What are qualitative methods?
Analysis of social data are crucial for health programme managers to understand why and how people make choices and the way they behave. In contrast to quantitative methods that rely on counts of the occurrence of a phenomenon, qualitative methods look for patterns in opinion and on subjective explanations of human behaviour.
Qualitative approaches are vital to understand behaviours relating to health: preference, choice, constraints, motivation, limitations, decision making, utilization etc.; thus helping to develop solutions that are ground-up, community-based and context-specific. When triangulated with quantitative data, qualitative information provides the insights that begin to explain global health.
Why qualitative inquiry?
Public health practitioners need to study disease within its cultural and social setting and to understand patients as individuals whose health decisions are subject to multiple constraints. This bio-psycho-social model of disease raises new questions, for example: what factors influence a patient’s delay in seeking medical intervention or, will a patient follow the recommendations of a professional? Similarly, social factors such as training and motivation of staff influence the implementation of health policy. We cannot assume that an intervention that succeeds in one place will work the same way in another town, region or country. Qualitative methods, in combination with other types of inquiries, can help explain these issues.
Policymakers require information that tells them not only what has worked but, crucially, why. Realising, for example, that people and communities best understand their problems and can tailor solutions to their contexts, in 2013, the United Nations (UN) called for a global conversation bringing ‘the voices of the people to the table’.
In response to the West African Ebola outbreak, the World Health Organization, observed that: ‘Communities have been, and will continue to be, the most critical part of an effective response’. WHO partnered with anthropologists to understand and work with – rather than against – local cultures, beliefs and practices. For example, a year after the epidemic started, medical anthropologists found that technical safety guidelines for health workers, such as wearing gloves and masks, contradicted a cultural view of compassionate care, one in which professionals should prioritise the treatment of medical emergencies over all else. Their findings revealed the need for new guidelines that respect the context in which medical professionals routinely place urgent patient care ahead of the recommended protocols for safeguarding their health.
The qualitative approach aims to reduce the distance between the expert and respondent. The approach can be advantageous when trying to obtain information about marginalised and hard-to-reach groups who are distrustful of outsiders.
But despite its potential, the health community has not wholeheartedly embraced the qualitative approach, neglecting to draw upon the subjective experiences of patients and frontline colleagues. There is a preference for hard evidence – read numbers.
Functions of qualitative inquiry
Qualitative inquiry serves three distinct but overlapping functions:
Exploration
Exploration is an inductive approach to identify areas that need further investigation; Rather than looking for data to validate or disprove a given hypothesis, the researcher builds theories during data collection and analysis using an iterative approach which prioritises respondents’ views. Findings can provide exploratory evidence about issues which others may not have considered. With time, some exploratory findings inform routine data collection, for example it is now ubiquitous that researchers ask about socio-economic factors in health research.
Explanation
Qualitative data can explain phenomena already identified through quantitative or anecdotal evidence. This deductive approach can help policymakers understand how policies and interventions work in everyday contexts by answering questions such as: What was it about this intervention that worked? What were the weaknesses of the intervention? Were any other factors responsible for its success or failure?
Triangulation
Triangulation is a process of confirming results observed through one source by referring to results from another. While triangulation can add rigour to all investigations, the approach is particularly useful for mixed methods studies employing both qualitative and quantitative approaches.
Four classes of qualitative data
Facts
Facts are pieces of information that the researcher has verified to be objectively true, often through document review. Researchers can obtain facts from people in authority or from multiple respondents who all report the same information, but they must independently verify the information. Facts provide structure to the situation being studied and are the unchanging constraints into which subjective data fit.
Assertions
Assertions are opinions that could be facts if verified. It is important to understand that respondents treat assertions as fact, whether or not others can verify them. Assertions are the structural elements in their view of the situation and are unlikely to change. When it is not possible or worthwhile to verify an assertion, researchers should treat it as an opinion.
Opinions
Opinions are the most common type of qualitative data. They are the perspectives, views and beliefs of the respondents, usually heralded by terms such as ‘I think’, ‘I feel’ or ‘I believe’. Opinions are subjective without objective verification. They are phenomenological insights into how respondents experience the situation.
Narratives
Narratives are a mix of opinions, facts and assertions. They are the stories that people tell when asked to explain a situation or an action. Narratives don’t exist solely at an individual level but involve broader social narratives. Narratives are ethnographic; they do not just provide information on an individual respondent but also on the culture in which the respondent operates. Understanding how cultural narratives shape an individual’s opinions and actions makes it possible to identify implicit deep-rooted social issues.
Gathering data
Researchers’ theoretical approaches shape their choice of methods. Their choice of theory influences the topics that researchers investigate, the type of data they elicit and how they approach data collection and analysis.
Key theories underlying qualitative data collection
- Ethnography systematically studies people within a particular cultural setting. It involves long-term observation and discussion with research subjects, to capture cultural meanings from their perspectives.
- Phenomenology studies the structures of consciousness and how to construct meaning. It aims to capture how individuals experience and interpret social phenomena.
- Grounded theory is an inductive methodology for generating theory through a cycle of research, analysis and theory refinement. The researcher develops theories that directly explain the social phenomena they are studying.
- Action Research is a form of applied research which aims to find effective solutions for practical issues through a spiral of planning, action and learning. It involves close collaboration with research participants, with the researcher participating in, or facilitating, the process of change.
Common forms of data collection
- Document reviews: the investigator systematically studies relevant texts. Documents may include peer-reviewed articles from published journals or grey literature that describe programmes, initiatives and small studies, or people’s reactions to existing policies and programmes. Increasingly, analysts trawl news content and social media feeds to capture trends in prevailing opinion.
- Focus group discussions: a trained facilitator brings together a small group of respondents to discuss a topic of interest, and leads and documents the discussion. The facilitator chooses respondents purposively to embrace differing views so that the discussion captures the full range of experiences and explanations for the topic.
- In-depth interviews: the interviewer has a deep dive one-on-one conversation with a respondent. The interviewer selects respondents purposively to provide information on the subject of inquiry. Interviews may be structured, semi-structured or unstructured; the choice depends on the need to put the respondent at ease and on the purpose of the inquiry.
- Questionnaires: researchers quantify attitudes, values and opinions through questionnaires. They may use Likert Scales to find out how much respondents agree or disagree with a series of statements on a topic. The researcher assigns numeric values to respondents’ opinions, most commonly using a scale of options of agreement, frequency, value, relevance, importance, quality, or likelihood, and then analyses the values using statistical methods. As this method yields less in-depth information than other qualitative techniques, researchers can use it in conjunction with open-ended questions to allow respondents to explain their choices. Researchers may use other quantifiable techniques such as ranking or ordering opinions.
Other methods to collect information
- Asking a person to narrate their life history, for example about their reproductive journey, can reveal unexpected events and information on types of contraception they have used at different points in their lives and the circumstances which triggered reproductive decisions.
- Protocol analysis asks respondents to share their decision-making process with the researcher – it can be used to see what factors patients consider when choosing a doctor, or deciding to seek help for a medical issue.
- Closely related to protocol analysis is the méthode clinique (also called experimental phenomenology) which documents, through observation or questions, how respondents handle a situation of interest. For instance, a researcher may sit with diabetic patients while they organise their weekly medicines to gain insight into how they approach illness and treatment.
- During participant observation the researcher completely immerses in a culture and sometimes spends several years in the field. Seminal medical anthropology studies such as these have yielded valuable insights for global health.
Participatory Rural Appraisal (PRA) captures information for public health with considerable accuracy. Researchers work with relevant community actors, for example, women, frontline workers and local self-government officials. They combine several techniques to construct a map of the community that portrays the environmental factors contributing to the health outcome. Techniques include diagramming using venn diagrams, taking transect walks, and drawing flow diagrams and daily routine charts; interviewing individuals and conducting focus group discussions; guiding preference ranking using matrix-ranking, proportion piling and wealth ranking; and mapping and modelling for social factors, resources, physical and hazard maps.
Sampling
Sampling and case selection work differently for the qualitative than for the quantitative approach, with researchers taking effort to identify persons most relevant to the issue they are studying. Exploratory studies capture a wide range of views so that they can construct a theory. Explanatory studies aim to understand why people think a particular way, rather than to capture the most prevalent views. Qualitative researchers tend to use small samples of respondents selected through:
- Purposive sampling, that is deliberately selecting respondents who have characteristics of interest to the survey;
- Convenience sampling, that is choosing respondents who are easily accessible; or
- Snowball sampling, that is asking respondents to recommend other respondents.
Researchers stop collecting data when they are no longer capturing new types of views, that is when the data have reached saturation. It is seldom necessary to obtain a large sample.
Analysing qualitative data
Qualitative researchers use several methods to analyse data, but most commonly they use:
- Coding to categorise, sort and organise their data. Researchers read the data, often in the form of transcripts and field notes, and assign codes to segments of text according to themes of interest. They iteratively refine the codes and explore how they relate and link to each other. Software such as ATLAS.ti and NVivo can streamline the coding process, allowing for the sorting of codes and quotes as well as the coding of data in audio and video formats.
- Recursive abstraction in which the researcher writes sequential summaries of the situation. Recursive abstraction, while time-consuming, creates a brief and clear summary.
There are ways to analyse qualitative data, quantitatively. Likert scales, ranking and ordering of responses yield numeric data that permit limited quantitative analysis. For large datasets, researchers rely on specialized computer methods. For example, they use:
- Content analysis to sort through text, categorise themes and count the number of times keywords appear.
- Social network analysis identifies and maps networks between social structures, for example providing visuals of social media groupings and connections.
- Factor analysis is useful to identify broad factors or components of interest and understand the frequency with which they occur.
How to judge the findings
While quantitative research has established methods to verify its findings, qualitative research has no standard quality measures. There are three camps: 1) those who think that quantitative measures for assessing rigour should apply to qualitative research; 2) those who think measures more suited to qualitative research are needed; and 3) those who think that any attempt to establish standardised quality measures for qualitative research will fail. The latter argues that it is not possible to judge all qualitative research by one set of criteria because there is no unified qualitative research paradigm and its nature, methods and outputs are diverse. They conclude that the user should assess the credibility and usefulness of each study on its merits.
We take a pragmatic approach and suggest that users ask the following questions about qualitative research findings:
- Who are the researchers, what are their interests in the study and how might these have influenced the results?
- Have the researchers provided reasonable justification for the sample they have selected?
- What were the data collection and analytical methods? Are these appropriate for the research questions and context?
- Is it possible to trace how the findings relate to the original data? Are there any cases that deviated from the others? If so, does this undermine the conclusions drawn from the data or is there a reasonable explanation for any outliers?
- Have the researchers double checked their analyses, and triangulated their findings; are the findings backed up by information from other sources and existing theories?
Users of qualitative research can make informed judgments about the credibility of the work and its relevance to their situations by carefully reading the qualitative text.
Rigorous qualitative researchers lay out their biases for the user’s scrutiny – a process known as reflexivity explaining how these biases may have influenced their interpretation.
Researchers document their respondents’ views and highlight when their interpretations of the data go against a respondent’s stated position, explaining why they do not accept the respondent’s statement at face-value. When researchers share their findings with their respondents, it is important to find out if the respondents agree with the researcher’s interpretations and if they think the findings give voice to their views. Sharing findings not only disseminates the research back to the community but adds credibility and ethical value to the study.
Challenges and opportunities
The qualitative approach operates in a different epistemological framework to that found in the hard sciences. The skills required to launch a qualitative inquiry are technical and take time and effort to learn – anthropologists, for instance, can spend a decade or more in post-graduate studies conducting ethnographies. Those trained in other knowledge systems may feel that the qualitative approach flouts their discipline’s key assumptions about objectivity, methodology and validity, and find the qualitative approach challenging. Because qualitative inquiry requires reflexivity, some health professionals, deeply embedded in their field, may find it difficult to engage with their own biases.
These challenges are not insurmountable; the divide between qualitative and quantitative methods is often artificial. Policymakers and medical professionals unwittingly deploy qualitative techniques all the time, such as when a policymaker has an unstructured discussion with potential beneficiaries of a public health scheme, a hospital administrator conducts rounds of patients’ rooms or a doctor has an in-depth conversation with a patient’s family. The challenge is to recognise that these interactions produce data which are as legitimate as hard numbers. Greater collaboration between qualitative and quantitative researchers can expand the range of methods used to answer policy questions. Such partnerships are critical to improving training to promote better use of qualitative and mixed methods.
New technologies provide innovative ways to collect and synthesize qualitative data. Open Space Technology workshops, for example, allow any number of stakeholders to hold discussions on a topic. Participants set their agenda by identifying priority sub-topics of interest, and split organically into small groups to consider the issues. The format results in multi-level prioritising of sub-topics, issues within it and the emergence of the most important findings.
Big data provide more information than ever before and open possibilities to understand trends in individuals’ health-seeking behaviour. They give much-needed context to variables, relationships and patterns. Big data offer new opportunities for qualitative research – traditionally rich in depth but, until recently, limited in breadth.
Contents
Source chapter
The complete chapter on which we based this page:
Singh S., Krishan A., Telford M. (2019) Seeking Insight: Using Qualitative Data for Policymaking. In: Macfarlane S., AbouZahr C. (eds) The Palgrave Handbook of Global Health Data Methods for Policy and Practice. Palgrave Macmillan, London.
Additional resources
Berg BL, Lune H. Qualitative research methods for the social sciences.
Emerson RM, Fretz RI, Shaw LL. Writing ethnographic fieldnotes.
Weiss RS. Learning from strangers: The art and method of qualitative interview studies.
Seale C. Introducing qualitative methods: The quality of qualitative research
Hannes K. Critical appraisal of qualitative research.