The study of relationships is a complex field incorporating a wide range of themes, including emotional aspects, functions, patterns and processes, cognitions as well as the cultural and ideological structure of relationships within society. It seems inevitable that some studies will cut across different methodological perspectives. For example, some studies of friendship have involved in-depth interviews, experimental manipulation of variables and rating scales (Allan, 1989). Research methods in psychology have been typically divided into quantitative and qualitative approaches, although it may be said that the treatment of data derived from both may be put together in a variety of ways according to the particular study being undertaken. It will be argued that although all research methods have their inherent shortcomings, the complex, dynamic and interactional nature of relationships are best explored through a qualitative orientation, particularly within the social constructionist model.
The quantitative approach is generally associated with traditional experimental and survey methods where measurement in terms of commonly understood and replicable standard units is undertaken, usually through the use of rating scales (Dey, 1993; Breakwell et al, 2000). Research questions within this approach are typically formulated prior to the collection of data, and then data is analysed to generate answers to the initial questions, generally known as the deductive method. In the study of relationships, then, research questions within this approach often need to be framed into fairly precise hypotheses regarding certain key features of relationships and the interactions taking place within them. Hypotheses, thus, may be tested in terms of the correlations between pairs of key variables, such as relationship duration and relationship commitment, or satisfaction as compared with intimacy in relationships.
The qualitative approach, by contrast, is generally associated with ethnographic methods of data collection which seek to produce data which are freely defined by the subject under study. Typical methods under this approach include participant (and non-participant) observation, unstructured interviewing, group interviews and the collection of documentary materials such as diaries, video or tape recordings (Miell and Wetherell, 1998). Research questions tend to be more open-ended, allowing the data to emerge more freely so that it may then be shaped into particular groupings. Questions, thus, are likely to be expressed as follows: ‘What are the different ways in which people express commitment in relationships?’, or ‘how do couples maintain harmony in their relationship?’ It is argued that the data elicited in this way is likely to be richer and more valid because it is derived from freely expressed, subjective experiences and not according to pre-determined, researcher-led categorisations (Banister et al, 1994; Hammersley and Atkinson, 1995).
In reality, however, as Dey (1993) points out, the distinctions between these two approaches are not quite so clear-cut. For example, certain methodological tools, such as survey questionnaires, can resist rigid classification into ‘structured’ (quantitative) and ‘unstructured’ (qualitative) techniques. A series of closed questions in a survey questionnaire may still include, in its range of response categories, the option ‘Other – please specify’ which invites the respondent to give his/her own subjective response. Furthermore, even the most non-directive interviewer must implicitly direct an interview to some extent of it is to cover certain topics within the time available. It would seem that data collection, whatever the methodological stance, will always involve selecting data, and also, the techniques of data collection and transcription are likely to affect what finally constitutes ‘data’ for the purposes of the research (Dey, 1993).
When we examine the different types of method employed by relationship researchers, it is clear that each is characterised by its own particular strengths and weaknesses. Examples of experimental studies carried out in controlled laboratory settings include those of Byrne (1971) which examined the relationship between perceived similarity and liking in couples and the study by Norman and Aron (1995) which explored the links between couple relationship satisfaction and shared activities according to their degree of novelty and excitement (Ickes and Duck, 2000). In the latter experiment, couples in ongoing relationships are asked to participate together in a task in which their interaction is video-taped and also to complete questionnaires both before and after completing the task. The questionnaire responses essentially constituted a measure of relationship satisfaction and the independent variable was the assigned activity which was manipulated to be either novel and arousing or boring and sedate. The researchers found significantly greater increases in satisfaction for the group participating in the arousing/novel activities as compared to the other group, and that this increase was significantly greater for those who had been together longer. These results were consistent with the original predictions (Ickes and Duck, 2000).
One advantage of such an experimental study is that if it is well-designed, it can establish clear associations between independent variables, manipulated and controlled by the experimenter, in this case the different tasks set, and dependent variables, the questionnaire responses, which are measured by the experimenter (Miell and Wetherell, 1998; Argyle, 1996). However, a disadvantage is that ‘staged’ interaction situations may be perceived by participants as artificial or unnatural. Furthermore, the specific conditions set up may not match the every-day lived conditions of participants’ lives and as such, may not have ecological validity or generalise well to other settings (Ickes and Duck, 2000).
Questionnaires designed to elicit self-reported responses regarding personal relationships are widely used yet have been subject to a number of criticisms. Firstly, as Miell and Dallos (1996) have noted, participants are known to record their relationship development as being more positive in retrospect than they had actually experienced it as it was happening. Secondly, questions such as “to want extent would you describe your conversations with your partner/friend/colleague as generally rewarding?” can be seen as highly problematic. For example, respondents may interpret ‘generally rewarding’ in different ways, a single tick on a scale is unlikely to describe what generally happened and also, the questions asked may not necessarily accurately reflect features of the participants’ relationships.
Peer Report methods and diary studies are examples of research tools which may avoid some of the problems described above. Peer reports from respondents provide data in the form of verbal or written accounts about the relationship(s) of people with whom they are acquainted. Researchers may seek evidence of consensus in peers’ perceptions of a given relationship, and one advantage of this is that “peer report research tends to focus on the shared, inter-subjective reactions of a set of peers who all view the relationship from the outside, as observers and knowledgeable informants” (Ickes and Duck, 2000, p.167). Diary studies are seen as a way of describing how everyday relationships are actually experienced. Ickes and Duck (2000) cite the influential early diary research of Wheeler and Nezleck (1977) in the USA, finding, for example, that life satisfaction, for males and females alike, could be predicted by the number of female acquaintances with whom the participant had interacted with during the study period. Diary methodology may involve responses to structured questionnaires or take the form of open-ended, accounts to record a participant’s daily experiences of their social interactions and/or personal relationships over a specified period of time.
Diary researchers who use the structured questionnaire approach tend to favour quantitative analyses of data, as described earlier, hoping to combine the rigour of this approach with the naturalistic setting of the participant’s everyday life. Use of the open-ended diary method, however, is more akin to a social-constructionist perspective (Wetherell and Maybin, 1996). It will elicit subjectively constructed statements from participants, obtaining a richness of data which may be described as high on external validity, that is to say applicable to the lived experiences of real relationships. However, a trade off seems inevitable in that as external validity is achieved, internal validity is compromised since the data generated is less controllable by the researcher.
Miell and Dallos observe that relationships seem to present us with a dilemma, “on one hand partners engage in repetitive and predictable sequences of actions and on the other hand, each person has the potential to act autonomously, to make choices based upon their understandings” (1996, p.150). This dichotomy seems to echo the contrasting methodological approaches in that quantitative methods tend to be more suited to studying those repeated and predictable actions, whilst the exploration of the unique, subjective experiences of people may best be explored through qualitative means.
The social constructive perspective espouses the view of people as actively constructing meanings about their relationships and the associated actions that they choose to take. Mutuality and inter-subjectivity are features which many researchers see as the central focus in their enquiry into relationships. Miell and Dallos (1996), for example, comment that early relationships “form the basis of individual beliefs, attributions and narratives which partners carry into the melting pot of later relationships” (p.151). Our core constructs, comprising our understandings and beliefs about people, are seen as highly influential in how we interact with and form relationships with other people.
Thus, it seems that research methods which can shed light on these constructions and the way they are manifested between people within relationships could be very productive.
Personal Construct and Attribution Theories (Kelly, 1970; Harvey, 1992) both seek to explain how people try to make sense of and manage their relationships and PCT, in particular, can capture both the uniqueness of partners’ meanings and the essentially interpersonal nature of our constructions. Within the same genre of research enquiry, there are those who study accounts from people who express their reflections about their relationships often take the form of a story or narrative. This method of studying relationships in terms of partners’ constructed accounts raises important issues about the extent to which it may be considered rigorous and ‘scientific’, especially for those who favour the empirical tradition. However, there seems to be a strong argument for suggesting that the nature of relationships is too complex and idiosyncratic to be reduced to attempting to establish generalised models or make predictions. As Shotter (1987) asserts
“What we need is a better story, a better way of formulating the nature of personal relationships than the current ‘causal story’………a story that fits with the practice of personal relationships, rather than with the established practices of science – for in personal relationships, too, we can check, evaluate, and elaborate the truths we make as we see them” (Miell and Dallos, 1996, p. 132).
The role played by language and discourse is seen as important in the accounts given by people and three further elements are also highly relevant namely, that a story is structured by and unfolds over time, an account always has a purpose to ascribe blame or responsibility and accounts often change according to our situation or the person to whom the account is given (Miell and Dallos, 1996). All these features of personal accounts constitute a richness and level of authenticity that would be difficult to capture through a reliance on data produced within the confines of a quantitative or experimental approach as described earlier.
This theme of relationships as actively constructed at the level of actions and of meanings is extended through the systemic perspective. This approach is derived from systems theory which essentially rejects traditional linear explanations of cause and effect in terms of one person’s feelings, thoughts and actions in relation those of another. Instead, circular explanations are emphasised in which causes reside in the interactional processes between two or more people, with all those involved reacting to each other and, at the same time, proactively stimulating the other to react (Bateson, 1972; Minuchin, 1974; Haley, 1976). Systemic perspectives argue that the interactions between people constitute an essentially creative process in which actions and meanings are jointly constructed, emerging and evolving over time, rather than simply being determined by the characteristics of the people involved. Systems theory, used in clinical circles, has been effective in addressing seemingly rigid and intractable problems within families, in particular.
In terms of methodology, researchers, and therapists, within this tradition could be regarded as participant observers, engaged in a process of careful observation of the interlocking patterns of actions and meaning unfolding between participants over time. The observer him/herself is also acutely aware of his or her own part in the co-construction of meaning within the system and reflexivity is thus an essential part of the process. Systemic approaches are also able to take account of wider societal discourses and their interactional processes as discussed by Miell and Dallos (1996). The ways in which these perspectives are able to address the complex, ongoing, dynamic and interactional qualities of relationships renders them highly persuasive.
In conclusion, then, quantitative methods have been used widely in relationship research, particularly when there are clearly and narrowly defined research questions where variables, such as relationship satisfaction or pre-defined personality traits, can be controlled and comparisons made. Research instruments such as surveys and questionnaires typically use rating scales, the responses from which may be examined through statistical analyses to produce data which may either refute or support initial research questions. In reality, many research studies have used a combination of both quantitative and qualitative methods, for example, a questionnaire may be used as an initial tool which may later be combined with an ethnographic study. Qualitative methods, on the other hand, notwithstanding its own limitations, can generate a richness of data, from diary studies and personal accounts, for example, which seems better suited to the complexities of ‘real’ relationships. The data obtainable from those methods subsumed under social constructionist approaches, particularly, Personal Construct and Attribution Theories and the systemic perspective, however, seems to offer the most promising insights into the dynamic, ongoing and interactional nature of relationships.
References
- Allan, G (1989) Friendship: Developing a Sociological Perspective, Harvester-Wheatsheaf, Hemel Hempstead
- Argyle, M (1996) the Psychology of Leisure, Routledge, London
- Banister, P, Burman, E, Parker, I, Taylor, M, Tindall, C (1994) Qualitative Methods in Psychology, Open University Press, Buckingham
- Bateson, G (1972) Steps to an Ecology of the Mind, Ballantine, New York
- Breakwell, G, Hammond, S, Fife-Schaw (2000) Research Methods in Psychology, Sage, London
- Dey, I (1993) Qualitative Data Analysis: A user-friendly guide for social scientists, Routledge, London
- Haley, J (1976) Problem Solving Therapy, Jossey Bass, San Francisco
- Hammersley, M, Atkinson, P (1995) Ethnography: Principles in Practice (2nd Edition), Routledge, London
- Harvey, J.H., Orbuch, T, Weber, A (Eds) (1992) Attributions, Accounts and Close Relationships, Springer-Verlag, London
- Ickes, W, Duck, S (2000) The Social Psychology of Personal Relationships, John Wiley & Sons Ltd, West Sussex
- Kelly, G (1970) ‘A brief introduction to personal construct theory’, in Bannister, D (ed) Perspectives in Personal Construct Theory, Academic Press, London
- Miell, D, Dallos, R (1996) Social Interaction and Personal Relationships, Sage, London
- Miell, D, Wetherell, M (1998) Doing Social Psychology, Sage, London
- Minuchin, S (1974) Families and Family Therapy, Harvard University Press, Harvard
- Wetherell, M and Maybin, J (1996) ‘The distributed self: a social constructionist perspective’, in Stevens, R (ed) (1996) Understanding the Self, Sage/The Open University, London.