Abstract:

With the methodological changes in psychology, more and more researchers tend to accept qualitative research as an effective way to solve psychological problems and serve the public. In a qualitative study, sufficient sample is the guarantee of research validity, and saturation is an indicator used to assess the adequacy of research data. Saturation means that on the basis of the currently collected and analyzed data, further data collection will not help researchers develop a deeper understanding of the story or theory, so there is no need to continue to collect data. The concept of theoretical saturation was first proposed in grounded theory. Then with the development of qualitative research methods, researchers have further created more saturation concepts, including data saturation, code or thematic saturation, meaning saturation, etc. Due to the diversity of saturation and its judgment standards, the relationship between different kinds of saturation are complicated and ambiguous. In addition, previous studies lack operational description and practical guidance for the evaluation of saturation, which leads to the vagueness of the concept of saturation and many difficulties in evaluation. In order to solve these problems, this study clarified the concepts and evaluation methods of four levels of saturation, and provided suggestions for researchers' operations based on comparison and analysis. The four types of saturation occur at different stages of the research process, and each has its own specific connotations. Data saturation, code or thematic saturation focuses on the breadth of collected data, while meaning saturation and theoretical saturation focus on the depth of research data. In terms of evaluation methods and criteria, researchers usually judge data saturation based on the repeatability of initial data; code or thematic saturation is determined based on empirical research results, the emergence of new codes or themes, or saturation coefficients; the results of retrospective empirical analysis or tables of meaning unit are normally used to evaluate meaning saturation; while the assessment of theoretical saturation relies on a process called "continuous comparison" in grounded theory, which focuses on the continuous improvement of the theory. Some problems are discussed in this study. 1) The sample size standard for reaching saturation should be embedded in the specific research process instead of being uniformly set in advance. Because each study has its own uniqueness in terms of questions, purposes, methods, etc., which saturation is extremely sensitive to, the evaluation of saturation should be based on the characteristics of the current research to select an appropriate level of saturation. 2) Due to the logical uncertainty of saturation, a little oversampling would be helpful. The logical uncertainty here means that researchers can only predict the necessity of continuing data collection based on the information that has been collected, which relies on the subjective judgment of researchers, and its accuracy can never be further proved. Oversampling may be an effective way to solve this problem, which means that even if saturation has been achieved, the researcher is recommended to add 2 to 3 personal interviews or 1 to 2 focus group interviews to further confirm. 3) As an important index to evaluate the quality of qualitative research, saturation is not suitable for all qualitative research, such as psychobiography, narrative analysis, etc., which focus on single or a few cases and pay more attention to the integrity of individual stories. In the future, researchers should further focus on the evaluation and testing of saturation in different kinds of qualitative research.

Key words: qualitative research, saturation, evaluation methods, judgment criteria