126.96.36.199.1 Introduction Observation, particularly participant observation, is seen as an excellent means of generating ‘rich’ or ‘thick descriptions’ of social phenomena, although not necessarily sufficient for any generalisations or for developing theory. Where research subjects are not susceptible to surveying, observational (ethnographic) techniques might be applied instead to get some information. For example, where potential research subjects are:
1. inarticulate, such as people with learning difficulties, or people who have an infirmity, or young children;
2. resistant to or sensitive about research, such as drug users, criminals, prisoners, sufferers from certain types of illness as well as the powerful élite and celebrities (such as politicians, senior civil servants, company managing directors, pop stars), who have employees screening out non-essential communication or other intrusions;
3. few in number, such as appeal judges, or premiership football mangers;
4. difficult to locate geographically, and therefore not susceptible to surveying, such as traveller communities.
Using observation as a descriptive tool is sometimes referred to as the empirical approach to (participant) observation.
Zelditch (1962), for example, regarded participant observation as an opportunity for in-depth systematic study of a particular group or activity and outlined three elements of an empirical approach:
1. Count the frequency of occurrence of different types of observed behaviour. Ideally this would involve a clear proforma (i.e., schedule or list) that specified types of behaviour linked to predefined hypotheses. An iterative approach would go further and allow for the observation to modify the proforma as the hypotheses themselves are adjusted in the light of the data.
2. Interview informants to establish the social rules that operate in the observed situation and the status of the various participants. Ideally, this would involve systematic sampling of informants to be interviewed, as well as content analysis of documents (see Section 5). It may also involve observation, recorded in formal and systematic ways, of the observed status hierarchy.
3. Participation to observe and provide detailed examples of events and actions that illustrate types of behaviour, status and social rules.
Thus, the empirical approach emphasises systematic observation and recording of data. This remains a concern of those preoccupied with the reliability, validity and generalisability of data, who claim that qualitative study can be made systematic by using qualitative sampling, agreed and consistent ways of doing the fieldwork, team-based approaches and using computers to do the analysis.
188.8.131.52.2 Qualitative sampling Qualitative sampling refers to the particular adaptation of standard sampling principles to qualitative research settings. It differs from the standard positivist approach to sampling, when undertaking quantitative research, which is to find a representative group that would allow empirical generalisations (see Section 1.10).
Instead, qualitative sampling uses available knowledge to identify cases (individuals, organisations, events) that are ‘theoretically rich’ and thus provide data appropriate to the purpose of the qualitative research.
…qualitative researchers perform sampling with a purpose. It is incumbent on the researcher to describe the sample in regards to gender, ethnicity, age, socioeconomic class, and any other relevant criteria so research consumers can understand how and why this sample was chosen. Many strategies influence sample size and selection. The researcher must document the decision-making process involved in qualitative sampling to provide credibility for the research findings. (Byrne, 2001)
In many cases, qualitative sampling is actually what is known as purposive sampling, or theoretical sampling. Often, as ImeldaCoyne (1997, p. 623) notes ‘The terms purposeful and theoretical are viewed synonomously and used interchangeably in the literature’. However, they are not the same.
184.108.40.206.2.1 Purposive sampling Purposive sampling is, in essence, the identification of criteria for section of respondents that fits the purpose of the research. The key is not the number of people interviewed or a random selection but knowing which people or types of people will be able to advance the analysis of the research question.
There are two main ways of doing this. First, by selecting people who reflect the diversity and breadth of the sample population, although the aim is not to produce a
statistically representative sample or draw statistical inference. This approach does though, offer some basis for suggesting indicative generalisations.
Second, by identifying people who may have a particular perspective or knowledge, such as key informants. In some situations these may not be in any way representative but perhaps extreme cases. The idea is that if a situation occurs in an extreme setting, where one would perhaps not expect it, then it is indicative of a general trend.
The Affluent Worker study (see Critical Social Research 2.5) is an example; affluent workers in Luton were observed to see if they adopted middle-class norms and values and it was shown that working-class culture remained strong, thus casting doubts on views on the embourgeoisement thesis (Goldthorpe et al., 1968).
In either event, an understanding of the population being investigated is advantageous when drawing up purposive samples.
Puposive sampling is also sometimes referred to as criterion sampling. In short, the sample is selcted on the basis of some criterion or criteria that ensure the respondents have the potential to provide inforamtion relevant to the study. Maria Menon (2016, pp. 7–8), for example, undertook a study of recently employed graduates.
Criterion sampling was used to identify respondents based on their years of labour market experience. The aim was to interview graduates who would be in a position to provide information on the link between higher education programmes, and content, and labour market productivity. It was considered important for graduates to have some experience in the labour market but at the same time to be in a position to recall the course content element of their higher education experience.... Only one interviewer was used in an attempt to minimise problems associated with the use of interviews for data collection. The interviewer was trained by a member of the research team and was observed during the three pilot interviews.
220.127.116.11.2.2 Theoretical sampling Theoretical sampling might start off purposively but is directed to developing a theory iteratively. The idea is that data is collected and analysed and new notions emerge that lead to the need to collect more empirical data and so on until ‘theoretical
saturation’ is achieved and no new ideas or theories emerge (LeCompte and Priessle, 1993).
This approach is closely associated with ‘grounded theory’. Indeed, Glaser (1978, p. 36) defines theoretical sampling as:
the process of data collection for generating theory whereby the analyst jointly collects, codes, and analyses his data and decides which data to collect next and where to find them, in order to develop his theory as it emerges. This process of data collection is controlled by the emerging theory.
Theoretical sampling of this kind has no fixed limit; there could always be another interesting case to be found.
18.104.22.168.2.3 Qualitative sampling frames Wilmot (2005 p. 1) argues that:
A well-defined sampling strategy that utilises an unbiased and robust frame can provide unbiased and robust results. There is a tendency, particularly within a quantitative environment, to consider that the sampling strategy for qualitative research is of lesser importance to that where statistical inference is required. Indeed, it is not unknown for those unfamiliar with qualitative research methods to suppose that no more than a convenience sampling strategy is applied. That is to say the researcher makes no attempt, or only a limited attempt, to ensure that the sample is an accurate reflection of the population.
Wilmot explained how the UK Office for National Statistics (ONS) used a qualitative ‘Respondent Register’ to ensure robust sampling from a well-constructed sampling frame is used in qualitative social research. The ONS is in a privileged position given its resources and extensive research portfolio undertaken on behalf of government. The ‘Respondent Register’ derived from the National Statistics Omnibus (subsequently the Opinions Survey now (2013) called the Opinions and Lifestyle Survey), which is conducted monthly, covers the whole of Great Britain, and selected the sample on a random basis. As part of the survey, respondents were asked if they would be prepared to be contacted again. Those who agreed were added to the respondent register. Data kept on the sample includes:
Around fifty classificatory variables including both household and individual level information such as household type and composition; tenure; number of cars; respondent age and sex; marital, education, health and employment status; socio-economic classification; ethnic group and income band.
Geographical information such as postcode and postcode sector.
Number of times respondent contacted via the register, number of times respondent participated and details of project participated in.
Strict protocols related to the Data Protection Act are in force.
Individuals are deleted from the register after four years, as their details are likely to be out of date, or if they have been contacted twice in that period for purposes of qualitative surveys.
The major benefit of having a qualitative potential respondent register such as this is the considerable efficiency gains from having a sampling frame to select from. However, the decision to use the register depends on the research objectives ‘and requires an understanding of the strengths and weaknesses of the register’ (Wilmot, 2005, p. 10). Advantages included a large, diverse, up-to-date, constantly refreshed list (45000 cases), with contact telephone numbers (in 96% of cases), many demographic and classificatory variables that can be matched with research section criteria, and can, if necessary be clustered to minimise qualitative interviewer travel time. In addition, as respondents have already been interviewed for the Omnibus survey they are likely to be more amenable, require less by way of introduction, be relatively easy to locate and be safe to interview. The Register was used, for example, in recruiting respondents to the ONS study entitled Developing survey questions on sexual identity:
Cognitive/in-depth interviews(Betts, 2009).
However, despite is diversity, there is a potential bias in the register: it is, to some extent self-selecting as it is only 80% of those who responded to the Omnibus survey, which itself has a response rate, on average, of 65%. It is also limited to people aged 16 years or over, living in private households in Great Britain. Those living in institutions such as old peoples’ homes are excluded. Numbers of some subgroups of the population may still be small despite the size of the register. Nonetheless, this would be a resource that any qualitative researcher would appreciate.
Can you think of any other disadvantages of the register? This activity involves critical reflection. Click here for an outline answer
22.214.171.124.3 Agreed and consistent ways of doing fieldwork Developing agreed and consistent ways of doing fieldwork is another way that qualitative research can supposedly made more systematic. Using agreed ways of doing fieldwork simply means that observation or in-depth interviewing is guided by an agreed schedule of questions, or proforma of aspects to observe. This has characterised much observational research from whatever perspective.
Even when agreed procedures are in place, though, the situation of every encounter is unique and the schedule or proforma only provides a framing guide. If the area is not well known in advance, the proforma will provide only a limited aid and will have to be refined as the research unfolds. In some cases, the whole point is to get the subject’s perspectives, as discussed below, in which case it will be difficult to use consistent approaches unless the intention is to constrain the boundaries around which notes will be made about what subjects say or do.
126.96.36.199.4 Teams to undertake the qualitative research Using teams to undertake the qualitative research allows for a wider geographic spread of observation but the same issues apply about the agreed and consistent approach to field work.
188.8.131.52.5 Computer-based analysis packages Computer-based packages to analyse qualitative data are sometimes assumed to enhance the 'objectivity' of the date. However, such packages are no more or less inherently systematic than paper-based analyses. They do nothing more than aid coding, retrieval, sorting and grouping of data. They are usually quicker and easier to use with large data sets than shuffling bits of paper.
However, the data has all to be input into the computer package, which itself can be time-consuming. Unless care is taken with computer coding, data might be classified differently at different times by by different data inputters. Having input the data, the speed and relative ease of use of computer qualitative data analysis packages may lead to more systematic analysis simply because they are less exhausting to use than shuffling paper. On the other hand, there may be a temptation to ‘let the computer do the work’ in effect assuming that the coding frame and the computer-sorting algorithms will provide the outcomes without the researcher having to keep conceptually alert.
Answer to Activity 3.3.1: Although refreshed each month and containing large numbers of potential respondents, the register may, in the longer term suffer from over use and resulting bias. This is because respondents are removed from the register after taking part in research on two occasions. So if a number of studies are carried out using the same selection criteria this could deplete the frame of certain types of respondent.