Credibility refers to whether you can believe your results. When you evaluate credibility, you are checking whether your research tests what it means to test, and asking if there are any exterior factors that would cause your data to be inaccurate. For instance, prolonged engagement is a technique for confirming credibility when researching an unfamiliar culture or group. Researchers spend enough time with members of the group so that they can be sure that they are not simply getting a rote answer, or can account for cultural differences in the responses they receive. For a study to be credible, it must also provide a detailed description of the environment under which it was conducted so that the reader understands the conditions under which data was obtained.
Qualitative research, unlike quantitative research, is usually based on smaller sample sizes and uses data gathered from interviews and personal responses. Qualitative data is analyzed non-statistically, so you cannot depend on statistical analysis to determine if your research produces trustworthy results. Based on the method devised by education professors Yvonna Lincoln and Egon Guba, you can evaluate a study's trustworthiness based on four qualities -- credibility, transferability, dependability and confirmability.
A study's transferability is the degree to which its data can be applied to contexts other than the context under which the research is performed. A in-depth, or thick description of the research's setting can help a reader determine how much of the research's results relate directly to the methods used by the researchers. For instance, a study of construction workers' reactions to stress might indicate that this stress is compounded by the danger involved in working with heavy machinery. This means the results might be transferable to other jobs that use heavy machinery, but not to construction workers who do not use heavy machinery. A research study does not have to be fully transferable to be trustworthy -- however, it should include the information needed for a reader to determine what portion of the research is transferable.
Even as researchers collect qualitative data, the environmental circumstances surrounding their research are constantly changing. For instance, a community's trust in town government may change dramatically if a new mayor is elected. A study's dependability measures whether the same results would be obtained if the study were repeated. Part of evaluating a study's dependability is, again, contextual detail -- researchers should note any changes that occur in the population or society while they are collecting data. In addition, external audits may be used to establish dependability. In an external audit, a researcher not involved in the project reviews the data to see if he comes to the same conclusions, or can challenge any of the assumptions, made by the primary researchers.
An audit may also be used to evaluate a study for confirmability, or objectivity -- the degree to which the study's results are independent of the researchers' biases. An audit allows an external expert to review the data and determine if it supports the researchers' findings. To aid in this process, researchers should create an audit trail, a record of the steps used in their research for the external auditor to review. Contextual description can also be used to boost confirmability by providing reflexive information. Reflexive information identifies the background and education of the researchers performing the survey, which will provide the reader with a greater understanding of the biases and assumptions that might shape the study's conclusions.
- Snap Surveys: Qualitative vs. Quantitative Research
- Northern Arizona University: Qualitative Research
- Research Methods: Qualitative Validity
- Robert Wood Johnson Foundation: Lincoln and Guba's Evaluative Criteria
- Andrew K. Shenton: Strategies for ensuring trustworthiness in qualitative research projects
- Photo Credit JackF/iStock/Getty Images