Copyright Sociological Research Online, 1999

Odette Parry, Carolyn Thomson and Gerry Fowkes (1999) 'Life Course Data Collection: Qualitative Interviewing using the Life Grid'
Sociological Research Online, vol. 4, no. 2, <http://www.socresonline.org.uk/4/2/parry.html>

Abstract

The life grid, has recently been acclaimed as a accurate method for collecting retrospective data from elderly respondents. Accounts of using this method, which are based upon quantitative studies, however, have not adequately captured the dynamics of grid interviewing. This is, we suggest, because it is much easier to describe technical aspects of a research method than it is to convey how that method works in practice. In this article we set out to portray some of these more 'indeterminate' aspects of the life grid interviewing, drawing on our own experiences of using this method in an qualitative study of lifecourse patterns of smoking behaviour among elderly respondents who have a smoking related illness. Focussing upon interaction between researcher and respondent the article explores the reconstruction of the life course as a mutual endeavour and the implications of this for the interview structure and data collection among this respondent group.

Keywords:

Arterial Disease; Indeterminacy; Life-grid; Qualitative Methods; Smoking

Introduction

1.1

The paper draws on our experiences of using a structured life history method for collecting qualitative data on lifetime smoking behaviour. The life grid method was first used in Britain for The Social Change and Economic Life Study (Gallie 1988). Since then method has been developed and tested by researchers at the Imperial College School of Medicine who have recently used life grids to study the association between cumulative exposure to health-damaging residential and occupational environments, and socio-economic variations in physical health (Holland et al 1999, Berney & Blane 1997, Blane 1996). In addressing some of the problems which have been traditionally associated with the collection of retrospective data this work has been very important in establishing the life grid as a respectable research method.

Problems associated with retrospective methods

1.2

The main criticism of studies which use retrospective methods, such as the life grid approach, has been that, because respondents' memories fade with time, the data that they yield are prone to inaccuracy. Thus, in the case of life history research with older people, the data may be both partial and biased as past experiences are repressed or merged with others. Recall bias is further accentuated in that accounts of past events will reflect, and may be distorted by, the position of the teller and the importance with which he/she wishes to imbue particular occurrences.

1.3

It should be pointed out, however, that all interview data, and not just that obtained through retrospective accounts, reflect those stories which respondents wish either consciously or subconsciously to tell. Furthermore, recall may not be as inaccurate as is generally supposed. There is evidence to suggest that forgetting is a gradual rather than an accelerated process and that a considerable amount of information may persist for a lifetime (Squire 1989). For example, the accuracy of recall about past height, weight and body build has not found to be influenced to a significant extent by the passage of time (Casey et al 1991, Must et al 1993) and neither have factors such as education, occupation or place of birth (Field 1981). Respondents' recall about their smoking behaviour has also been found to have high levels of agreement with their earlier accounts, although admittedly it has been found to be rather less accurate in the case of ex-smokers who tend to under-estimate the amount smoked (Krall et al 1989).

1.4

Concern about retrospective data has tended to draw attention from, or undermine, their potential accuracy particularly around life events which are meaningful to the respondent (Blaxter & Paterson 1982). Also, while it has been argued that accuracy of recall is influenced by the emotional load attached to particular life events, recent research suggests that those events, or occurrences, which are less emotionally charged are likely to be most accurately recalled (Blane 1996). If this is the case, then methods relying upon respondent recall should be particularly apt for exploring those routine behaviours which Blane describes as the 'hum-drum' background routines of daily life. In the case of smoking this is particularly pertinent because while there is an abundance of data on starting and stopping (two crucial events in smoking biographies) there is a dearth of data on the middle or uncharted years of smoking behaviour.

1.5

Further evidence suggests that the ways in which data are collected are very important in minimising recall problems. For example, it has been shown, that when respondents are prompted precisely, there is higher agreement between the dates which are remembered and the actual dates when events occurred Livson & McNeill 1962). Furthermore, it has been found that people often recall facts about public events along with information about the personal context in which those facts are encountered (Brown 1990). Hence, probing around these events is an effective way to prompt respondent recall of personal experience. It is also suggested that the identification of key life transitions and turning points can provide a framework within which socio-biographical detail can be usefully explored (Humphrey 1993). This approach has been found to be particularly useful in research with older respondents who are often able to construct their biographies using a series of meaningful reference points (such as the war) which serve to reinforce memory (Walter et al 1987).

The life grid as a qualitative research tool

1.6

While the literature demonstrates that the life grid is a method facilitating the collection of more accurate recall data, less attention has been paid to other methodological advantages which it offers. To date, the use of the grid method has been illustrated solely through reference to its contribution to quantitative research. Here we explore the part that life grids can play in qualitative methods by focusing upon the context of the life grid interview and implications which this method may have for the relationship between researcher and respondent.

1.7

In order to capture or portray the dynamics of the grid interview it is necessary to describe some of the more 'indeterminate' aspects of interviewing work. 'Indeterminacy' is a concept which was first used to demonstrate particular aspects of occupational work (Fox 1957, Polanyi 1958, Strauss et al 1964) and was later developed by Jamous and Peloille (1970) who juxtaposed it against 'technicality' to provide a classificatory device for the professions. On the one hand, technicality can be described as that part of occupational work which lends itself to documentation, and at its extreme this aspect of work can be conveyed by a list of specifications graspable through memory and physical dexterity. Indeterminate knowledge, on the other hand, implies a kind of tacit, implicit knowledge which remains the personal property of the practitioner. It can be argued that, in a similar way to other occupational work (Atkinson et al 1977, Umstron-Phillips 1982, Atkinson 1984, Parry 1993) research interviewing incorporates both technical and indeterminate aspects. Whereas research instruments can be described and instructions for their use can be specified, it is the more indeterminate aspects of the interview which defy translation into set formulae. Because of this, texts which describe how to use a particular method are often unable to portray to the reader a feel for how that particular method is actually accomplished in practice.

Methods

2.1

The focus of the study, on which this paper draws, is life-course influences on patterns of persistent smoking. The study aims to facilitate an interdisciplinary understanding of smoking by exploring the relative influences of nicotine exposure, intrinsic factors and life events/occurrences upon long term patterns of smoking behaviour.

2.2

The rationale behind the facilitation of an interdisciplinary approach is that smoking research has traditionally been highly discipline bound. The different research paradigms of, for example, addiction, psychology and sociology, have tended to carve up smoking behaviour into their respective interests at the expense of permitting a more holistic understanding of the behaviour.

2.3

In order to reach an holistic understanding the research explores different aspects of smoking behaviour traditionally prioritised by the respective disciplines. Hence in order to ascertain levels of nicotine exposure data is collected on patterns of usage such as the number and strength of cigarettes smoked, length of time smoking, time of day first cigarette is/was smoked, quitting, periods of abstinence and return to smoking. Cognitive factors explored include self efficacy, perception of risk, control over performance and perceived social pressures. Social factors include social position, social support and life events/occurrences including personal life events (such as a status transition from employed to unemployed, or married to divorced) and external factors such as tobacco pricing and health initiatives.

2.4

The study sample comprises seventy long-term smokers (and ex-smokers) between the ages of 65 and 85 years. The sample was drawn from participants of a much larger and continuing longitudinal study of arterial disease. The selection criteria included smoking status (respondents were current smokers or were ex-smokers with lengthy smoking histories) and health (respondents had a smoking related arterial illness). This paper is based on twenty qualitative interviews carried out in the first eight months of the research study.

2.5

Because we wanted to explore associations between life experiences/events and smoking it was important that we collect detailed data on the respondents lives. The retrospective interview enabled respondents to reflect upon and reconstruct their biographies, giving adequate time for subjective reworking and reinterpretation of past events. Published work by Blane and colleagues (Blane 1996, Berney & Blane 1997) led us to consider using the life grid to facilitate collecting fairly structured data on respondents' lives at the same time being flexible so that events and issues which respondents raised could be explored with them in-depth. We were fortunate to receive excellent training on the life grid method, geared specifically to our own needs, from David Blane and his colleagues.

Technicality: the structure and method of using the life grid

2.6

In the interview the grid was used to cross-reference the dates of any changes experienced by respondents, throughout the life course, to their smoking behaviour. The grid comprised one sheet of A3 size graph paper, divided into 5 columns. A grid was prepared for each respondent before the interview. An example of a cross section of a prepared grid is illustrated in figure 1.

External Events	age	family	work/leisure	housing	health	smoking
1915
1960	45
Royal Col. Phys Report
US Surgeon Gen. Report
1965	50
World Cup
Apollo
1970	55

Figure 1:This table illustrates a section of grid prepared for a respondent born in 1915. Follow the hyperlink to see completed grid for a second respondent [195k].

2.7

The vertical axis displayed dates (at intervals of five years) starting with the respondent's date of birth. On this axis the respondent's age (at intervals of five years) was shown. To the left of this axis, several external events (indirect indicators), including, for example, the second world war, the coronation, when England won the World Cup etc. were noted. The indirect indicators also included several external events relating to smoking, such as publication of the Surgeon General's report, the date when health warnings on cigarette packets were introduced, budgets which introduced significant price increases for cigarettes etc.

2.8

Each of the five columns to the right of the vertical axis related to an area of the respondent's personal experience. Events recorded in these columns comprised the 'direct indicators'. These were family, work/leisure, housing, health and smoking. During the interview, events under each heading, such as marriages, divorces, redundancies etc. were entered chronologically into the grid. In the column for smoking, significant dates were recorded, such as when the respondent started smoking and any quit attempts. During the interview, which was tape recorded, the researcher was able to cross reference direct and indirect indicators with the respondent's reported smoking behaviour and probe in-depth around these associations.

2.9

At the beginning of the interview, the researcher described how the grid would be used to capture different experiences and events over the respondent's life course. Permission was also sought to tape record the interview so that the researcher could concentrate on the grid rather than on manually recording discussions around events and issues.

Indeterminacy and life grid interviewing

Introducing the grid

3.1

None of our respondents had previous experience of the grid, yet each appeared satisfied with the researcher's brief explanation of how it would be used. Understanding of how the grid actually works, however, is best reached through 'hands on' experience, and from the onset the grid assumed the focal point of the research encounter. Because the grid was prepared immediately prior to the interview, the researcher had an indication about the respondent's lifecourse before they met. The first date on the grid was the respondent's date of birth and a glance the age of the respondent at each of the external events charted on the grid was apparent. The demonstration of, albeit limited, familiarity with the respondent's lifecourse was advantageous, especially in the very early stages of the interview when respondents were unsure what was expected of them. In one interview, when the researcher first looked at the grid she was able to say to the respondent 'you had your birthday just last week', to which the respondent replied 'well yes', showing a mixture of both surprise and appreciation. It is interesting that, although the respondents were fully aware of how the grid was used in the interview, they were nevertheless often surprised during the interview when the researcher made a connection between two events. For example, a respondent who claimed to have given up smoking in his late thirties, was clearly impressed when the researcher attempted to pinpoint the date accurately by asking, 'Was that before or after you moved here from Liverpool?'

The grid as a joint endeavour

3.2

Completing the grid was a joint endeavour which necessitated the co-operation of the respondent. As illustrated in the transcript extract below, this collaborative effort began at the start of the interview, when the researcher began to record lifecourse data at the same time as introducing and familiarising the respondent with the grid. She did this by asking simple questions and verifying dates:

Interviewer: 'How old were you when you left school in Edinburgh?'
R08: '16 '
Interviewer: 'So that would be about 1947? Does that sound right?'
R08: 'Oh I need paper and pencil for that one'.
Interviewer: 'According to the, I've got the grid here, it says 1947, does that sound right?'
R08: 'Yes'.

3.3

Because the respondents had lengthy life courses to recall, in some cases it was helpful to recreate the occurrence and order of a range of events in their lives before turning to their smoking behaviour. The rationale for this approach was that, by the time smoking was introduced, the respondents were used to working with the grid and, with the help of the researcher, had constructed the context in which their smoking biographies could be explored. As we shall see later, however, some respondents preferred to work through their lifecourse in a different way and the grid was equally able to accommodate their preferences.

3.4

The information required from respondents was detailed and, for many, completing the grid was a challenging task. However, like many assignments, which although difficult are ultimately realisable, with assistance, the grid became an enjoyable and achievable task. Putting dates to events and ordering them chronologically was not unlike doing a jigsaw puzzle, where through a slow yet deliberate process a finished picture emerged:

Interviewer: 'So she was born when?'
R01: 'Well I'm nearly five years older than her.'
Interviewer: 'So she was born in 1923, would that be right?'
R01: 'My other sister.'
Interviewer: 'Yes your half sister...'

3.5

By working slowly through the grid the interviewer could cross reference events under the different areas and build up an image of the respondent's lifecourse:

R16: 'I got married in 1957.'
Interviewer: '1957 right, you were 30 ?'
R16: '30 yeah, mm hm.'
Interviewer: 'So you stopped, work when you got married ?'
R16: 'I stopped work when, well I worked for 2 years, because I didn't have David till 1960.'
Interviewer: 'Right so you stopped work then when he was born ?
R16: 'Mm hm, that's right.'

3.6

In the above extract the interviewer entered the date of marriage and child birth in the 'family' column and the date she stopped work under 'employment'. When the various columns were beginning to fill up the grid represented, at a glance, detailed information about the respondent. The depth and particularly the breadth of these data, immediately available to the researcher, constituted an excellent resource on which the interview discussion could draw.

The cross-referencing of events

3.7

We found that respondents tended to reconstruct their lifecourses by recourse to a range of significant life experiences. Because of this the grid was particularly useful in that it facilitated the identification of one event through reference to one or more other meaningful events. For example, as shown in the transcript extract below and depicted on the interview grid for this respondent (see attached grid for respondent R14) the respondent attempted to pinpoint the date of a heart attack by recourse to, among other things, the marriages of her two daughters and her son's death;

Interviewer: 'Right when did you have the heart attack ?'
R14: 'Well my worst one was 68 I think.'
Interviewer: '68, right heart attack then that's when you were..'
R14: '62'
Interviewer: 'Oh right I've got, that's right.'
R14: 'No wait a minute it must have - 71, wait till I calculate this out, my oldest son was killed in 1970.'
Interviewer: 'Oh right.'
R14: '71, I've a feeling I had it then.'
Interviewer: '1971 you would have been 55.'
R14: 'No I'm all mixed up I need to, my oldest son was killed in 1970 I work everything back from that.'
Interviewer: 'Oh right, right.'
R14: 'Alison was married in 1971, Sheila was married in 1979 .... wait a minute 69 no 79, she couldn't have married in 69 don't be ridiculous, if she was born in 81, how old would her oldest daughter be ?'
Interviewer: '17.'
R14: '17, she was born in 79, I had my first heart attack in 81.'
Interviewer: '81 right.'
R14: 'That works, sorry.'

3.8

Often, as shown above, the biography is constructed through reference to personally traumatic events and we have found that respondents deal with this in a very pragmatic way, initially focussing on the outline skeleton of their lives and not dwelling on what are sometimes very painful experiences. Because the respondents used these events themselves as markers it was easier for the interviewer to return to them in more depth later in the interview. It seemed almost as if early acknowledgement of an occurrence granted permission to the interviewer to pursue a more detailed discussion once the grid had been completed.

3.9

Where respondents had difficulty with the chronology of events the grid could be used to help them clarify and confirm the accuracy of dates. In the excerpt below the researcher used the grid data, already entered, to assist the respondent in identifying the exact year she gave up smoking. Although in this instance the respondent did not actually look at the grid herself, on other occasions respondents examined the grid themselves to help solve similar problem;

Interviewer: 'Was this before you had the breast cancer then ?'
R16: 'Just a minute now, no it was after I had the breast cancer. I had the breast cancer in 1990 eh, so I'd be 60, right ?'
Interviewer: '63.'
R16: '60 cause this is 1998. So it's 6 er 8 years since I've had the breast cancer that was 1990.'
Interviewer: 'So you'd have been 63.'
R16: 'That's right I'm getting you all mixed up.'

3.10

Rather than being disconcerted about discrepancies in their accounts, the respondents found the challenge of solving a grid problem quite rewarding;

Interviewer: 'I see from the grid that you stopped smoking when you were living in Leeds, was that right ? You said it was between 64 and 68 ? So it would be the same time you were living in Leeds.'
R19: 'Ah no, here it's a good job you pointed that... actually em, no, I stopped smoking before I went to Leeds, it must have been between 62, 64 then this happened. Sorry about that.'
Interviewer: 'That's fine.'
R19: 'It's just that you bring, I should have had my diary, well that wouldn't be in my diary anyway because I wouldn't show that one.
Interviewer: 'No that's fine, this is the very reason why we use the grids. They are clever.'
R19: 'No you're quite right, I didn't smoke while I was in Leeds. Quite right, yes.'
Interviewer: 'You stopped before then?'
R19: 'Yes I stopped before then.'

3.11

Hence, although sometimes respondents did experience difficulties in detailing the order of events in their lives, when this order was accomplished to their satisfaction it was regarded by both respondent and interviewer with a mutual sense of achievement. The grid therefore not only assisted in obtaining accurate grid data but also contributed to a productive rapport between researcher and respondent. Rapport is very important because the quality of the data in the interview transcripts largely depended upon discussion around different events entered into the grid and associations made between them:

R18: 'Eh 5 years I moved to Edinburgh.'
Interviewer: 'Moved to Edinburgh, right, okay and did you carry on working in the Civil Service ?'
R18: 'Civil Service, yes uh huh.'
Interviewer: 'How long did you work there for ?'
R18: 'Oh dear, 27 years.'
Interviewer: 'So you left in 1977.'
R18: 'That's correct.'
Interviewer: 'Is that right, yeah.'
R18: 'Yes uh huh.'
Interviewer: 'That would be when you were 60, that's when you retired.'
R18: 'That's when I retired, that's correct.'
Interviewer: 'Got it all worked out here you see.'
R18: 'Well done, aye, it's a massive sheet that.'

3.12

To reiterate, although there were five different columns into which life events were recorded, the course of the interview was often directed by the respondent. Hence while some respondents were happy for other areas of their lives to be entered before their smoking histories, others wanted to talk about smoking right away. Irrespective of which course the interview took, the data in one column could be crossed referenced to recorded events in other columns;

Interviewer: 'Now how old were you when you left school ?'
R13: '17.'
Interviewer: '1947 according to my grid. So what did you do when you left school ?
R13: Eh I went into insurance and then I was called up of course, and that's when I started smoking.'
Interviewer: 'Right, right so when were you called up ?' R13: 'In 1948.'
Interviewer: '1948 right and that's when you started smoking ?'
R13: 'Everybody did.'

3.13

During the above exchange, data about employment, national service and smoking were simultaneously entered into the appropriate columns. The researcher could then move across columns and between topics without losing a grasp of the emerging patterns. The grid thus allowed the interview encounter to be very flexible; in effect respondents could reconstruct the lifecourse in the way which they felt was most appropriate.

Accessing life events

3.14

Through using the grid we have found that, while significant events in the area of our research interest were recalled with relative ease, the grid also assisted recall about respondents' smoking behaviour which seemed less readily accessible to them. While for many of our respondents starting and stopping smoking (where applicable) were memorable dates in themselves, the respondent in the extract below reached the latter through its association with another significant life event;

Interviewer: '1981 was when you gave up, right?'
R21: 'Yep, I had my heart attack. Before that I was a heavy smoker'

3.15

Furthermore, the identification of important events, which had implications for smoking, also appeared to facilitate recall about smoking behaviour even where no behaviour change occurred. In the extract below a respondent recalled smoking behaviour when she had been specifically advised to stop for health reasons:

R01: 'Because well, they told me in 1968 of course when I got my first breast off to stop smoking and all that, but I just carried on (smoking).'

3.16

Our data show clearly that life events may have a differential impact on behaviour and how in combination they may ultimately provoke change. Through using the grid respondents were therefore able to assess the relative impacts of different life events on their smoking:

R16: 'I smoked them (super kings) and they were getting more expensive too, but it's when I moved here that I stopped smoking.'
Interviewer: 'When did you move to this house ?'
R16: 'Here, well, 1990 I think it would be, no, no, wait a minute, 1990 em I think, 1994, when I took the first heart attack I was here so it must have been 1993 I was here.'

3.17

As illustrated in the following account of stopping smoking, the respondent drew upon more than one life occurrence. The main reason which he gave for stopping was the expense. The respondent was under financial pressure following the birth of his son and soon after this event, increases in taxation led to increases in the cost of cigarettes:

R27: 'My lad was born on the October, so I gave it up between Christmas and October.'
Interviewer: 'Right, did you'
R27: 'I only smoked the pipe after about twice I think.
Interviewer: 'Really, right, so what was it made you give up ?'
R27: 'Mostly money.'
Interviewer: 'Yeah, Did cigarettes become more expensive around that time ?
R27: 'Oh aye, I mean within a year the tax went on them.'

3.18

Although the grid focuses upon the respondent's particular lifecourse, by probing issues relating to his/her smoking behaviour the researcher was able to locate this in a wider social context or network:

Interviewer: 'When you first tried to give up in the 60's were most of your friends still smoking then or not ? Had any of them given up?'
R10: 'Yes I think they would be (smoking). I don't think any of them gave up. Not until later on.'

3.19

In dating her quit attempt and exploring whether or not the respondent's friends were also attempting to change their smoking behaviour, this exchange led on to a discussion about the cultural climate of smoking at that time and when and how this climate began to change.

Discussion

4.1

In this article we set out with the explicit intention of illustrating some of the more indeterminate aspects of grid interviewing. In other words we have sought to capture and portray a flavour of the interview momentum facilitated by this method. As a result we have concentrated on how the life grid interview is accomplished rather than on the detailed data collection of smoking behaviour around specific events and activities in our respondents' lives.

4.2

We have found the life grid a useful tool for qualitative interviewing in five ways. First, it facilitated recall in the area of the research interest, in this case smoking behaviour, by referencing it to a range of events and experiences in people's lives. Although starting and stopping (where applicable) were often two quite memorable events for smokers, it was more difficult for them to recall their patterns of smoking in between these significant dates. We have found that recall of other, and often non-smoking related, meaningful events in their lives was a useful way for respondents to access less memorable aspects of their smoking biographies. Some, for example, clearly remembered that the onset of a serious illness did not alter smoking behaviour, or that the advice of a GP was disregarded. Likewise, others were able to recall when the price of cigarettes rose, how they 'got by', but still smoked.

4.3

Second, we have found that because the grid involved the co-operation of respondents who collaborated in the reconstruction of their biographies, a high level of respondent engagement was necessitated. This endeavour could be rewarding for respondents because, although at times the task was quite difficult, the completed grid became evidence of successful joint accomplishment. The interviews we have completed demonstrate a high level and good quality of researcher-respondent rapport.

4.4

Third, as Kitzinger (1990) has noted in her work on focus groups, data collection can be made more 'focussed' by asking respondents to perform specific tasks. It can be argued that the life grid method accomplishes for interviewing, what Kitzinger achieves through focused group techniques used, for example, in the study of audience understandings of AIDS media messages. In our interviews we found that it was quite easy to initiate discussion around the different aspects of smoking experience, which were the focus of our study, during the process of grid completion. Furthermore these discussions themselves contributed to the successful accomplishment of the grid.

4.5

Fourth, the grid method appeared to enable respondents to draw on personally traumatic experiences in a way which diffused what were potentially emotionally charged areas. This is particularly pertinent for data collection with respondents who may be currently experiencing severe symptoms. Furthermore, the selection and use of their own life experiences to construct the grid effectively seemed to grant permission for the interviewer to return in more depth to these events and experiences later in the interview.

4.6

Fifth, related to this, is that grid interviewing allowed respondents to take some control over both the course of the interview and in the construction of their biographies. During grid completion respondents were quite likely to discover associations between different events in their lives that they had not previously considered. They made these associations at the same time as the researcher and were able to reflect upon and discuss these issues as they arose. This is quite different from more 'traditional' interviews where the majority of associations are discovered subsequently during analysis.

4.7

Apart from the fact that the grid proves to be a fairly rigorous method for recording life course detail, one of the main attractions for qualitative researchers is its potential to alter traditional interview dynamics. There have been recent moves within qualitative research, and particularly within feminist research, to challenge the distinction upon which social science research has traditionally rested, between the researcher as knowing subject and the respondent as knowable object (Lal 1996). Such challenges start from the premise that social science research is a social interaction in its own right and therefore encapsulates or reflects the social world in which it is situated. For example, feminist researchers argue that if the social world being studied is sexist and hierarchical the processes of research will be sexist and hierarchical as well. For feminist researchers the most central research dilemma is power and the unequal hierarchies or levels of control that are often maintained, perpetuated, created and recreated during and after fieldwork (Wolf 1996). To address this dilemma it becomes necessary to acknowledge the respondent's agency and treat the researched as a subject with whom the researcher engages in a mutual, though unequal, power-charged social relationship (Haraway 1988).

4.8

Although we are not suggesting that using the grid will shift the power balance in favour of the respondent, his/her authority is clearly recognised through the mutual reconstruction of the lifecourse. It is through the achievement of this shared task that the researcher relinquishes some control over the data collection. In so doing the dynamics of the interview are changed as respondents assert influence over their biographical accounts.

4.9

None of the twenty respondents objected to either the grid or the tape recorder. This may be because, as participants in the longitudinal study, they were familiar with research demands. These seasoned respondents were used to researchers not only asking questions but also carrying out clinical examinations. Because our study involved only the interview, respondents found it relatively easy. However, we acknowledge that previous good experiences as participants in the larger longitudinal study probably accounted in large measure for their willingness and enthusiasm for our study. Notwithstanding, all twenty interviews appeared to have been enjoyable encounters for both researcher and respondent. To many of our respondents this came as a surprise. This was illustrated by the respondent who at the end of the interview, looked at the completed grid with some amazement and, said: 'The story of my smoking I didn't think I could fill up as much time as that'.

4.10

Our findings suggest that factors other than recall ability may be as important, if not more so, when collecting data from older respondents. We are very sensitive to the fact that physical impairments will affect the success of data collection and that the respondents should be accommodated accordingly (McKinlay et al 1990). In the life grid interview, although respondents were aware from the outset that there was a task to be completed, the route of accomplishment was negotiable. Respondents could move back and forward in their biographical reconstructions, progressing at a pace which they themselves set. This was a considerable advantage in the collection of retrospective data from older respondents who not only had a longer period to recall, but who were also variously affected by the conditions of poor health.

4.11

In this article we have focussed upon the grid method as a qualitative tool and its impact upon the dynamics of the interview. While we acknowledge that all methods and techniques affect the relationship between respondent and researcher we recognise that quantitative and qualitative approaches do not share the same understandings of the research endeavour. Quantitative or positivistic methodologies tend to distinguish between the objects of study and the methods and techniques which are used to research them, they encourage distance and non-involvement between researcher and researched and assume that the researcher can objectively see, judge and interpret the life and meanings of his/her subjects. Qualitative (and particularly feminist) researchers on the other hand celebrate, rather than diminish, the importance of the relationship between researcher and researched, and the methods themselves are not treated as distinct from the data which are collected.

4.12

Through excerpts from interviews we have therefore attempted to capture, portray and celebrate how the grid method affects the interview dynamics and the implications of these dynamics for the data collected. We also expect the life grid method to have further implications for the way in which we interpret our qualitative data. Our initial analysis suggests that the completed life grids, when read in conjunction with the interview transcripts, may prompt us to explore new ways of understanding and presenting qualitative findings. This, however, must be the subject of future publication.

Acknowledgements

The study on which this paper draws is funded by the ESRC. We also thank members of the Advisory Group to the study: Amanda Amos, Hilary Graham, Paula Holland and Sally Haw for their assistance with the research, and Stephen Platt who commented on an earlier draft.

The Research Unit in Health and Behavioural Change is jointly funded by the Scottish Office Department of Health and the Health Education Board for Scotland. All opinions expressed in the article are those of the authors and are not necessarily those of the funders.

References

ATKINSON, P. (1984) The Clinical Experience: The Construction and Reconstruction of Medical Reality. Farnborough: Gower.

ATKINSON, P., REID, M., & SHELDRAKE, P. (1977) 'Medical Mystiques', Sociology of Work and Occupations, 4(3), pp. 243-281.

BLAXTER, M., & PATERSON, E. (1982) Mothers and Daughters: A three generational study of health attitudes and behaviour. London: Heinemann.

BERNEY, L. R., & BLANE, D. B. (1997) 'Collecting Retrospective Data: Accuracy of recall after 50 years judged against historical records', Social Science and Medicine, 45(10), pp. 1519-1525.

BLANE, D. (1996) 'Collecting Retrospective Data: Development of a reliable method and a pilot study of its use', Social Science and Medicine, 42(5), pp. 751-757.

BROWN, N. R. (1990) 'Organisation of Public Events in Long-Term Memory', Journal of Experimental Psychology, 119(3), pp. 297-314.

CASEY, V. A., DWYER, J. T., BERKEY, C. S., COLEMAN, K. A., GARDNER, J., & VALADIAN, I. (1991) 'Long-Term Memory of Body Weight and Past Weight Satisfaction: A longitudinal follow-Up Study', American Journal of Clinical Nutrition, 1(53), pp. 1493-8.

FIELD, D. (1981) 'Retrospective Reports by Healthy Intelligent Elderly People of Personal Events of Their Adult Lives', International Journal of Behavioural Development, 4, pp. 77-97.

FOX, R. (1957) 'Training For Uncertainty' in R. K. Merton (editor) The Student Physician. Cambridge, Mass: Harvard University Press.

GALLIE, D. (1988) The Social Change and Economic Life Initiative: An Overview, SCELI Working Paper No 1. ESRC Swindon.

HARAWAY, D. (1988) 'Situated knowledges: The science question in feminism and the privilege of partial perspective', Feminist Studies, 14(3), pp. 575-599.

HOLLAND, P., BERNEY, L., BLANE, D., DAVEY-SMITH, G. (1999) 'The Life Grid Method in Health Inequalities Research', Health Variations, 3, pp. 8-9

HUMPHREY, R. (1993) 'Life Stories and Social Careers: Ageing and social life in an ex-mining town', Sociology, 27(1), pp. 166-178.

JAMOUS, H., & PELOILLE, B. (1970) 'Professions or Self Perpetuatory Systems: Changes in the French University Hospital System' in J. Jackson (editor) Professions and Professionalism. Cambridge: Cambridge University Press.

KITZINGER, J. (1990) 'Audience understandsing of AIDS media messages: a discussion of methods', Sociology of Health and Illness, 12(3), pp. 319-335.

KRALL, E. A., VALADIAN, I. M. D., DWYER, J. T., & GARDNER, J. (1989) 'Accuracy of Recalled Smoking Data', American Journal of Public Health, 79(2), pp. 200-202.

LAL, J. (1996) 'Situating Locations, and "Other" in Living and Writing the Text' in D. L. Wolf (editor) Feminist dilemmas in fieldwork. Colerado: Westview Press, pp. 185-213.

LIVSON, N., & McNEILL, D. (1962) 'The Accuracy of Recalled Age of Menarche', Human Biology, 34, pp. 218-221.

McKINLAY, J. B., TENNSTEDT, S. L., BRAMBILLA, D., CASWELL, C., & THORINGTON, B. (1990) 'Adapting Research Methods for Socio-medical Studies of Older People', Advances in Medical Sociology, 1, pp. 299-323.

MUST, A., WILLETT, W. C., & DIETZ, W. H. (1993) 'Remote Recall of Childhood Height, Weight and Body Build by Elderly Subjects', American Journal of Epidemiology, 138(1), pp. 56-64.

PARRY, O. (1993) 'The Reproduction of an Occupation', British Journal of Education and Work, 6(4), pp. 45-56.

POLANYI, M. (1958) Personal Knowledge. London: Routledge and Kegan Paul.

SQUIRE, L. R. (1989) 'On the Course of Forgetting in Very Long-Term Memory', Journal of Experimental Psychology, 15(2), pp. 241-245.

STRAUSS, A. L., GLASER, B., & QUINT, J. (1964) 'The Non-Accountability of Hospital Care', Hospitals, 38, pp. 73-78.

URMSTRON-PHILLIPS, S. (1982) 'The Language Socialization of Lawyers: Acquiring the "Cant"' in Doing the Ethnography of Schooling. Holt, Rineheart and Winston; New York, pp. 176-210

WALTER, F.S., JAMES T, MATANOWSKI, M.D. (1987) 'The Validity of Questionnaire-reported work history in live respondents', Journal of Occupational Medicine, 29 (10), pp. 795-800.

WOLF, D. L. (1996) Feminist Dilemmas in Fieldwork. Colorado: Westview Press.

Copyright Sociological Research Online, 1999