There is a need for more explicit discussion of qualitative sampling issues. This article will outline the guiding principles and rationales, features, and practices of sampling in qualitative research.
It then describes common questions about sampling in qualitative research. In conclusion it proposes the concept of qualitative clarity as a set of principles analogous to statistical power to guide assessments of qualitative sampling in a particular study or proposal.
Questions of what is an appropriate research sample are common across the many disciplines of gerontology, albeit in different guises. The basic questions concern what to observe and how many observations or cases are needed to assure that the findings will contribute useful information.
Throughout the history of gerontology, the most recognized and elaborate discourse about sampling has been associated with quantitative research, including survey and medical research. But concerns about sampling have long been central to social and humanistic inquiry e. The authors argue such concerns remained less recognized by quantitative researchers because of differing focus, concepts, and language.
Recently, an explicit discussion about concepts and procedures for qualitative sampling issues has emerged. Despite the growing numbers of textbooks on qualitative research, most offer only a brief discussion of sampling issues, and far less is presented in a critical fashion Gubrium and Sankar ; Werner and Schoepfle ; Spradley , ; Strauss and Corbin ; Trotter ; but cf.
The goal of this article is to extend and further refine the explicit discussion of sampling issues and techniques for qualitative research in gerontology. Throughout the article, the discussion draws on a variety of examples in aging, disability, ethnicity as well as more general anthropology. The significance of the need to understand qualitative sampling and its uses is increasing for several reasons. First, emerging from the normal march of scientific developments that builds on prior research, there is a growing consensus about the necessity of complementing standardized data with insights about the contexts and insiders' perspectives on aging and the elderly.
These data are best provided by qualitative approaches. In gerontology, the historical focus on aging pathology obscured our view of the role of culture and personal meanings in shaping how individuals at every level of cognitive and physical functioning personally experience and shape their lives.
A second significance to enhancing our appreciation of qualitative approaches to sampling is related to the societal contexts of the scientific enterprise. Shifts in public culture now endorse the inclusion of the experiences and beliefs of diverse and minority segments of the population. A reflection of these societal changes is the new institutional climate for federally funded research, which mandates the inclusion and analysis of data on minorities.
Qualitative approaches are valuable because they are suited to assessing the validity of standardized measures and analytic techniques for use with racial and ethnic subpopulations. They also permit us to explore diversities in cultural and personal beliefs, values, ideals, and experiences. It describes the scientific implications of the cultural embeddedness of sampling issues as a pervasive feature in wider society. It concludes by proposing an analog to statistical power, qualitative clarity , as a set of principles to guide assessments of the sampling techniques in a study report or research proposal.
The term clarity was chosen to express the goal of making explicit the details of how the sample was assembled, the theoretical assumptions, and the practical constraints that influenced the sampling process. Qualitative clarity should include at least two components, theoretical grounding and sensitivity to context.
The concept focuses on evaluating the strength and flexibility of the analytic tools used to develop knowledge during discovery procedures and interpretation. These can be evaluated even if the factors to be measured cannot be specified. A wide range of opinions about sampling exists in the qualitative research community. The authors take issue with qualitative researchers who dismiss these as irrelevant or even as heretical concerns.
The authors also disagree with those quantitative practitioners who dismiss concerns about qualitative sampling as irrelevant in general on the grounds that qualitative research provides no useful knowledge. It is suggested that such a position is untenable and uninformed. This article focuses only on qualitative research; issues related to combined qualitative and quantitative methods are not discussed. The focus is on criteria for designing samples; qualitative issues related to suitability of any given person for research are not addressed.
Criteria Two issues relate to the individual subjects in terms of cooperativeness, rapport, and suitability for qualitative study methods. Although this article may appear to overly dichotomize qualitative and quantitative approaches, this was done strictly for the purposes of highlighting key issues in a brief space. The authors write here from the perspective of researchers who work extensively with both orientations, singly and in combination, in the conduct of major in-depth and longitudinal research grants that employ both methods.
It is the authors' firm belief that good research requires an openness to multiple approaches to conceptualizing and measurement phenomena. Attention to sampling issues has usually been at the heart of anthropology and of qualitative research since their inception. Much work was devoted to evaluating the appropriateness of theory, design strategies, and procedures for sampling. Important contributions have been made by research devoted to identifying and describing the nature of sample universes and the relevant analytic units for sampling.
Kinship studies aim to determine the fundamental culturally defined building blocks of social relationships of affiliation and descent e. Ethnographic investigations document the diversity of kinship structures, categories of kith and kin, and terminologies that give each culture across the globe its distinctive worldview, social structure, family organization, and patterns to individual experiences of the world.
Concerns with sampling in qualitative research focus on discovering the scope and the nature of the universe to be sampled. Thus the important contributions of qualitative work derived from concerns with validity and process may be seen as addressing core concerns of sampling, albeit in terms of issues less typically discussed by quantitative studies.
Two examples may clarify this; one concerns time allocation studies of Peruvian farmers and the other addresses a census on Truk Island in the South Pacific.
The Andes mountains of Peru are home to communities of peasants who farm and tend small herds to garner a subsistence living. To help guide socioeconomic modernization and to improve living conditions, refined time allocation studies see Gross were conducted in the s to assess the rational efficiency of traditional patterns of labor, production, and reproduction.
Seemingly irrational results were obtained. A systematic survey of how villagers allocated their time to various activities identified a few healthy adults who sat in the fields much of the day.
Only after interviewing the farmers to learn why the men sat in the fields and then calculating the kilocalories of foods gained by putting these men to productive work elsewhere was an explanation uncovered. It was discovered that crop yields and available calories would decline , not increase, due to foraging birds and animals. Because the farmers sat there, the events of animal foraging never occurred in the data universe.
Here, judgments about the rationality of behaviors were guided by too narrow a definition of the behavioral universe, shaped by reliance on analytic factors external to the system e. An important message here is that discovery and definition of the sample universe and of relevant units of activity must precede sampling and analyses.
On Truk Island in the South Pacific, two anthropologists each conducted an independent census using the same methods. They surveyed every person in the community. Statistical analyses of these total universe samples were conducted to determine the incidence of types of residence arrangements for newlywed couples.
The researchers reached opposite conclusions. Goodenough argued that his colleague's conclusion that there are no norms for where new couples locate their residence clearly erred by classifying households as patrilocal near the father , matrilocal, or neolocal not near either parent at one time as if isolated from other social factors.
Goodenough used the same residence typology as did his colleague in his analysis, but identified a strong matralineal pattern wife's extended family. Evidence for this pattern becomes clear when the behaviors are viewed in relation to the extended family and over time. The newlyweds settle on whatever space is available but plan to move later to the more socially preferred e. Thus different analytic definitions of domestic units led to opposite conclusions, despite the use of a sample of the total universe of people!
Social constructions of the lived universe, subjectively important temporal factors have to be understood to identify valid units for analyses and interpretation of the data. The Peruvian and the Truk Island examples illustrate some of the focal contributions of qualitative approaches to sampling. Altering the quantitatively oriented sampling interval, frequency, or duration would not have produced the necessary insights. The examples also suggest some of the dilemmas challenging sampling in qualitative research.
These will be addressed in a later section. Both cases reveal the influence of deeply ingrained implicit cultural biases in the scientific construction of the sampling universe and the units for sampling. Sampling issues are not exclusive to science. Widespread familiarity with sampling and related issues is indicated by the pervasive popular appetite for opinion and election polls, surveys of consumer product prices and quality, and brief reports of newsworthy scientific research in the mass media.
Sampling issues are at the heart of jury selection, which aims to represent a cross section of the community; frequent debates erupt over how to define the universe of larger American society e. We can shop for sampler boxes of chocolates to get a tasty representation of the universe of all the candies from a company. Debates about the representativeness, size, and biases in survey results because of the people selected for study or the small size of samples are a part of everyday conversation.
Newspapers frequently report on medical or social science research, with accounts of experts' challenging the composition or size of the sample or the wording of the survey questions. Critical skills in sampling are instilled during schooling and on-the-job training. Such widespread familiarity with basic sampling issues suggests a deep cultural basis for the fascination and thus the need for a more critical understanding.
The concept and practices of sampling resonate with fundamental cultural ideals and taboos. It is perhaps the case that sampling is linked, in American culture, to democratic ideals and notions of inclusion and representation. What does that mean for qualitative researchers designing sampling strategies? We need to be aware that the language of science is ladened with cultural and moral categories.
Thus gerontological research may potentially be shaped by both cultural themes masked as scientific principles. Basic terms for research standards can simultaneously apply to ideals for social life Luborsky We construct and are admonished by peers to carefully protect independent and dependent variables; we design studies to provide the greatest statistical power and speak of controlling variables.
At the same time, psychosocial interventions are designed to enhance these same factors of individual independence and senses of power and control. We examine constructs and data to see if they are valid or invalid; the latter word also is defined in dictionaries as referring to someone who is not upright but physically deformed or sickly. Qualitative research, likewise, needs to recognize that we share with informants in the search for themes and coherence in life, and normatively judge the performance of others in these terms Luborsky , b.
The ideals of representativeness and proportionality are not, in practice, unambiguous or simple to achieve as is evidenced in the complex jury selection process. Indeed, there is often more than one way to achieve representativeness. Implicit cultural values may direct scientists to define some techniques as more desirable than others.
Two current examples illustrate how sampling issues are the source of vitriolic debate outside the scientific community: Lani Guinier, was withdrawn, in part, because she suggested the possibility of an alternative voting system giving citizens more than one vote to cast to achieve proportional representation for minorities. We see in these examples that to implement generalized democratic ideals of equal rights and representation can be problematic in the context of the democratic ideal of majority rule.
Another example is the continuing debate in the U. Supreme Court over how to reapportion voting districts so as to include sufficient numbers of minority persons to give them a voice in local elections.
These examples indicate the popular knowledge of sampling issues, the intensity of feelings about representativeness, and the deep dilemmas about proportional representation and biases arising within a democratic society.
The democratic ideals produce multiple conflicts at the ideological level. It is speculated that the association of sampling issues with such core American cultural dilemmas exacerbates the rancor between qualitative and quantitative gerontology; whereas in disciplines that do not deal with social systems, there is a tradition of interdependence instead of rancor.
For example, the field of chemistry includes both qualitative and quantitative methods but is not beset by the tension found in gerontology. Qualitative chemistry is the set of methods specialized in identifying the types and entire range of elements and compounds present in materials or chemical reactions.
A variety of discovery-oriented methods are used, including learning which elements are reacting with one another. Quantities of elements present may be described in general ranges as being from a trace to a substantial amount. Quantitative chemistry includes measurement-oriented methods attuned to determining the exact quantity of each constituent element present. Chemists use both methods as necessary to answer research problems.
The differences in social contextual factors may contribute to the lower level of tension between quantitative and qualitative traditions within the European social sciences situated as they are within alternative systems for achieving democratic representation in government e. The preceding discussion highlighted the need to first identify the ideal or goal for sampling and second to examine the techniques and dilemmas for achieving the ideal. The following section describes several ideals, sampling techniques, and inherent dilemmas.
Core ideals include the determination of the scope of the universe for study and the identification of appropriate analytic units when sampling for meaning. This is simultaneously one of qualitative research's greatest contributions and greatest stumbling blocks to wider acceptance in the scientific community. As the examples of the Peruvian peasants and Trukese postmarital residence norms illustrated, qualitative approaches that can identify relevant units e.
Probability-based approaches do not capture these dimensions adequately. The problem is that the very nature of such discovery-oriented techniques runs counter to customary quantitative design procedures. This needs to be clearly recognized. Because the nature of the units and their character cannot be specified ahead of time, but are to be discovered, the exact number and appropriate techniques for sampling cannot be stated at the design stage but must emerge during the process of conducting the research.
One consequence is that research proposals and reports may appear incomplete or inadequate when in fact they are appropriately defined for qualitative purposes. One technique in writing research proposals has been to specify the likely or probable number of subjects to be interviewed.
Evidence that a researcher devoted sufficient attention to these issues can be observed in at least two dimensions. First, one finds a wealth of theoretical development of the concepts and topics. In qualitative research, these serve as the analytic tools for discovery and aid in anticipating new issues that emerge during the analyses of the materials. Second, because standardized measurement or diagnostic tests have not yet been developed for qualitative materials, a strong emphasis is placed on analytic or interpretive perspectives to the data collection and data analyses.
Expository styles, traditional in qualitative studies, present another dilemma for qualitative discussions of sampling. The traditional format for grant applications places discussions of theory in the section devoted to the general significance of the research application separate from the methods and measures.
However, theoretical issues and conceptual distinctions are the research tools and methods for qualitative researchers, equivalent to the quantitative researchers' standardized scales and measures. Qualitative researchers look for the analytic refinement, rigor, and breadth in conceptualization linked to the research procedures section as signs of a strong proposal or publication.
Thus basic differences in scientific emphases, complicated by expectations for standardized scientific discourse, need to be more fully acknowledged. The logic or premises for qualitative sampling for meaning is incompletely understood in gerontology.
At the same time, and partly in reaction to the dominance of the quantitative ethos, qualitative researchers have demurred from legitimating or addressing these issues in their own work. Understanding the logic behind sampling for meaning in gerontological research requires an appreciation of how it differs from other approaches. By sampling for meaning, the authors indicate the selection of subjects in research that has as its goal the understanding of individuals' naturalistic perceptions of self, society, and the environment.
Stated in another way, this is research that takes the insider's perspective. Clearly, the qualitative approach to meaning stands in marked contrast to other approaches to assessing meaning by virtue of its focus on naturalistic data and the discovery of the informant's own evaluations and categories.
For example, one approach assesses meaning by using standardized lists of predefined adjectives or phrases e. The difference between the me of that night and the me of tonight is the difference between the cadaver and the surgeon doing the cutting. Flaubert, quoted in Crapanzano , p. It is important to understand that meanings and contexts including an individual's sense of identity , the basic building blocks of qualitative research, are not fixed, constant objects with immutable traits.
Rather, meanings and identities are fluid and changeable according to the situation and the persons involved. Gustave Flaubert precisely captures the sense of active personal meaning-making and remaking across time.
Cohler describes such meaning-making and remaking as the personal life history self, a self that interprets, experiences, and marshals meanings as a means to manage adversity. A classic illustration of the fluidity of meanings is the case presented by Evans-Pritchard who explains the difficulty he had determining the names of his informants at the start of his fieldwork in Africa.
He was repeatedly given entirely different names by the same people. In the kinship-based society, the name or identity one provides to another person depends on factors relative to each person's respective clan membership, age, and community.
Now known as the principle of segmentary opposition, the situated and contextual nature of identities was illustrated once the fieldworker discovered the informants were indexing their names to provide an identity at an equal level of social organization.
For example, to explain who we are when we travel outside the United States, we identify ourselves as Americans, not as someone from Oakdale Road. When we introduce ourselves to a new neighbor at a neighborhood block party, we identify ourselves by our apartment building or house on the block, not by reference to our identity as residents at the state or national level.
Themes and personal meanings are markers of processes not fixed structures. Life stories, whose narration is organized around a strongly held personal theme s as opposed to a chronology of events from birth to present day, have been linked with distress and clinical depression Luborsky b.
Williams suggests that the experience of being ill from a chronic medical disease arises when the disease disrupts the expected trajectory of one's biography. Some researchers argue that a break in the sense of continuity in personal meaning Becker , rather than any particular meaning theme , precedes illness and depression Atchley ; Antonovsky Another example of fluid meaning is ethnicity.
Ethnic identity is a set of meanings that can be fluid and vary according to the social situation, historical time period, and its personal salience over the lifetime Luborsky and Rubinstein , Ethnic identity serves as a source of fixed, basic family values during child socialization; more fluidly, as an ascribed family identity to redefine or even reject as part of psychological processes of individuation in early adulthood; sometimes a source of social stigma in communities or in times of war with foreign countries e.
From the qualitative perspective, there are a number of contrasts that emerge between sampling for meaning and more traditional, survey-style sampling, which has different goals. Those who are not familiar with the sampling-for-meaning approach often voice concerns over such aspects as size Lieberson , adequacy and, most tellingly, purpose of the sampling.
Why, for example, are sample sizes often relatively small? What is elicited and why? What is the relationship between meanings and other traditional categories of analyses, such as age, sex, class, social statuses, or particular diseases? What is perhaps the most important contrast between the sampling-for-meaning approach and more standard survey sampling is found in the model of the person that underlies elicitation strategies.
From this perspective, individuals are viewed as sets of fixed traits and not as carriers and makers of meaning. Sampling for meaning, in contrast, is based on four very distinct notions. The first is that responses have contexts and carry referential meaning. Thus questions about events, activities, or other categories of experience cannot be understood without some consideration of how these events implicate other similar or contrasting events in a person's life Scheer and Luborsky This is particularly important for older people.
Second, individuals often actively interpret experience. That is to say, many people—but not all—actively work to consider their experience, put it in context, and understand it. Experience is not a fixed response. Further, the concern with meanings or of remaking meaning can be more emergent during some life stages and events or attention to certain kinds of meanings than others. Examples of this include bereavement, retirement, ethnic identity, and personal life themes in later life.
Third, certain categories of data do not have a separable existence apart from their occurrences embodied within routines and habits of the day and the body. Consequently, qualitative research provides a context and facilitates a process of collaboration between researcher and informant. Fourth, interpretation, either as natural for the informant or facilitated in the research interview, is basically an action of interpretation of experience that makes reference to both sociocultural standards, be they general cultural standards or local community ones, as well as the ongoing template or matrix of individual experience.
Thus, for example, a person knows cultural ideals about a marriage, has some knowledge of other people's marriages, and has intimate knowledge of one's own.
In the process of interpretation, all these levels come into play. These issues occur over a variety of sampling frames and processing frameworks. There are three such sampling contexts. First, sampling for meaning occurs in relation to individuals as representatives of experiential types.
Here, the goal is the elucidation of particular types of meaning or experience personal, setting-based, sociocultural , through inquiry about, discussion of, and conversation concerning experiences and the interpretation of events and social occur-rences. The goal of sampling, in this case, is to produce collections of individuals from whom the nature of experience can be elicited through verbal descriptions and narrations.
Second, sampling for meaning can occur in the context of an individual in a defined social process. An example here could include understanding the entry of a person into a medical practice as a patient, for the treatment of a disorder.
Qualitatively, we might wish to follow this person as she moves through medical channels, following referrals, tests, and the like.
Even beginning this research at a single primary physician, or with a sample of individuals who have a certain disorder, the structure of passage through a processing system may vary widely and complexly. However, given a fixed point of entry a medical practice or a single disease , sampling for meaning is nested in ongoing social processes. Researchers wish to understand not only the patient's experience of this setting as she moves through it e.
Finally, researchers may wish to consider sampling for meaning in a fixed social setting. An example might be a nursing home unit, with a more or less fixed number of residents, some stability but some change, and regular staff of several types representing distinctive organizational strata and interests administration, medicine, nursing, social work, aides, volunteers, family, or environmental services. It is important to note that even though qualitative research focuses on the individual, subjectivity or individuality is not the only goal of study.
Qualitative research can focus on the macrolevel. One basic goal of qualitative research in aging is to describe the contents of people's experiences of life, health, and disability. It is true that much of the research to date treats the individual as the basic unit of analysis. Yet, the development of insights into the cultural construction of life experiences is an equal priority because cultural beliefs and values instill and shape powerful experiences, ideals, and motivations and shape how individuals make sense of and respond to events.
Studying how macrolevel cultural and community ideologies pattern the microlevel of individual life is part of a tradition stretching from Margaret Mead, Max Weber, Robert Merton, Talcott Parsons, to studies of physical and mental disabilities by Edgerton , Esteroff , and Murphy For example, Stouffer's pioneering of survey methods revealed that American soldiers in World War II responded to the shared adversity of combat differently according to personal expectations based on sociocultural value patterns and lived experiences.
These findings further illustrate Merton's theories of relative deprivation and reference groups, which point to the basis of individual well-being in basic processes of social comparison. The notion of stigma illustrates the micro- and the macrolevels of analyses.
For example, stigma theory's long reign in the social and political sciences and in clinical practice illustrates the micro- and macroqualitative perspectives. Stigma theory posits that individuals are socially marked or stigmatized by negative cultural evaluations because of visible differences or deformities, as defined by the community.
Patterns of avoidance and denial of the disabled mark the socially conditioned feelings of revulsion, fear, or contagion. Personal experiences of low self-esteem result when negative messages are internalized by, for example, persons with visible impairments, or the elderly in an ageist setting. Management of social stigma by individuals and family is as much a focus as is management of impairments.
Stigma is related significantly to compliance with prescribed adaptive devices Zola ; Luborsky a. A graphic case of this phenomenon are polio survivors who were homebound due to dependence on massive bedside artificial ventilators.
With the recent advent of portable ventilators, polio survivors gained the opportunity to become mobile and travel outside the home, but they did not adopt the new equipment, because the new independence was far outweighed by the public stigma they experienced Kaufert and Locker In addition, this type of sampling method does not require that a simple random sample is generated, since the only criteria is whether the participants agree to participate .
Convenience sampling is not often recommended for research due to the possibility of sampling error and lack of representation of population. But it can be handy depending on the situation. In some situations, convenience sampling is the only possible option. For example, a college student who is doing a term project and wants to know the average consumption of beer in that college town on Friday night will most probably call some of his friends and ask them how many cans of beer they drink, or go to a nearby party to do an easy survey.
There is always a trade-off between this method of quick sampling and accuracy. Collected samples may not represent the population of interest and therefore be a source of bias.
In the example above, if said college town has a small population and mostly consists of students, and that particular student chooses a graduation party for survey , then his sample has a fair chance to represent the population.
Larger sample size will reduce the chance of sampling error occurring. Another example would be a gaming company that wants to know how one of their games is doing in the market one day after its release. Its analyst may choose to create an online survey on Facebook to rate that game. The major challenge of this approach will be reaching to the people who play games.
As social media is a vast place, it's always difficult to collect samples from the population of interest. Most people may not be interested or take the survey seriously while completing it, which results in sampling error.
The survey may be improved greatly if the analyst posts it to fan pages dedicated to game-lovers. He may find a lot more people in that group who would be inclined to judge and rate the game critically. Convenience sampling can be used by almost anyone and has been around for generations. One of the reasons that it is most often used is due to the numerous advantages it provides.
This method is extremely speedy, easy, readily available, and cost effective, causing it to be an attractive option to most researchers. When time is of the essence, many researchers turn to convenience sampling for data collection , as they can swiftly gather data and begin their calculations. It is also useful when researchers need to conduct pilot data collection in order to gain a quick understanding of certain trends or to develop hypotheses for future research.
By rapidly gathering information, researchers and scientists can isolate growing trends, or extrapolate generalized information from local public opinion. For researchers who are not looking for an accurate sampling, they can simply collect their information and move on to other aspects of their study. This type of sampling can be done by simply creating a questionnaire and distributing it to their targeted group.
Through this method, researchers can easily finish collecting their data in a matter of hours, free from worrying about whether it is an accurate representation of the population. Since most convenience sampling is collected with the populations on hand, the data is readily available for the researcher to collect.
Convenience sampling (also known as availability sampling) is a specific type of non-probability sampling method that relies on data collection from population members who are conveniently available to participate in study. Facebook polls or questions can be mentioned as a popular example for convenience sampling.
This is the least desirable sampling method, and researchers should typically avoid using it. More rigorous alternatives include purposeful and other strategic sampling methods.
Qualitative research is designed to explore the human elements of a given topic, while specific qualitative methods examine how individuals see and experienc. 2 Ilker Etikan et al.: Comparison of Convenience Sampling and Purposive Sampling. include every subject because the population is almost finite. This is the rationale behind using sampling techniques like convenience sampling by most researchers .
Sampling is a very complex issue in qualitative research as there are many variations of qualitative sampling described in the literature and much confusion and overlapping of types of sampling, particularly in the case of purposeful. Convenience sampling is a method in which, for convenience sake, the study units that happen to be available at the time of data collection are selected in the sample. Many health facility or drug-outlet-based studies use convenience samples.