Phenomenography in the “Lived” Context of Parental Learning

Abstract

Phenomenography is a qualitative research methodology that is often adopted by researchers to investigate people’s lived experience of the phenomena around them. Within the phenomenographic research tradition, there has been much discussion of the influence of the context on the phenomenon under study. Time and again, both phenomenographers and critics of phenomenography have stressed the importance of the researcher being mindful of the context when using phenomenography as a research methodology. In this paper, two phenomenographic studies of the learning of adults under the context of parental learning are reported. We attempt to illuminate empirically the consequences of ignoring the cautionary advice about context in the pilot study and contrast its outcomes with the quality of the research results of the subsequent main study that heeded the advice. It is important to clearly delimit the phenomenon in question by taking careful consideration of the relevant context, as this ensures that phenomenography is conducted on the same, target phenomenon, rather than on different phenomena.

Phenomenography as an Empirical Research Paradigm

Phenomenography is an empirical research paradigm developed by Ference Marton, Lars Owe Dahlgren, Roger Säljö and Lennart Svensson during the early 1970s that aims at describing the qualitatively different ways in which people experience the same phenomenon (Pang & Ki, 2016). Etymologically, the word “phenomenography” derives from the Greek words “phainemenon” (appearance) and “graphein” (description). Phenomenography is concerned about the description of how people experience the world and how the world appears to people (Pang, 2003). The object of research is the different meanings people bring to the specific phenomenon in question at a collective level (Marton, 2015). The main outcome of phenomenographic research is an “outcome space,” which describes the variation in the possible ways in which a specific phenomenon is experienced on a collective basis. It comprises a system of categories of description of the phenomenon in question, with each category denoting a distinctively different way of experiencing the same phenomenon (Åkerlind, 2018; Pang & Ki, 2016).

Phenomenography holds a non-dualist stance of human consciousness in which the experiencer and the world are considered inseparable. “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” (Marton & Booth, 1997, p. 13). The world will not be the same world without the experiencer, and reciprocally the experiencer will not be the same person if the world that he/she is experiencing is not there. The experience of a person about the world comprises functional entities describing both the experiencer and the phenomenon to be experienced. As argued by Marton (2000), “Our world is a world which is always understood in one way or in another, it cannot be defined without someone defining it. On the other hand, we cannot be without our world. Still, we can focus on the object or on the subject aspect of the subject-object relations that experiences are. When focusing on the former, we concluded that an object is the structured complex of all the different ways in which it can be experienced. When focusing on the latter, we concluded that we are always aware of everything, although the way in which we are aware of everything is situationally variable” (p. 115).

Instead of taking on a first-order perspective, which examines the essence and fundamental nature of the phenomenon in question, phenomenography adopts a second-order perspective, which investigates how the phenomenon in question is apprehended by people (Pang, 2003). According to Marton (1981), “from the first-order perspective we aim at describing various aspects of the world and from the second-order perspective we aim at describing people’s experience of various aspects of the world” (p. 177).

The epistemological stance of phenomenography, as espoused by Marton (1992), is premised on the principle of intentionality advocated by Brentano that “no hearing without something heard, no believing without something believed, no hoping without something hoped, no striving without something striven for, no joy without something we are joyous about, etc” (Spiegelberg, 1982, p. 37). Any act of consciousness needs to be intentional, and all that is psychological needs to be directed to something. Phenomenography conceptualizes human experience as man-world relationship, which is directed to the phenomenon in question. People cannot experience “something” without “something” being experienced; human experience necessitates the existence of something to be experienced (Pang, 2003).

Based on the findings of a huge number of empirical studies conducted, phenomenography asserts that people’s qualitatively different ways of experiencing a particular phenomenon are limited in number. This assertion is premised upon the notion that for a phenomenon to be seen as such, it must contain a limited number of discernible aspects, and the discernment of these aspects would give rise to a particular way of experiencing. If this condition does not stand, it implies that different persons would experience a particular phenomenon in different ways at different times. People would then not be able to identify that particular phenomenon and therefore not able to communicate with each other effectively as every one of us would have been living in a different world (Marton & Booth, 1997).

Phenomenography is a qualitative research methodology that is often adopted by researchers to investigate people’s lived experience of the phenomena around them. Marton (1981, 1992) proposed phenomenography not as a superior research methodology, but as a complementary approach to research on learning. He argued that studying learners’ perceptions, conceptions, or understandings of a phenomenon is ultimately beneficial to educators or teachers in designing programs to facilitate learning about that phenomenon. Phenomenography has since been adapted as a research methodology for a wide variety of subjects and for learners of various ages and conditions (Tight, 2016).

Two studies have provided thorough descriptions of the use of phenomenography as a research methodology and suggested ways to address some issues surrounding the quality of the application of this method. Sin (2010) tried to address the quality issues in phenomenographic research in three ways. First, she discussed the criteria for rigor and quality in qualitative research. Second, she drew on the literature to examine issues related to the conceptual underpinnings of phenomenography when these criteria are applied to phenomenographic studies. Finally, she analyzed the phenomenographic process and made suggestions for enhancing quality at each stage, with the hope that applying those suggestions could mitigate the issues surrounding the quality of phenomenographic studies. Åkerlind (2012) further argued that the primary source of the questions that have been raised about the quality of phenomenographic research is a lack of understanding among critics of the data analysis stage. She therefore focused on this stage of phenomenographic research, attempting to clarify the process through an analysis of the commonalities and variations in accepted practices.

Qualitative research is a set of complex interpretive activities, which does not privilege one single methodological practice over another. Denzin and Lincoln (2018) highlighted that “there are multiple qualitative research communities, each having its own criteria for evaluating an interpretation” (p. 22). Also, they argued that “each interpretative paradigm makes particular demands on the researcher, including the questions that are asked and the interpretations that are brought to them” (p. 19). As posited by Åkerlind (2012), while phenomenography has much in common with the underlying assumptions of other qualitative research paradigms, it has its own set of practices to enhance the rigor of research. Unlike the constructivist paradigm which adopts the evaluative criteria of credibility, transferability, dependability and confirmability to replace the positivist criteria of internal validity, external validity, reliability and objectivity (Lincoln et al., 2018), the notions of validity and reliability are fundamentally reframed within the context of phenomenography to evaluate the quality of phenomenographic research.

Kvale (1996) proposes two types of validity checks, namely communicative and pragmatic validity, which have been adopted as one of the major ways to evaluate the rigor of phenomenographic research. For communicative validity checks, the phenomenographic researcher needs to demonstrate how she can convince the relevant research community, other members of the population represented by the sample of participants, and the intended audience for the research outcomes that the research methods to be employed and her interpretation of the data collected are appropriate. Regarding pragmatic validity check, the researcher needs to show that the research outcomes are considered to be useful to the practices in the areas relevant to the study. In a similar vein, the notion of reliability is also reframed, which refers to the extent to which appropriate methodological procedures have been undertaken to ensure quality and consistency in data interpretation (Kvale, 1996). Under the phenomenographic research paradigm, some studies may adopt the coder reliability check where two researchers independently code all or a sample of the data collected and compare their outcomes of categorization; and/or dialogic reliability check in which open discussion and deliberation of each researcher’s interpretation of the data is to be conducted until an agreement is reached for the final interpretation (Åkerlind, 2012). An alternative for these reliability checks which is commonly accepted by phenomenographers is for the researcher to fully demonstrate and document how he/she has taken a self-critical approach toward his/her own interpretation and detail the ways in which he/she has conducted the data analysis (Åkerlind, 2005), This is in line with the practice of other qualitative research methodologies in enhancing the rigor of research (e.g., Creswell & Poth, 2018; Merriam & Tisdell, 2016).

This paper aims to complement the discussions of the quality of the phenomenographic research methodology by revisiting the issues of context. The pilot study as well as the main study of a phenomenographic study of the learning of adults under the context of parental learning is reported, which aims at demonstrating the consequences of not taking full consideration of the influence of context in the collection and analysis of data in the pilot study and contrast its outcomes with the quality of the research findings of the main study which has paid due attention to the significance of context.

Attention to Context in Research on Conceptions of Learning

According to Marton (1981), phenomenography aims to describe qualitatively different ways in which the same phenomenon is experienced, apprehended, conceptualized, or understood. However, phenomenography has often been criticized for its tendency to investigate conceptions of a phenomenon without careful reference to the context in which these conceptions were formed (Entwistle, 1997). This is indeed a concern for scholars within this research tradition, because learning in different contexts may actually represent more than one phenomenon but be treated mistakenly as a single phenomenon. When researchers do not pay sufficient attention to the context of the phenomenon they are studying, the data collected from interviews or other methods that seem to be utterances or phrases about the same phenomenon can actually be making reference to different phenomena. These misunderstandings compromise the quality of the phenomenographic research conducted.

Although he would later become a strong critic of phenomenography, Säljö (1979) alerted researchers to the need to pay attention to the context of an experimental situation that aimed to describe people’s subjective conceptions of learning. Instead of relying on the psychological tradition, which explains the behavior and characteristics of people, Säljö (1979) believed that an alternative way to understand learning is to find out and to describe as accurately as possible what people actually do when they are learning. He believed that people’s conceptions of learning have some relationship to their approach to learning, although he denied that there was a direct causal relationship between the two. The primary question of his research was “What do you actually mean by learning?”

In his study, Säljö (1979) demonstrated how to delimit the context of the phenomenon being investigated. The 90 people that he interviewed represented a highly diverse range of learners, with ages ranging from 16 to 73. By highlighting the fact that all interviewees were taking part or were about to take part in some kind of academic education in formal institutions at the time of the study, he established clearly the context of the research as learners’ understandings of learning academic subjects in a formal learning environment. Therefore, the personal views expressed by interviewees, such as that learning in schools was not “real learning,” were not considered relevant to the context of this specific study.

Säljö (1979) pointed out that the context of an experimental study influences the outcome and therefore the interpretation of the results. He believed it was crucial that researchers define clearly the context of the phenomenon under study and keep it constantly in mind in the whole process of investigation. Learning through reading a book and learning in a classroom lecture involve learning in different contexts, and people’s conceptions of learning in these contexts would certainly be fundamentally different. In collecting empirical evidence through interviews, it is therefore important for the researcher to ensure that the interviewees are referring to a phenomenon in the same context, which can be secured through careful structuring of the interview itself.

Later, Säljö (1994) questioned the extendibility of the results and findings from phenomenographic research of conceptions of learning in a particular context to learning situations outside of that context. Ekeblad and Bond (1994) addressed this criticism by clarifying the notion of the internal relation between the person and the thing experienced. Using the example of a student making a distinction between educational learning and social learning, or learning for assessment and learning for “knowing,” Ekeblad and Bond (1994) illustrated the importance of being mindful of the relation of the person and the phenomenon experienced and espoused that phenomenography as a method brings value to educational research, as it is adept at gaining an understanding of the learners’ experience of learning in a particular context, which is otherwise difficult to obtain.

Marton and Booth (1997) posited that a way of experiencing something is related to how a person’s awareness is structured, and they described the relationship between the phenomenon and the context of research on learning by ascribing structural and referential (or meaning) aspects of human awareness to the experience of learning. Marton (1988) elucidated this relationship with the use of gestalt theory (Gurtwistch, 1964), in which a gestalt refers to a collection of items that each determine and support one another. In order to experience something as an identifiable whole from its surrounding context, the experiencer has to perceive that something as a gestalt, a thematic whole, which is discerned from its context. Since a whole is made up of different parts, it implies that the different parts, the relationships between the different parts and that between the parts and the whole must also be discerned by the experiencer simultaneously (Marton & Booth, 1997).

The structural aspect of a way of experiencing refers to the structural relationship between the different aspects of a phenomenon, which in turn constitutes the meaning of the phenomenon. Marton and Booth (1997) used the notion of the internal horizon and the external horizon to account for this intertwined relationship. The internal horizon refers to the parts and their inter-relationships as well as the ways in which the interplay of each part contributes to the whole, which carries an overall meaning. The external horizon refers to “the way in which the phenomenon we experience in a certain way is discerned from its context…[and] how it is related to its context as well” (Marton & Booth, 1997, p. 89). To experience a phenomenon in a certain way, the experiencer must discern the whole from its context, and simultaneously understand its relationship to the context. Nevertheless, to discern something as such from its context, the experiencer would have identified that which is discerned as a certain thing and assigned a certain meaning to it; otherwise, the experiencer would not be able to delimit it from its context. This is what is referred to as the referential aspect, which denotes the overall meaning of the phenomenon. In the act of experiencing a phenomenon, the structural and referential aspects are dialectically intertwined (Marton & Booth, 1997).

Using the “deer in the forest” analogy in Marton and Booth (1997), the structural aspects of a phenomenon comprise the internal horizon, which is the deer itself, the different parts of the deer (the antlers, the legs, the body and the head), and the relationship of these parts to the whole deer, and the external horizon, which is the woods. In other words, the deer is the phenomenon, and the forest is the context. The referential aspect, which is intertwined with the structural, is the meaning of that phenomenon. This includes the posture, the stance, and the activity of the deer. On the basis of Gurwitsch’s (1964) description of human awareness and the distinctions he made between the theme, thematic field, and margin, Marton and Booth (1997) expounded further on the dynamics between the phenomenon and the context. In the language of human awareness, the deer and its parts form the “theme,” and the forest forms the thematic field and the margin. For Marton and Booth (1997), the thematic field or constituent thematic fields and the margin belong to the external horizon, which is the context. Marton (1993) expounded that “awareness has a particular structure also as far as the theme is concerned. The theme appears to the subject in a certain way; it is seen from a particular point of view. The specific experience (or conception) of a theme—or of an object,…can be defined in terms of the way in which it is delimited from, and related to, a context and in terms of the ways its components parts are delimited from, and related to each other, and to the whole” (p. 10).

Adawi et al. (2001) pointed out the importance of making a distinction between the “prepared context” and the “experienced context.” The focus of phenomenography is not the experience of the researcher but that of the research participants, and their experience of the phenomenon is the experienced context. Building on this, Collier-Reed et al. (2009) distinguished three levels or domains of context: the researcher, the collective, and the individual participant. In the domain of the researcher, the researcher needs to be aware of his or her own role and the aim of the study. In setting up the study, the researcher should pay close attention to the object of research, to the particular context of the phenomenon, and to the structuring of the data collection according to the context and object of the study. In the domain of the collective, it is important to ensure that there is a collective understanding that the different parts of the phenomenon refer to the same phenomenon. In the domain of the individual participant, it is important to ensure that the context of the individual is framed to enrich the prepared context of the study through the introduction of a shared experience to structure the data collection process (Collier-Reed et al., 2009).

Reed (2006) reiterated the critical role of context in the investigative process. He also pointed out that both Marton (1996) and Säljö (1997) recognized the importance of context in relation to the investigation of a phenomenon, with both emphasizing that putting participants in situations with meaningful contexts makes a phenomenographic investigation more meaningful.

Phenomenography in the “Ill-Defined” and “Lived” Context of Parental Learning

To illuminate the significance of context in phenomenography as a qualitative research methodology, two phenomenographic studies of the learning of adults under the “ill-defined” and “lived” context of parental learning are reported.

In this study, parental learning refers to learning to be a better parent. To understand the meaning of “ill-defined” context, one may contrast the context of learning in academic subjects with those in parental learning. Learning the subject matter of scientific subjects (such as solid-state physics, complex numbers, and organic chemistry) in the university setting is not influenced by the situation at home or in other social venues. In contrast, the learning of effective parent–child communication skills and the management of the parent–child relationship, both of which are taught in parent education programs, are directly influenced by the situation at home and in other “lived” situations, such as taking a child to a children’s party or responding to a teenage child who is pleading to be allowed to stay out late with friends on the weekend. Parents enrolled in parent education programs often find themselves in situations where the content is necessary for the immediate resolution of ongoing parenting problems. An adult education researcher put it this way: “…the primary concern of much formal education appears to involve transferring the knowledge, skills, and values of an academic discipline to the learner. The majority of adults [in this case, parents], however, do not live in a world that is based on academic disciplines; rather, they live in contexts where they must apply what they have learned in practical situations” (MacKeracher, 2004, p. 14). Moreover, the evaluation of learning in “lived” contexts is normally not through written examinations, essays or term papers, or presentations, but on the outcome of the application of learning in “lived” parenting situations outside formal learning venues.

Unlike quadratic equations or the chemical properties of elements, the influences on parental learning are not limited to the conditions of classroom learning but also include the personal history and current socio-economic status of the parents, the size of the family, and the characteristics of the children, spouse, and any other persons in the household. Parenting solutions have as many permutations as the variables present in each family situation. In contrast to learning science in an academic institution, there are no model answers, no single and exact solution to a given problem, no formula for plugging in the variables, and often no absolutely right or wrong answers. These characteristics of parental learning are the factors contributing to its complexity and are what make the context more “ill-defined.” Additionally, the solutions to parenting problems require knowledge from a wide range of subjects. Unlike academic learning, where the content of subjects is fairly compartmentalized and clearly defined, what is to be learnt in the context of parental learning is an amalgamation of many disciplines.

To illustrate what is meant by the “lived” context, we present a typical parenting and parental learning scenario: Parent learners, after attending a 2-hr parent education workshop on parent–child communication, return home at 10 in the evening to find that their 13-year-old has still not come home. A curfew of 9 o’clock has been previously agreed upon with the child. The parent learners ring the child’s mobile number, but the call is not answered. The learners find themselves immediately confronted with a situation in which they may have to apply certain parenting skills and knowledge to decide what the most appropriate responses are when and if the child comes home that evening. There is no time to review notes, and it is not a multiple-choice test. The parent learners would be experiencing some emotional agitation, perhaps anger mixed with anxiety, which would cloud their thought and make the process of evaluating the situation and deciding on the proper response even more difficult. Perhaps the parent learners are already caught up in some argument or engaging in mutual reproaches and accusations. Moreover, the consequences of what will be said, or blurted out, to the teenage child in the moment of a highly charged confrontation could have a long-lasting undesirable effect on the parent–child relationship. At that moment, the parent learners hear someone at the door, fumbling with the keys.

In the “lived” context of parental learning, learning takes place in parent education program and in real parenting situations. The learners are transformed in both of these settings. Learning continues after the program, such as during and after the situation illustrated above. There is a spillover of the learning process from the formal parent education program to the “lived” family situation, and vice versa. The understanding of parenting and of parental learning also changes with each new situation in which learning occurs. Moreover, parental learning has direct implications for family life, with the application of parental learning often having an effect on the learners and their families, sometimes immediately and sometimes in the medium term, and sometimes with long-lasting repercussions. Unlike failing a math or physics exam, in which case the exam can be retaken or the subject discontinued, the failure to cultivate a good relationship with a child affects everyday living and the quality of family life.

The learner and the context of learning, in and out of the classroom, are inextricably intertwined and continuously and dynamically influence each other. It is not easy to distinguish where parental learning starts and where it stops, as drawing a line between classroom and situational learning in parental learning is exceedingly complex. Delimiting the context of research on a “lived” situation of learning could therefore be more difficult than when conducting research into learning in an academic milieu.

In the next sections, we share some empirical experiences of a researcher’s lack of attention to context in the pilot study and the ambiguous outcomes of such an oversight. We present the lessons learned from this experience and illustrate how these have improved the quality of research in the main study. The importance of context in phenomenographic research did not dawn on the researcher until only after he had presented the results of the pilot study to experienced phenomenographers. The pilot study aimed to identify parents’ different ways of understanding some key concepts in parent education, while the main study focused on discovering the conceptions of learning held by parents in parent education programs.

The Lack of Attention Paid to Context in the Pilot Study

Through interviews and interactions with parent education facilitators and parents, and through observations of the parent education programs, certain themes, such as the awareness of parents as primary educators and the use of freedom and self-regulation among children, were recurrent and persistent in the parent education programs of the school. The researchers decided that these were the key concepts in parent education; from these, the concept “parents as primary educators” was chosen for the pilot study.

In the pilot study, 18 parents were interviewed, 14 of whom were enrolled in the school’s parent education program at the time of the interviews, and four of whom were not. The four non-enrolled parents were included because they were considered rich sources of information and thought likely to provide interesting perspectives. The interviews used scenarios to frame a shared experience from which to tease out the parents’ understandings of the concept of “parents as primary educators” in a single common context.

Interview protocol:

How many children do you have? How are they getting on in their studies? What are their favorite subjects? etc.

What do you understand by “parents as primary educators”? What do you think “primary” means? “Primary educator” of what?

Could you give some examples that illustrate “parents as primary educators”?

Scenarios:

Parents of students in your son’s class complained to the Form Teacher that your son was bullying his classmate. After some investigation, the Discipline Master confirmed that your son did bully his classmate because the classmate teased him. The school intends to give both students a major demerit. How do you feel?

When your child was in primary school, he liked maths or some other subject, and performed well in it. In fact, he received high marks. However, when he joined Form 1, he seemed to lose interest in the subject, habitually did not hand in homework in that subject, and even received poor grades. How do you feel about this?

The scenarios were used in the interviews to frame the context of the concept in terms of the child’s academic and character development. However, in the interview itself, the discussion often shifted away from the phenomenon framed by the scenarios and onto personal experiences and various family circumstances, such as having or not having domestic helpers in the household, leaving the care of the child to grandparents or relatives, whether the parents had to work or not, etc. The researcher who conducted the interview did not pay the necessary attention to directing the discussion back to the scenarios, which were supposed to serve as a “shared experience” from which to tease out the parents’ understandings of the concept of “parents as primary educators” in a single common context.

At some moments in the interviews, the interviewees disagreed with the concept of “parents as primary educators.” One parent considered her uncle as her primary educator because of the love and affection he showered upon her and her siblings. Another thought that it was the child’s grandmother, with whom she was left during her early childhood, followed by the domestic helper, who spent the most time with her while she was in primary school, and finally the teachers in high school, whose advice was heeded by the child more than that of the parents. Another dissenting parent believed university professors, in providing the guidance for the career of a young person, to be the primary educators.

“Phenomenography is not about what different things people see, but about in what different ways the same thing is seen. This is why it is an advantage if the interviewer formulates their question about a certain phenomenon (or an object of learning) in terms of a particular instance (a task, a text, a scenario, etc.). The instance chosen can be considered as a kind of context defined by the researcher. Otherwise, the researcher has to distinguish between the interviewees talking about the target phenomenon and when they are talking about something else. In accordance with the idea of phenomenography, the researcher attempts to separate (and to each other relate) the different ways of perceiving the phenomenon (the object of research) on the one hand and the context in which they are identified on the other” (F. Marton, personal communication, November 14, 2020).

On the basis of the framework proposed by Collier-Reed et al. (2009), what we learnt from the pilot study was that the researcher failed to pay attention to the research context. In the domain of the researcher, there was a rudimentary lack of control of the interview process such that interviewees were allowed to freely deviate from the scenarios. In the data analysis, the researcher mixed up phrases and utterances concerning the target phenomenon and those phenomena other than the target one, mistakenly categorizing them together. At the root of this was the researcher’s simple lack of awareness of the impact of the context of research on the results of the study. Such an awareness is vital and should be emphasized to newcomers to the field of phenomenography.

In the domain of the collective, the lack of sensitivity to the research context at the beginning of the process had already rendered the data collection problematic. The researcher treated all the phrases related to the concept as relevant to the data analysis because there was little appreciation of the implications of the context for the phenomenon being investigated.

In the domain of the individual participant, the scenarios were designed to bring to focal awareness the concept of “parents as primary educators” in relation to the child’s academic and character development. However, due again to the lack of awareness of the significance of confining the data collection within the research context, the attempt to frame the interviews within the “prepared context” was ineffective.

To apply the two evaluative criteria, i.e. communicative validity and dialogic reliability mentioned above, to evaluate the quality of the phenomenographic research conducted in the pilot study, it was found not to be satisfactory. In the data analysis, utterances and phrases referring to the concept beyond the confines of the scenario were mixed in with the data. The researcher failed to distinguish between the participants talking about the target phenomenon (i.e., parents as the primary educator) and when they were talking about some other phenomena (e.g., the impact of a university professor on one’s lives), which deviated from the phenomenon to be investigated. Therefore, the categories of description derived from the data did not refer to the same, target phenomenon, and the findings of the pilot study was called into question by experienced phenomenographers when the researcher presented his analysis of the data to the phenomenographic research community at an international conference. In particular, they found it questionable whether the interviewees were talking about their experiences of the same target phenomenon, as some shared their recollections of how a university professor or senior impacted on their lives when they were young whilst some others voiced out the nitty-gritty of their own parenting experience. The experiences shared were situated under very different contexts and the interviewees may not refer to the same phenomenon (i.e., parents as primary educator). Some members suggested that it would be highly useful to make use of the prepared scenarios to serve as a “shared experience” to elicit the parents’ understandings in a single common context and to minimize the possibility of different people talking about different phenomena situated in different contexts. Obviously, the communicative validity of the pilot study was impeded, as it could not get the endorsement from the relevant research community. Furthermore, without careful consideration of the significant impact of context on the quality of phenomenographic study, it was rather difficult for the two researchers involved to fully agree upon whether the interviewees referred to the phenomenon in question in their experience-sharing. As such, there were discrepancies between the two researchers involved in their interpretation and categorization of the data collected, and thus the dialogic reliability was not well established in the pilot study.

Lessons Learned and Its Application to the Research on “Conceptions of Parental Learning” in the Main Study

After experiencing the negative consequences of ignoring the cautionary advice about context on the quality of research in the pilot study, the researcher took on the suggestions of the experienced phenomenographers and made amendments in the design and implementation of the main study, in which the phenomenon to be investigated was “parental learning.”

Eighteen participants in the Basic Parenting Program were interviewed, all of whom were parents of Grade 7 pupils of a secondary school in Hong Kong. In the interviews, the participants were asked what they understood as “learning” in the parent education program that they were attending. Specifically, they were asked, “What do you understand by learning?” “What do you mean when you say you have learned something?” “How do you know you have learned something?” and “What do you mean when you say you have understood something?” The interviews averaged 60 minutes in length. Each interview was audio-recorded and transcribed verbatim.

Following the phenomenographic tradition, data were collected through open-ended, semi-structured interview, which is the major data collection method for phenomenograhic research, aiming at providing a deeper understanding of the phenomenon in question as experienced by humans (Kvale & Brinkmann, 2015; Marton, 1988). Although an interview protocol was developed, the researcher does not follow the exact structure in the protocol. The focus of the conversation between the interviewee and the interviewer is to unfold the lived experience of the interviewee regarding the target phenomenon.

Having understood the role of context in the domain of the individual participant and recognized the importance of the shared experience in delimiting the context from the pilot study, the researchers took extra care to select individuals only from among the current participants of a specific parent education program for the main study on “conceptions of parental learning.” Furthermore, among the selected interviewees, the researcher used the shared experience of the program to delimit the context of the research and paid attention to the need to distinguish phrases and utterances that were directly related to the shared experience on the target phenomenon from those that referred to learning experiences outside the program. We thus remained focused on the internal relation between the subject and the object of learning, which Ekeblad and Bond (1994) pointed to as the object of phenomenographic study. In this study, we distinguished between educational learning and social learning, or learning in the “lived” context. It was not uncommon during the interviews for utterances referring to parental learning in the daily circumstances of life to be mixed with utterances referring to learning in the prepared context of the study. Staying vigilant to the internal relation of the person and the phenomenon is essential for ensuring the delimitation of the research context. Following the approach proposed by Collier-Reed et al. (2009), the context of the individual was framed through the introduction of a shared experience. The parenting education program as the prepared context of the study provided the delimiting structure for the data collection process.

In the domain of the collective, longer excerpts with embedded phrases were selected to ensure that we did not lose the context in which a phrase was uttered. Regarding utterances that referred to parental learning outside the classroom setting, a judgment was made regarding whether they were connected directly to what had been learned in the program. For example, one parent learned some techniques to improve communication with their child and, in the process of applying them at home, reflected upon their initial failure, made some modifications, and reapplied what had been learned with better results. Because this change was precipitated by a parent’s heightened awareness of the child’s viewpoint after attending the class, this was included as part of the experience of the collective in the parent education course. Utterances relating to parental learning from books, television programs, and parenting websites, however, were considered to be outside the research context and were thus excluded from the pool of phrases for analysis.

The researchers first read the transcripts while listening to the audio recordings and checking the accuracy of the transcriptions. This process served to re-familiarize the researchers with the data, which was necessary because of the time lapse between the interviews and the analysis of the transcripts. In the second reading of the transcripts, utterances or phrases that were deemed relevant to the parents’ conceptions of their learning were selected. These were then treated as a single pool of phrases. Because the aim of phenomenography is to reach a collective understanding rather than an individual perception (Marton & Booth, 1997), “the researcher shifts the focus from the individual subjects to the collective ‘pool of meanings’ to be discovered in the data” (Marton, 1986). The analysis of data was iterative, with the data being sorted and resorted. The selected phrases or quotes, through constant comparison, were grouped together according to the similarities found among them. What emerged were groups of phrases that, upon further analysis, could be merged into a single group when similarities were found or divided into multiple groups when relevant differences were found to justify such a division. Through the grouping and regrouping of the relevant phrases from the interview transcripts, a limited number of qualitatively different groups of conceptions, understandings, or ways of seeing emerged into what in phenomenographic terminology are called “categories of description” (Marton, 1986).

Drawing upon Newberger’s (1977) theory of parental awareness and Knowles et al. (1998) principles of adult learning, the relationships between categories were defined and arranged in a hierarchical order. The final outcome was a set of categories of description organized into a structured framework showing their relationships and reflecting the logical complexity of the domain and the different depths (Marton & Pang, 2008) at which parental learning is experienced.

The study identified six conceptions of parental learning among parents participating in an in-school parent education program: [a.] recalling experiences, [b.] acquiring problem-solving skills, [c.] applying problem-solving skills, [d.] awareness of a parent–child relationship, [e.] noticing alternative understandings of parenting, and [f.] change in a person (Go, 2013). The first three of these conceptions [a, b, and c] were focused on regaining control of the child’s behavior, with a learning strategy of acquiring and applying quick fixes, while the last three conceptions [d, e, and f] were focused on the development of the parent–child relationship, and the adopted learning strategy involved a more systematic and critical reflection on assumptions (Brookfield, 2010; Mezirow, 1998). Furthermore, in the first three conceptions [a, b, and c], the problem was treated as external to the learner, while in the last three conceptions [d, e, and f], the learner was treated as part of the system and therefore also part of the problem, which often leads to transformative learning (Mezirow, 2000). A hierarchy of attention to relationships and depth of learning was found among the conceptions, with conception [a] considered the lowest level of learning and [f] the highest. The first three conceptions were considered to indicate a surface learning approach and the last three a deep learning approach (Pang & Go, 2015).

After paying close attention to the role of context and undertaking the abovementioned measures, such as using the parenting education program attended by all interviewees as a shared context, the researchers realized that the responses of the interviewees were more focused on the same target phenomenon when compared to the pilot study, with their utterances demonstrating qualitative variation in their ways of experiencing parental learning. During the interviews, once the interviewees were found to be digressing from the target phenomenon, the researcher learnt the lesson from the pilot study and was more alert in guiding the interviewees to come back to talk about the phenomenon in question. As a result, fewer utterances outside the target phenomenon were identified, which made it easier for the researchers to come to an agreement that the interviewees were referring to the same phenomenon in the same context in the conversations. Consensus could be reached among the three researchers (i.e., one more researcher was involved in the main study) in the analysis and categorization of the data, which in turn enhanced the dialogic reliability of the study. Finally, when the researchers presented the results of the main study to the research community at an international conference, the methods of collecting and analyzing the data were judged by the expert members to be appropriately and rigorously conducted and the research outcomes were widely accepted. As such, the communicative validity of the main study was greatly enhanced, when compared to the pilot study.

Conclusion

Phenomenography is a qualitative research methodology that is often adopted by researchers to investigate people’s lived experience of the phenomena around them. Within the phenomenographic research tradition, there has been much discussion of the influence of the context on the phenomenon under study. Time and again, both phenomenographers and critics of phenomenography have stressed the importance of the researcher being mindful of the context when using phenomenography as a research methodology. In this paper, two phenomenographic studies of the learning of adults under the context of parental learning are reported. We attempt to illuminate empirically the consequences of ignoring the cautionary advice about context in the pilot study and contrast its outcomes with the quality of the research results of the subsequent main study that heeded the advice. It is concluded that it is important to clearly delimit the phenomenon in question by taking careful consideration of the relevant context, as this ensures that phenomenography is conducted on the same, target phenomenon, rather than on different phenomena. The need to pay attention to the context of learning in the phenomenographic research tradition is essential, and it is critical to the quality of phenomenographic research. By putting due attention to the role of context when conducting phenomenography, the quality of research in terms of the communicative validity and dialogic reliability of the study can be substantially enhanced.

The framework proposed by Collier-Reed et al. (2009) is useful for helping researchers to be mindful of the relationship between the phenomenon and its context. As people’s experience of phenomena is context sensitive, it is essential for new practitioners of phenomenography to be made aware of the importance of distinguishing the different contexts in which a phenomenon may occur and to learn how to delineate the constituent thematic fields in relation to the theme. The influence of the context on the outcome of the research is indeed crucial, and oversight in this respect can lead to ambiguity and to the questionable quality of results. This is even more crucial to the study of phenomena which are under “ill-defined” and/or “lived” contexts such as parental learning.