A consistent research process supports dependability
There are several ways through which a researcher can strengthen the dependability of their data analysis. These include independent coding-recoding, peer examination, dialogue among co-researchers, panel discussion, and face validity. A researcher can re-analyse their data (either the entire data set or a smaller part of it) to check the consistency of the data analysis technique. For example, a researcher can analyse their data twice and assess how the results answer the research question. However, this approach may not be effective as it is highly plausible that the researcher will remember how they conducted the analysis the first time.
Another alternative is peer examination, in which another researcher analyses the data and assesses how their results compare to the original findings. This includes a certain level of risk because both researchers will analyse the data from their own perspectives. Hence, the peer reviewer needs a detailed introduction that will cover the motivation for data analysis, along with the approaches that were used. If more than one person analyses the data, it may be beneficial to calculate the data agreement coefficient (ICR). A value >80% reflects a valid assessment by both researchers . However, an ICR assessment cannot always be performed. Furthermore, an inductive content analysis is usually only performed by one researcher because it is time-consuming and tedious. A peer examination may not be relevant for inductive content analyses because this technique is used to identify concepts based on subjective interpretation of the data.
Dialogue among colleagues is also relevant to credibility, and researchers should ask colleagues who are familiar with the research subject to read through the findings and share their candid opinions about study credibility. In these situations, tables or pictures that depict the development process of each main category are useful. The results section should start with examples of identified open codes, for example, quotations from the collected data, and end with the main categories. Having another researcher read through the research report can be useful because another set of eyes may notice overlap between the identified categories that the primary researcher missed. When the steps of the data analysis are presented clearly, another researcher can notice flaws in the research, for example, incomplete data abstraction or the grouping of too many items under one category. Furthermore, research that presents a large array of main concepts may indicate that the researcher was not able to group the data under the correct categories. For this reason, the researcher should always specify the number of identified categories and/or concepts—preferably through clear tables or figures—when describing the analysis process.
In essence, the issue underlying the choice to test face validity or give the research to a peer for evaluation is a lack of confidence, and certain scholars argue that a researcher should not need someone else to analyse their data. A detailed description of the analytical process is a good starting point from which a researcher can build confidence about the trustworthiness of their research.
5.4 Trustworthiness: Confirmability
Confirmability is a measure of how well the study findings are supported by the collected data . This aspect of trustworthiness is concerned with the connection between the data and the results. Hence, when considering confirmability, a researcher should evaluate whether their findings are solely shaped by the data collected from respondents, or do the results reflect some of the researcher’s bias, motivation, or other interests . The reader should be able to examine the data to confirm that the results or author interpretations reflect the data. A researcher can enhance confirmability by using ‘audit trails’, which means that the researcher will include written field notes, memos, or excerpts from a field diary to support the connection between the data and findings. However, this practice includes the same problems that were described earlier, i.e., written notes and diary entries are intended for the researcher rather than for outsiders. As such, researchers should understand that including ‘audit trails’ can also potentially harm the trustworthiness of their research. This criterion is closely related to the concept of authenticity, which is described in the next section and can also be used to gauge the connection between the data and results.
5.5 Trustworthiness: Authenticity
Authenticity describes the extent to which researchers fairly and faithfully show a range of realities . Research that has sufficient authenticity will include various citations that clearly demonstrate the connection between the results and data. These citations should be used systematically throughout the text, for example, each identified category should include at least one relevant citation. Furthermore, it is important to include citations from different participants, as several previous studies have presented citations that reflect only one participant. In this situation, the reader may wonder whether this was the only participant who expressed something that was relevant to the research question. The researcher should also be able to demonstrate that the citation originates from the original data, for example, by using an ‘identification’ code. For example, the code ‘BC35’ could demonstrate that the participant is a woman (B), a teacher (C), and 35-years old. However, the researcher must ensure that the identification codes are in line with current data protection guidelines and cannot be used to identify the participant. There is also a risk of including too many authentic citations. To avoid this, the researcher should ensure that there are not more citations than text in the results section, as this may cause readers to question the researcher’s ability to interpret the collected data. A researcher should always consider the value of including a certain citation. If the citation simply repeats what has been mentioned earlier, it might be boring for the reader and does not add any value.
5.6 Trustworthiness: Transferability
Transferability describes the degree to which research findings will be applicable to other fields and contexts . Researchers who are concerned about transferability should question whether their results will hold in another setting or group of participants. It is important to note that transferability is not the same as generalisation in quantitative research. It is important to note that transferability is not the same as generalisation in quantitative research because transferability is also concerned with how readers will extend the results to their own situations, whereas generalisation covers the extension of results from a sample to a broader population. Transferability is affected by every stage of research, including the choice of research context and topic. For example, the results from a study that focuses on the interactions between nurses and patients in an orthopaedic ward may not be transferable to the medical ward setting. This is because care and treatment in orthopaedic wards differs from that in internal medicine wards, so it can be assumed that the interactions between nurses and patients in these two settings focus on different issues. However, the results from the orthopaedic ward study may be transferable to another surgical ward because these wards have some similar elements, for example, patients are waiting for operations, which means that they may have some fears about their situation and/or they need assistance in basic daily activities. During the research planning phase, a researcher should consider transferability by clearly describing the sampling techniques, potential inclusion criteria, and participants’ main characteristic so that other researcher can assess whether the results drawn from this sample are applicable to other contexts. Transparent reporting of the research process and results is critical to achieving sufficient transferability. Every researcher is responsible for providing enough information about their study so that the audience can evaluate whether the findings are applicable to other contexts. Hence, researchers who want to present transferable knowledge should consider the following question while writing their results and discussion: How, and to what extent, are these findings transferable to other settings?
A key element of trustworthiness is the sample. It must be appropriate and comprise participants and/or documents that are relevant to the research topic. Purposive sampling may be useful for building an appropriate sample, but data saturation is the most important measure of sampling adequacy because it provides the optimal sample size. Data saturation ensures that the gathered data can be organised into categories, concepts, and themes, which, in turn, verifies that the analysis is complete. Researchers who want to provide trustworthy analyses should consider performing a preliminary analysis after a few interviews or once they have collected some data from the study documents. Researchers should also keep in mind that the chosen unit of analysis will influence trustworthiness. A broad unit of analysis may be difficult to manage and can have various meanings, while a narrow unit of analysis may result in fragmentation. Both of these situations will negatively affect trustworthiness. Trustworthy research must be systematically reported and include clear indications of the connections between the data and results. The content and structure of concepts or narrative results should be clearly presented, and a researcher can provide figures to help the reader better understand the significance of the results. Failure to report the results in an appropriate way will threaten the trustworthiness of the study.
Elo et al.  published a checklist that researchers can use to improve the trustworthiness of studies that apply content analysis. This checklist is especially beneficial during the planning of a qualitative study, as it will ensure that the researcher pays attention to every issue that can affect trustworthiness. The checklist also provides valuable tips for the reporting of results, for example, researchers can use this guide to critically evaluate their research in terms of strengths and weaknesses to trustworthiness. Following the discussion of trustworthiness in this chapter, the next chapter will present ethical issues in the context of qualitative research and content analysis.