Sampling
One of the first issues in qualitative research is deciding who to interview. The decision will depend, to some extent, on the nature of the study. If one is doing case-studies, one will want to interview a wide range of people involved in those cases. For example, a series of case-studies on the ways in which different congregations had integrated welfare work with their congregational activities involved interviewing people in charge of those activities, people engaged in providing welfare and in organising congregational activities, and people who had participated in the various programs and activities.
People in different parts of an organisation will have different interests and different concerns and it will usually be important to listen to the range of these. Clergy, for example, usually have quite different interests and concerns from the lay leaders of a congregation. Members of the congregation who have no ‘official’ position offer different opinions again. People at the core of an organisation are often protective of the organisation. People on the fringes may be more interested in what are the benefits or costs of the organisation to their own circumstances. The people in charge may give a very different view on ‘how things are going’ from participants or observers. To get a full picture, from a research perspective, it will usually be important to listen to a variety of perspectives.
If the focus is a particular topic, such as worship or Christian education, then one must decide on the exact group about which one would like to draw conclusions. One should keep in mind, for example, that many people may have attended worship at a particular church in the past but may have withdrawn because they did not enjoy it or find it fulfilling. Thus, the remaining group of people will be a ‘biased sample’: the group which has remained committed to the patterns. The fact that almost the whole congregation enjoys traditional Christian music does not mean that the church should only use traditional Christian music, for example. Contemporary music may draw quite a different crowd altogether.
Nevertheless, there will be times when wants to examine a group such as: ‘those who worship regularly at a particular church’ or ‘the people who have spent time with a hospital chaplain within the last month’ or ‘the students who attend a particular college’. If one wishes to obtain information about such a group, it may be wise to choose people randomly to ensure that no particular biases enter the choice of who is interviewed and who is left out. In general, it is desirable to interview as many people as possible if randomly selecting them and desiring to draw conclusions about the whole group. Every individual will have their own perspective and their own story, and when the factors which affect people’s experiences are many and various, the greater the number of people in the sample the better. If the total group contains a hundred people, it would be desirable to interview seventy-five of those people if one had the resources to do it. If one finds over a number of interviews that no new information is becoming available, then it may be appropriate to reduce the number of interviews. On the other hand, it is generally considered that 1000 interviews are sufficient to gather a reliable sample of the population of a nation as long as one is looking for a relatively simple range of information. (See also the discussion of sampling in relation to quantatitive research.)
In order to draw a random sample, number each person (or congregation, or group) and then use the random number table included on this CD-Rom or use a random number generator in a spreadsheet program such as QuattroPro or Excel to ‘pull out’ the numbers of your sample.
However, in qualitative research, researchers are commonly most interested in understanding something of the range of responses or situations and in the ways in which the various factors inter-relate. In such projects, it is more appropriate to choose a stratified sample, that is, a sample which includes a particular range of sub-groups or strata. For example, a project on the ways in which Christian education occurs and what forms of Christian education are found to be most helpful might examine Christian education in a range of congregations. It would be important to look at large and small congregations, congregations in ‘professional’ areas and congregations in ‘working-class’ areas, rural and urban congregations, congregations where young people dominated and congregations made up mostly of older people.
If possible it might be helpful to distinguish between theologically liberal and theologically conservative congregations. Ideally, one may develop a multi-dimensional table of the various categories or strata, and seek to find several examples in each cell of the table. The following table, based on just three characteristics of congregations gives us a total of twelve categories of congregation.
A Table for Identifying Congregations for a Study | Rural | Urban | ||
---|---|---|---|---|
Working-class | Professional | Working-class | Professional | |
Large (Over 250 people) | ||||
Medium (100 to 250 people) | ||||
Small (less than 100 people) |
In identifying relevant ‘strata’, one would need to think through all the possible factors which may have a bearing on the topics of research. One would then look for people or groups which provided one with the possible range in relation to each factor.
Sometimes the sample is ‘given’. If one is examining the impact of a particular program, the sample will be the people who have participated in the program. Or one may be interviewing people who have had a particular experience such as having studied overseas, or who are married to people from a different religious background from their own. Grounded theory research seeks such a group of people who can contribute to the research, who can reflect on their experiences and their situation and who can contribute to the development of theory.
In some instances, it is helpful to find a ‘control sample’, a group of people who have not been through a program or have not had the relevant experience but who can talk about their own situations. For example, if one is examining the impact of divorce on the formation of new relationships, one might talk to a range of divorced people. However, it may also be helpful to talk to some people who are married and who have not been divorced about their relationships within and outside of marriage to provide some comparisons.
However, it is not always easy to find a sample. One may know very few people, for example, who are married to people from different religious backgrounds. In such situations, a snow-balling technique is often used. One asks the initial interviewees if they can suggest other people who fall into the category and who may be interested in sharing their experiences. For ethical reasons, it is generally better if the initial interviewees are prepared to ask their friends first before passing on their names to the researcher who will then contact them with more information and arrange the interviews.
Setting Up the Interview
Privacy laws have made it more difficult for the interviewer to set up interviews. One is dependent on the organisation or people who have the initial contacts to make the first contact and determine whether potential interviewees are happy for their names and contact details to be passed on to the researcher. Only then can the researcher make contact. For example, a church should not pass on to a researcher the names and contact details of people without their expressed permission.
When permission for contact is received, then the researcher may write to the interviewee and explain the project to them. Most ethics committees consider that a phone call from the researcher is not enough. Written information about the research project should be provided to the interviewee. Through the letter, the researcher will want to encourage participation while making it clear that involvement is entirely voluntary. The letter should include:
- details of the project
- who has auspised it
- who will be conducting the interview
- whether it is planned to use tape recordings
- how the information will be used
- what information will be published
- how the anonymity of the interviewee will be protected, including details of storage of tape recordings, for example
- what sort of time and any other costs will be required of the interviewee
- whether there are any risks involved
- where further information can be gained about the research project itself
- what ethics committee has given approval for the project
- how comments or complaints can be made directly to the ethics committee.
With the letter should be sent a consent form designed by the researcher and approved by the ethics committee in which the interviewee signs a statement indicating willingness to participate in the research, willingness for the interview to be recorded, and willingness for the information to be used in reports in such a way as will protect the anonymity of the interviewee.
When interviewing minors, there should be separate consent forms for guardians or parents. Note that the legal definition of a minor varies from one State to another in Australia. An interview conducted with any child under the age of 15 years should certainly be approved by a parent or guardian. Interviews which occur in a school, for example, will also require the permission of the school and possibly the school system. It will also be necessary to obtain police checks for each of the interviewers.
In encouraging people to participate, it is generally appropriate to explain how important their contribution would be. People are usually happen to assist if they feel that their contribution will be taken seriously and it will actually benefit organisations, structures or other people in a way they believe is important.
While there is a lot of information that needs to be relayed to people prior to an interview, it needs to be phrased in simple but non-patronising language. Technical language contained in the letter in the explanation of what the project is about will not encourage people to take part.
Finding a suitable location for an interview is not always easy. In some projects, homes can be used if the interviewee prefers that. Public spaces like cafes are usually too noisy and not sufficiently private. It is often most suitable if a space can be found in a cooperating organisation’s offices. A church may have a room which can be used. A library may have an area for conversations.
Interviews can be conducted over the telephone. This is being done increasingly with large projects and there are many private research and polling organisations which use this method. Software is available for making recordings of telephone interviews which uses information provided early in the interview to ‘personalise’ each interview. Again, it is important that, at the start of any interview, the interviewee is informed about the nature of the project, how long the interview will take, and how the information will be used. The interviewee must be informed that their participation is entirely voluntary. Interviewers usually offer to phone back if the present time is not convenient.
Telephone interviews save a lot of time in travel. However, they have some limitations. It is more difficult to develop rapport with the interviewee and it is unusual for interest to be maintained for telephone interviews for longer than about twenty minutes. It is also impossible to see all the body language of ther interviewee, and, to that extent, the amount of information available to the interviewee about the situation and about the responses of the person is not as great as in face-to-face interviews. Hence, telephone interviews are very suitable for polls or for relatively short, focussed interviews, but are not suitable for in-depth interviewing.
Interview Questions
Interview questions should be prepared well in advance of the interviews and it is usual for ethics committees to require copies of the interview schedule to be submitted and examined by the committee prior to permission being given for the project to proceed. Questions may vary greatly in how open-ended and how specific they are according to the purpose of the project. For example, in descriptive and exploratory research it may be helpful to have very general questions in order to find ‘what is out there’. In narrative research in which the interviewer wants to hear about experiences and about how those experiences are understood, some general questions may be used to encourage the interviewee to ‘tell their story’. On the other hand, when research is focussed by specific hypotheses or where detailed information is required, the questions may be much tighter.
In some interviews, including many telephone interviews, the questions are similar to those used in surveys. One technique used by interviewers which replicates the sorts of questions used in questionnaires is to give the interviewee a card with a range of options printed on it. The interviewer then asks the interviewee to read through the options and indicate which option is their best answer.
It is not easy to prepare good questions. The following are some desirable characteristics. Ideally, questions should be:
- linguistically simple, not using technical language which some interviewees will not be able to understand
- conceptually simple in that they require one set of information, rather than being double-barrelled, and avoiding the use of double negatives
- neutral in as far as they do not presume or lead interviewees into a particular response
- non-judgemental in that they do not contain or even suggest that one way of doing things is right and another wrong.
In terms of ordering questions, some researchers prefer to start with broad, general questions before focussing on more specific and personal ones. One technique is to ask people about how other people think or act before asking them how they do so themselves. Questions about activities are generally easier to answer than questions about ideas. Hence, it is often best to start with questions about activities then move to ideas later. Giving people some very difficult questions at the start of the interview may be discourage them from continuing with the interviewer and will make it harder to build rapport.
On the other hand, it is good to start with questions about the topic which the interviewee understands as the subject of the interview. If they feel that they making a significant contribution from the beginning and that they have something worthwhile to say that will be of interest to the interviewer, they will be encouraged to proceed.
Do not start with demographic questions about the interviewee’s age or occupation. Some researchers prefer to leave these until last. However, it is generally better if they can be slotted into the relevant sections of ther interview. For example, in talking about the interviewees work, one may ask, ‘Tell me, how exactly do you describe your occupation?’ In talking about family life, it may be appropriate to ask how many children they have.
In narrative methods, the interviewer will encourage the interviewee to ‘tell their story’. There may be just a very few questions with a few extra prompts. For example, the interviewer may ask ‘Tell me how you came to be involved in this activity / organisation’. The prompts may include:
- what happened before that?
- can you tell more more about that?
The task of the interviewer will be to listen to the story, and to re-focus the interviewee from time to time. The interviewer may add questions regarding whether the interviewee thinks the same way now, or how they feel about what happened. It should be noted that, in most interviews, the task is not to discover ‘the truth’ as such, but rather ‘the truth as the informant sees it’. The interviewer is not there to quiz the interviewees or to help them get their stories correct, but to listen to their personal viewpoints.
The Interview Itself
The body language and stance of the interviewer are most important in the success of the interview. The flow of information is dependent upon trust being established. In order for this to occur, the interviewer cannot be neutral even though the questions should not presuppose particular answers. Rather, the interviewer should show themselves to be interested, focussed and involved. In small ways, it is generally helpful to be affirming of the interviewees responses in the sense that the interviewee needs to feel that they are understood and appreciated. The interviewee needs to feel that the interviewer is entering their world, appreciating their circumstances, and understanding their responses.
Negative body language which indicates impatience or judgement can destroy the interview. Body language in the tight folding of arms and frowning will suggest that the interviewer is closed to what the interviewee has to say. At the same time, interviewers need to be seen as ‘genuine’ in their interest in what the interviewer has to say. To ‘over act’ the friendliness can also be counter-productive.
While the interviewer will follow the schedule or list of questions which has been prepared, the interviewer should be prepared with some prompts to assist in further exploring the responses. Possible nudges include:
- Tell me more.
- Is that so?
- Please continue.
- What was in your mind at the time?
- What happened after that?
- I’m not sure I understood your point.
- How did you react?
- Is that the way you think now?
- Is that what usually happens?
How, when, and what questions may be useful for exploring a situation more deeply. ‘Why’ questions may be used too, but more sparingly as they can be taken as questioning the integrity of the interviewee.
The interviewee may check on information given by saying ‘To summarise, you began by telling me … I have I undestood you correctly?’
If the responses are confusing, the interviewer should only hint at that indirectly: ‘I am a little confused. Can you clear this up for me?’
The interviewer may like to signal the end of the interview by saying something like ‘Just a couple more questions …’ or ‘Is there anything else you would like to add before we conclude?’ or by asking the interviewee if they have any questions they would like to ask. In concluding it is important that the interviewer does not heave a sigh of relief that it is all over, but continues to express thanks for the contribution of the interviewee:
‘Thank you for your time. I enjoyed our discussion and I appreciated your insights. I will send you a copy of the report when it has been written.’
The interview can be recorded in one or both of two ways. Most researchers like to record interviews using a small, unobtrusive tape-recorder, having first received permission from the interviewee to use one. The advantage is that the exact words are recorded. Even the ambiguities of the statements are available to the researcher for further reflection. Recording interviewees is particularly important if the focus of the interviewee is the interviewee themselves, how the interviewee thinks and feels, how they experience life or categorise their experiences. On the other hand, some interviewees feel inhibited by the use of tape recorders which capture their exact statements ‘for life’.
In some research projects, the focus is rather on the information that the interviewee can provide, for example, about a particular organisation, activity or program. In such cases, the exact words that the interviewee uses are not quite so important. In such situations, it may be sufficient for the interviewer to make notes during the interview recording the information that is sought. Notes can be analysed much more quickly. Doing a full transcription of lengthy interviews is a very time-consuming occupation if one does it oneself, and very costly if one gets others to do it. In writing notes, the interviewer actually begins the process of categorising responses. The interviewer can also make notes about body language and the context of the interview, about interruptions or times when the focus of the interviewee wanders. Nevertheless, writing full notes may mean that the interviewer spends more time looking down than looking at the interviewee, and may make it more difficult to establish rapport.
Content Analysis
Content analysis is used in many contexts within empirical research. For example, one might do content analysis of sermons to discover what are the recurring themes and the patterns in the ways that themes are developed. One might do content analysis on articles in church magazines or on letters to the editor. Content analysis is an important part of the processes of analysising individual or focus group interviews.
While content analysis is usually used with written word-based material, techniques of content analysis can be used with graphical material, sounds, pictures, film, or any other media. The usual practice, however, is to conduct content analysis on written material. It is common to seek to reduce other materials, such as interviews, to full written transcriptions, prior to the process of content analysis.
Transcribing interviews is a long and often expensive process. It can only be done if one has previously recorded the interviews and permission to do that depends on the interviewee. Transcribing is important if one is looking for more than factual information provided by the interviewee. If the research is seeking to understand how people think, then the exact words that are used and the ways in which they are used will be important clues. If, on the other hand, the research is about patterns of attendance in small groups then the words used in conveying the information will not be so important. In such cases, transcribing the interviews will not be necessary. Indeed, in such cases, notes written during the interview may be quite sufficient.
There are some computer programs to assist in content analysis. One such program developed in Australia at LaTrobe University and widely used has come in several forms, including NUDIST and NIVO. Another program is included on this CD-Rom in a trial form from SYSSTAT.
In brief, these programs allow one to place codes in the text and categorise text. They then allow one to retrieve the text related to particular codes and to look at where certain codes are used in the vicinity of each other. For example, supposing one was doing content analysis of hymns and was interested in learning how hymn – writers have understood salvation. One might code all references to salvation in the collection of hymns – looking not only at the explicit references to ‘salvation’ and associated words, but at references which used other words such as ‘redemption’. One might then ask the program to list all the references to salvation and look at proportion of them which were focussed on individual salvation and the proportion focussed on the salvation of groups of people such as nations, the Church, or all humanity. One might look further at the means by which salvation was envisaged to occur such as the parts played by Jesus Christ and the Holy Spirit. One might look at how salvation was linked in the body of hymns to the Second Coming.
Computer programs are not necessary for simple forms of content analysis when the number of pieces of text (interviews, sermons, letters, or whatever) are few in number. Content analysis can be done by hand, by reading through each piece of text and noting whether particular themes, ideas or expressions appear in the text.
In content analysis, there must be a balance in the framework of questions you bring to the materials and in the openness of the analysis to what the materials hold. That balance will vary according to the purpose and focus of the research. Some research projects are interested in what people experience in certain types of situations, such as in the process of retirement from work. Such research projects will be open to a wide variety of expressions. Other research frameworks will be highly focussed on particular questions. For example, in examining the benefits and limitations of a particular program of preparation for a sacrament, such as Baptism, one would bring a tight framework to the interview materials and it may be appropriate to disregard material which is not relevant to the issue of the benefits and limitations of that particular program.
In cases where the framework is open, such as people’s experience of retirement, it may be appropriate to begin the process of analysis by reading through the materials and looking for the sorts of themes which emerge. A list can be made of such themes. In the subsequent readings, careful counts can be kept of the various themes. It may be helpful to note those counts under various headings such as the sorts of occupations from which people have retired. One may be interested in whether the experiences of males and females is different, or people living alone or living in partnerships.
Where the framework for analysis is relatively closed, one may begin the process by making notes and counts within the established framework. For example, one may look through the materials noting the various benefits and limitations of a particular program.
There are a range of things one may look for in content analysis:
- Themes and patterns in the responses
- Patterns of relationships in the themes
- Components in the themes and patterns
- Contexts in which certain themes emerge
- How the themes and patterns vary among different groups of people
- How the themes and patterns relate to different types of situations
- The exceptions to the patterns noted
In looking at the themes and patterns, the researcher is seeking to develop
- typologies in the material
- identify relationships and dynamic structures
- seek for patterns which might provide the basis for wider generalisations beyond the specific material that is being examined.
It remains true that, in content analysis, as in most types of research, the theories and hypotheses that one brings to the process will have a major impact on the success of the research. The hypotheses provide the basis on which the researcher proceeds. In many instances, research can do little more than support or not support certain hypotheses. It is in the hypotheses that the imagination and understanding of the researcher play important roles.