The validity and reliability of qualitative research represent the key aspects of the quality of research. When handled meticulously, the reliability and validity parameters help differentiate between good and bad research. They also assure readers that the findings of the study are credible and trustworthy. This aspect becomes particularly vital in case studies involving primary data analysis. Here the researcher’s subjectivity can highly influence the interpretation of the data. This article clarifies how to establish the validity and reliability of qualitative research using various tools and techniques.
Establishing the validity of qualitative data
Qualitative data is as important as quantitative data, as it also helps in establishing key research points. However, since it cannot be quantified, the question of its correctness is critical. Validity relates to the appropriateness of any research value, tools and techniques, and processes, including data collection and validation (Mohamad et al., 2015). Validity also establishes the soundness of the methodology, sampling process, data analysis process, and conclusion of the study (Golafshani, 2003).
The main aspect that needs to be ensured is that the research philosophies fall in line with the research. To maintain the validity of the research, there is a need to understand the underlying needs of the research, the overarching process guidelines and the societal rules of ethical research. While establishing validity, there needs to be a consensus among the individual and the society on how to establish the correctness and accuracy of the research. If this aspect is kept in mind, the tools and techniques used are bound to be accepted by wider audiences. The main point to remember, thus, is to choose the tools wisely, which will establish how to correct the data of the research.
One of the major techniques that can be used for establishing the validity of qualitative data includes choosing a skilled moderator. Employing a moderator will help overcome personal bias. Thus, the researcher or the organisation can employ moderators to ensure that the data is genuine and is not influenced by “what the researcher wants to see or hear”.
Another way to promote the validity of research is by employing the strategy of triangulation. This basically involves that the research will be conducted from different or multiple perspectives. For example, this can take the form of using several moderators, in different locations or it could be multiple individuals who are analysing the same data. Basically, any technique through which the researcher can analyse the data from different angles.
Furthermore, the validity of qualitative research can also be established using a technique known as respondent validation. This basically involves testing the initial results with the participants in order to see if the results still ring true.
|To overcome personal bias.
|Analysing the data collected.
|To conduct research from multiple perspectives.
|Multiple individuals analyse the same data.
|To test that initial results still ring true for the participants.
|Testing the initial results.
Establishing the reliability of qualitative data
Quantitative research includes reliability measures where the researcher must prove that the process and the results have replicable outcomes. On the other hand, reliability in qualitative research includes very diverse paradigms, where the aspect itself is epistemologically counter-intuitive along with having a very difficult definition (Russell, 2014). Thus, what needs to be done to maintain and establish reliability in qualitative research is to be consistent.
Reliability tests for qualitative research can be established by techniques like:
- refutational analysis,
- use of comprehensive data,
- constant testing and comparison of data,
- use of tables to record data,
- as well as the use of inclusive deviant cases.
These techniques can help support the data sourcing, data validation and data presentation process of the research, as well as support the claim of reliability in terms of form and context.
Triangulation is another aspect which becomes very important in establishing reliability in the research. Also, as an additional note, it is very important for qualitative research to include a reference to a quantitative aspect. The use of a simple quantitative aspect in otherwise completely qualitative research creates a very positive attitude towards the overall concept of the research and helps to establish reliability in a much easier form. Also, the inclusion of at least two reliability tests, as per the type of research outcomes of a research, is a dependable way of establishing that the research process and results are reliable.
|To support the claim of reliability
|Involves exploring and explaining contradictions between individual studies
|Use of comprehensive data, constant testing and comparison of data, use of tables to record data
|To organise the data formally and establish its authenticity
|Use of tools such as MS Excel and NVIVO for analysis
|To establish reliability
|This is done by ensuring credibility, transferability, dependability and confirmability
Dos of validity and reliability of qualitative data
- While establishing validity and reliability, it is very important to decide the tools and techniques to be used in the research, before conducting the actual research. This helps in establishing the parameters for obtaining reliable and valid results from the beginning and does not impair the results of the research at the end of the process.
- To conduct efficient reliability and validity measures, an effective assessment of the literature must be done to understand which processes will work. Irrelevant approaches that compromise the reliability and validity of research should not be pursued.
- Reliability and validity processes should be conducted in tandem with the research processes being carried out to confirm the research objective, which provides another additional layer of authenticity to the research work.
Don’ts’ of validity and reliability of qualitative data
- Reliability and validity should not be taken as an extra element of the research. If they do not add value to research, they result in insecurity regarding the accuracy of the results.
- To be able to establish reliability and validity, researchers should not include excessive measures to support their research claims. This may add to unauthenticated results, as adding too many measures will overcomplicate the research.
An example statement of validity
The validity of this research was established using two measures, the data blinding and the inclusion of different sampling groups in the plan. The research included an assessment of the knowledge of traditional cuisine among the present population of a city. Where the sample was divided into two groups- to reduce biases. The sample included young adults, who have been mostly raised in an urban environment, along with middle-aged and elderly population who have had a partial upbringing in the rural area of India. The inclusion of greater diversity and a large number of sample respondents led the research to reduce its biasness towards only one type of outcome, creating a base for valid results. The other technique used was to restrict the amount of information shared with the respondents to make sure that the research was not biased with preconceived notions of the respondents. These steps helped to establish the validity of the results gained, proving the accurateness of the qualitative research. Further, the validity of the questionnaire was established using a panel of experts that reviewed the questionnaire. And hence the statements that did not go well with the subject of the study were removed.
An example statement of reliability
In terms of establishing reliability, the researcher conducted two processes. The first included recording the data in a table to provide an overall assessment of the data collection process and the updated assessment of the results, as they come. The use of the table for recording data provided the researcher with a chance to quickly interpret the results as per the record of every individual respondent and realize the progress of the research.
The table also helped in the concise construction of the conclusion of the research parameters. The reliability was also assessed through data triangulation. Among the various model of data triangulation such as methodological triangulation, data triangulation, investigator triangulation and theoretical triangulation. The study adopted theoretical triangulation, wherein other research works in the same arena were analysed and presented as a literature review to support the results claims of the data collection and analysis process. Moreover, the reliability measures relating to the triangulation of data provided an extensive understanding of the research objectives, which provided an additional layer of reliable stamping to the research.