of hierarchical inclusiveness. This also leads to the structure not necessarily being linear, but instead may contain forks or branches. However, the categories of description developed can never form an exhaustive system for the aspect of the world, but they should be complete for the experiences of the group of participants under consideration at a particular point in time (Åkerlind 2002, Marton and Booth 1997).

 

The following three criteria for judging the quality of the categories of description developed in a phenomenographic study are put forward by Marton and Booth (1997):

 

1. The individual categories should each stand in clear relation to the aspect of the world under investigation so that each category tells us something distinct about a particular way of experiencing the aspect of the world;

2. The categories have to stand in a logical relationship with one another, a relationship that is frequently hierarchical;

3. The system should be parsimonious, which is to say that as few categories should be explicated as is feasible and reasonable, for capturing the critical variation in the data.

 

Marton and Booth (1997)(p114) argue that the final categories of description and the outcome space they create is a depiction of variation on a collective level, and as such, “individual voices are not heard. Moreover, it is a stripped description in which the structure and essential meaning of the … [categories] are retained while the specific flavors, the scents, and the colors of the worlds of the individuals have been abandoned”. The categories are thus not necessarily ones that any one person in ‘real life’ would identify with; they are constructions that incorporate key variations of discussions with a specific number of people (Cherry 2005).

 

Data Collection in Phenomenography

 

Phenomenographic studies need to have a coherent purpose (Bowden 2000) and method throughout, from the initial planning stages through the collection of the data, to analysis and the presentation of the results. The research subjects are identified in the planning stage of the research due to their relationship with the specific aspect of the world under consideration, in this case sustainable design. They should also be selected to obtain as much variation in their experiences as possible, but still within the purpose of the study. Phenomenographic data collection usually revolves around interviews (Green 2005), which have an open ended format with interviewees responding to an initial question or problem. “The researcher and researched must begin with some kind of (superficially) shared topic, verbalized in terms which they both recognize as meaningful” (Ashworth and Lucas 2000)(p299). The purpose of the phenomenographic interviews is to reveal interviewees’ experiences with the aspect of the world under consideration. As such, interviewees are encouraged throughout the interview to reflect on and reveal their way of experiencing the aspect of the world in context. What is important is what the interviewees think these experiences reveal about the aspect of the world itself, and follow up questions in the interview should focus on eliciting this meaning (Åkerlind 2005). As Marton (1994)(p4427) argues “The interview has to be carried out as a dialogue, it should facilitate the thematisation of aspects of the subject’s experience not previously thematised. The experiences… are jointly constituted by interviewer and interviewee.”

 

Once the initial question or problem has been proposed, follow up questions ask interviewees to elaborate on their experiences and what they mean by certain concepts. All follow up questions are extracted from what the interviewee has said so far in the interview, and not formed through predetermined ideas and questions from the interviewer. As such, different interviews “may follow somewhat different courses” (Marton 1986)(p42). In this way, the interview is a dialogue or conversation, encouraging the interviewees to reflect on their experiences of the aspect of the world. Judgmental comments from the interviewer should never be made in the interview (Bowden 2005).

 

An important aspect of the phenomenographic interview is the use of empathy to further engage with subjects’ life-worlds (Ashworth and Lucas 2000). As the categories of description are derived from subjects’ experiences relayed in the interview, it is “a paramount requirement for phenomenography to be sensitive to the individuality of conceptions of the world” (Ashworth and Lucas 2000)(p297). This is achieved through the process of ‘bracketing’ the interviewer’s own assumptions and theories, and instead being empathetic to the subjects’ experiences of the aspect of the world under consideration. Three of the presuppositions that need to be bracketed which have been identified by Ashworth & Lucas (2000) are: (i) importing earlier research findings; (ii) assuming pre-given theoretical structures or particular interpretations; and (iii) imposing the investigator’s personal knowledge and belief. While interviewers can’t detach themselves from their own life-world, as that is what enables the researcher to interpret what is said, they do need to bracket their own theories and preconceptions, and focus on the experiences of the participant. Ashworth & Lucas (2000)(p299, their emphasis) use the following to illustrate this: “For instance, views and factual claims which the student expresses in an interview may well be regarded by the researcher as quite erroneous. The temptation would be to marginalise such material. But the researcher who adopts an attitude of empathy with the student should find such views and factual claims of immense interest”.

 

Another important aspect of the data collection process is conducting pilot interviews to enhance phenomenographic interviewing skills (Bowden 2005), and to test if the initial questions reveal the sorts of experiences (data) necessary to address the focus of the research (Green 2005). It is important that the pilot interviews are with people within the target group to obtain practice investigating the sorts of experiences that could be encountered in the final study. It is also important that they are discarded and not included in the final study (Bowden 2005), as the interviews may contain potential errors that might invalidate the results. Also, it is often the case that the follow-up prompts are more useful in eliciting meaning than the initial planned questions (Åkerlind 2005). As these follow-up questions have to be devised ‘on the fly’ based upon what the subjects say in the interviews, it is vital to practice identifying and asking this type of question during the pilot interviews.

 

Data Analysis in Phenomenography

 

While there is some variation in the methods used to analyze data in phenomenography, the overview presented here includes some of the generally accepted procedures and forms the basis of the approach used in this study. For a more detailed description of the commonalities and variations in the phenomenographic method of data analysis, see Åkerlind (2002, 2005). Interviews transcribed verbatim become the focus of the phenomenographic analysis when interviews have been used as the primary means of data collection. The set of transcripts represent a ‘snapshot’ of some of the experiences of a group of people with a particular aspect of the world in response to a particular set of questions at a particular time (Åkerlind et al. 2005). When data collection has relied only on interviews, no other evidence exists beyond the transcripts to inform the analysis process (Bowden 2005).

 

The analysis process is both one of ‘discovery’ (Hasselgren and Beach 1997) as well as one of ‘construction’ (Bruce 2002). The results are not known in advance and tested in the study, but must be discovered, or emerge from transcripts, and constructed in an iterative way from the transcripts. In this way, phenomenographic analysis is a ‘bottom up’, inductive way of working from the data to the results, rather than a ‘top down’ way of constructing then testing an hypothesis (Green 2005). It is important to keep an open mind during the analysis process (Åkerlind 2002), as the categories of description may change several times and the researcher cannot be closed off to already determined categories. To achieve this, a constant focus must be maintained on the transcripts as the only source of evidence. Further, the researcher needs to focus on the transcriptions and categories as a whole set, rather than on individual transcripts or categories in isolation (Green 2005). The researcher must be open to the fact that different people may see the same aspect of the world in different ways, which is an underpinning of phenomenography, but is counter-intuitive to our natural attitude (Bowden 2005, Marton and Booth 1997).

 

Some phenomenographers emphasize not analyzing the structural relationships between the categories until the categories themselves are finalized, as it may introduce the researcher’s relationship with the phenomenon into the categories (Ashworth and Lucas 2000, Bowden 2005). Others argue that focusing on the structure of the categories and outcome space too late could lead to the meaning and structure not being adequately co-constituted in the final outcome space (Åkerlind 2005), and that a strong emphasis on looking for structure in the phenomenographic analysis process is vital, as the focus on structure: is an epistemological underpinning of phenomenography; increases the potential for practical applications from the research; and provides a simultaneous focus on variation and commonality.

 

The analysis process involves identifying meaning or variation in meaning across the set of transcripts. As it focuses on describing qualitative similarities and differences across the transcripts, phenomenographic outcomes do not show the richness of the data, only variation for which there is clear evidence from the transcripts (Bowden 2005). This focus on facets that are critical in distinguishing the variation between categories of description allows the structural relationships to be highlighted to a degree that would not be possible if “the analysis focused on every nuance of meaning” (Åkerlind 2005)(p72). The analysis process starts by the researcher reading and re-reading all the transcripts as a full set of data (Green 2005). The researcher then tries to articulate the aspect of the world for each transcript. Transcripts with similar individual meanings are then grouped, with the similarities within and differences between the groups clarified. A description of each category is written with illustrative quotations from the transcripts. These descriptions form the preliminary categories for the set of transcripts. It needs to be understood that this first attempt will not necessarily be ‘right’ and will most likely change. It will, however, provide a different way to see the data, to then revisit and further develop the categories (Green 2005).

 

From the initial groups, the researcher identifies transcripts that do not seem to fit into any category, as such transcripts often show a different facet that needs to be considered. The descriptions of the categories are clarified with constant reference back to transcripts as wholes. During this process, the researcher must constantly be asking, ‘Is there another way of interpreting this statement?’ It is also important to constantly refer to the initial focus of the study, as it is easy to become distracted by particular aspects of the transcripts (Bowden 2005). In writing the descriptions of the categories, researchers can only rely on what is included in the transcripts, and cannot extend or speculate on this. The researcher can accomplish this by constantly asking, ‘Where in the transcripts does this come from?’ almost becoming their own devil’s advocate. The final descriptions of the categories should be self-contained, in that they are able to be understood as a set of separate, stand alone statements.

 

At the end of the analysis process, all of the transcripts are sorted into individual categories of description. The categories themselves should have clearly defined statements of what they are, backed up with illustrative quotations from the transcripts. Pictorial representations may also help to explain the categories. A label for each category of description can also be developed, but this labeling should be avoided until late in analysis, as it may limit further category development (Bowden 2005). The relationships between the categories of description should also be detailed, using illustrative quotations where appropriate. These relationships should specify the similarities and differences between the categories and help to reveal categories that are more comprehensive than others. The categories are then sorted into a hierarchy based on their increasing comprehensiveness. This hierarchical representation of the categories of description is known as an outcome space (Åkerlind et al. 2005).

 

Validity of Phenomenographic Results

 

Two types of validity are often used, those of communicative validity, and to some extent, pragmatic validity. Sandberg (1994) proposes three phases in the phenomenographic process where communicative validity is relevant: (i) within the interviews communicating with the subjects; (ii) in the analysis process communicating with the text; and (ii) in communicating the results to other researchers and professionals.

 

For the first phase of communicative validity within the interviews, subjects are informed prior to the interview that the researchers were interested in their experiences of the aspect of the world under investigation. They are also informed that there were no right or wrong answers, and that no personal judgments would be made about what was discussed. This is to start to develop a joint understanding between the subjects and the researchers about what is discussed in the interview (Sandberg 1994). The other aspect of communicative validity in this phase is establishing a dialogue within the interview, rather than the interview becoming a question and answer session. This is often achieved by having a specific interview protocol with a set of open ended initial and follow up questions to stimulate discussion, rather than asking questions of a closed nature. Also, within the interviews, the subjects are constantly asked to qualify the statements they made as a way to stimulate further discussion.

 

The second phase, during the analysis process, involves focusing on each transcript as a whole, rather than trying to extract parts of a transcript and to analyze these out of context. This focus is maintained by looking at the similarities and differences between whole transcripts, especially where a particular statement taken out of the transcript may appear to fit into one category, but when seen within the whole transcript fits into another category.

 

The third phase of communicative validity involves obtaining feedback from other researchers and professionals. The results are often communicated and developed with other researchers in the field.

 

Reliability of Phenomenographic Results

 

Reliability in phenomenographic studies often revolves around the researchers’ interpretive awareness, or how interpretations have been controlled and checked throughout the research process (Sandberg 1997, Sandberg 2005). The stages where this is done are:

 

1.         The formulation of the research questions;

2.         The selection of the subjects;

3.         Interviewing those subjects;

4.         Analysing the resultant transcripts;

5.         Reporting the final categories of description.

 

The research questions for a phenomenographic study are aimed at eliciting the variations that existed among participants’ experiences of the aspect of the world being investigated. The questions are formulated to elicit experiences, with a focus on exploring variation in these experiences, rather than trying to test or impose a preconceived theory.

 

In selecting the subjects, a set of specific criteria are often used to ensure variation in the experiences of the subjects. These ‘diversity’ criteria are developed from the literature and ensured that participants had diverse backgrounds and hopefully experiences. These could include gender, age bracket, years of experience, disciplinary background or the type of experience.

 

During the interview process, the researchers’ interpretations are controlled and checked in a number of ways. Before formal interviews begin, pilot interviews are usually conducted to develop interviewing skills as well as the interview protocol. The interview protocol was also developed so as not to ask leading questions, or questions that suggested a particular way of experiencing to the interview subjects. The interview protocol also aided in making the interviews as consistent as possible. Each interview starts with the same information and introductory questions, and ends with the same concluding questions. During the interviews, an open-ended but focused interviewing technique is used. This allows the subjects to focus on the aspects of the aspect of the world they believed were important rather than ‘fitting in’ to any preconceived theories.

 

In the analysis process, the main control of the researchers’ interpretations is a strict adherence to data, usually in the form of interview transcripts. This involves constantly going back to the data as a whole, and reading statements in context. It also involves admitting to inconsistencies between transcripts during the analysis process, rather than trying to constrain the data to appear consistent. The categories are developed in an iterative fashion, in which the inconsistent transcripts act as prompts for a different way of viewing the categories of description.

 

The final results are in the form of a set of categories of description that form a hierarchical outcome space. The descriptions of the categories are based on the transcripts, and include illustrative quotes taken from some of the transcripts to further check the researchers’ interpretations.

 

Generalisability of Phenomenographic Results

 

A conventional notion of generalisability is not applicable to phenomenographic research, as it examines the variations of the experiences of an aspect of the world for a specific group of people. The group of people is chosen from a population to maximize the variation of experiences, rather than trying to be representative of the population (Åkerlind 2002). A different sample group in a different context may provide different categories of description, just as different researchers may develop different categories from the same data. As Åkerlind (2002)(p12) notes, “Consequently, phenomenographic research outcomes have been described as not enabling generalization from the sample group to the population represented by the group, because the sample is not representative of the population in the usual sense of the term.

 

It is expected however that the range of variation in the sample reflects the range of variation in the population (Francis 1996, Marton and Booth 1997). As such, the results of a phenomenographic study are generalisable to a group with similar characteristics and experiences to the sample group. Further, the range of variation should still be relevant to groups with less in common with the sample group, though it is likely to be a less complete representation of the range of experiences in the larger group. While the range of variation may be generalisable, the distribution of people among the different categories may not be (Åkerlind 2002). As such, it is important in any phenomenographic study to specify the characteristics of the subjects included in the study so that readers are able to make up their own minds about the generalisability to the group in which they are interested (Cope 2002).

 

 

Comments (0)

You don't have permission to comment on this page.