Grounded theory là gì?

Có một từ mà bất kỳ ai đã từng trải qua một chương trình đào tạo SĐH trong các ngành KHXH-NV ở các nước nói tiếng Anh hẳn đều đã nghe qua, đó là từ “grounded theory”. Một từ rất hay gặp trong các nghiên cứu định tính. Vậy nó là cái gì thế, và dịch sang tiếng Việt như thế nào nhỉ?

Trước hết, cần giải thích grounded theory có nghĩa là gì. Xin đọc định nghĩa dưới đây, trích trong bài viết của The Yeshiva University Fatherhood Project (2001):

The grounded theory method allows the researcher to begin a research study without having to test a hypothesis. Instead, it allows her to develop hypotheses by listening to what the research participants say. Because the method involves developing hypotheses after the data are collected, it is called hypothesis-generating research rather than hypothesis-testing research. The grounded theory method uses two basic principles: (1) questioning rather than measuring, and (2) generating hypotheses using theoretical coding.

Tóm tắt:
Đây là một phương pháp cho phép nhà nghiên cứu tiến hành nghiên cứu mà không cần có trước ít nhất là một giả thuyết nghiên cứu. Những giả thuyết này sẽ được hình thành dần trong quá trình tiếp cận vấn đề, và cách nghiên cứu này được gọi là phương pháp “tạo sinh giả thuyết” (hypothesis generating) thay vì “kiểm nghiệm giả thuyết” (hypothesis testing) như trong cách tiếp cận định lượng. Grounded theory sử dụng hai nguyên tắc chính: (1) phát vấn (questioning) thay vì đo lường (measuring), và (2) tạo sinh giả thuyết bằng cách mã hóa theo lý luận (theoretical coding).

Hừm, nghe lùng bùng lỗ tai, phải không các bạn? Thôi thì cứ đọc đi, rồi sẽ quen. Nói nôm na bằng lời của tôi: các bạn cứ việc bắt đầu một cách lơ mơ, không lý luận và không giả thuyết, nhưng phỏng vấn, quan sát đến đâu thì hệ thống hóa đến đó, và suy luận để đưa ra những câu trả lời tạm thời. Sau đó lại tiếp tục quan sát, mã hóa những gì quan sát được, và tiếp tục chỉnh sửa những giả thuyết đã hình thành trước đó để xây dựng giả thuyết mới, sao cho cuối cùng cảm thấy đã đủ thông tin để kết luận điều mình muốn kết luận.

Vậy GT dịch sang tiếng Việt ra sao? Tra trên google, tôi thấy ít nhất là có ba cách dịch dưới đây:
i. Lý thuyết nền (hoặc nền tảng)
ii. Lý thuyết cơ sở
iii. Lý thuyết cơ bản

Tùy các bạn chọn nhé; tôi thì thấy cách dịch đầu tiên là tốt nhất.
-------------
Cập nhật ngày 27/10/2012

Hôm nay mở link dẫn đến bài đọc mà tôi đã dẫn ở trên thì không được, link hỏng, phải đi tìm và sửa lại. E rằng sau này link lại thay đổi và lại mất dấu vết nên tôi chép lại toàn bài viết mà tôi đã trích dẫn ở trên để lưu. Dưới đây:

Bronx, NY, Nov 1, 2003 — 2003 marks the ninth year of the Yeshiva Fatherhood Project, a qualitative study of the subject experiences of fathers from several subcultures in the U. S. To date, we have interviewed more than 400 fathers in focus groups. We have also developed a standardized procedure for designing a qualitative research study, coding and analyzing the data. This summer, we published a book outlining our method (Auerbach & Silverstein, 2003). Because qualitative research lends itself particularly well to clinical and applied research, we thought our alumni might be interested in learning more about it.

Qualitative research is hypothesis-generating rather than hypothesis-testing research. As a result, qualitative studies are conducted and evaluated differently than quantitative ones. In hypothesis-testing research, the hypotheses are stated in terms of the relations between independent and dependent variables. The research goal is to operationalize the variables in order to test whether the hypothesized relations exist. In hypothesis-generating research, the goal is to discover the appropriate variables (called theoretical constructs), and the abstract pattern describing the relations between these constructs. Thus the goal of qualitative research is to formulate theory.

Qualitative research is the method of choice when investigating an area about which not enough is known to formulate meaningful hypotheses. When we began the Yeshiva Fatherhood Project in 1994, there was very little empirical research on the subjective experiences of fathers. Much of the research that did exist was matricentric, comparing fathering to maternal behaviors. Within this context, we did not know enough about fathering to formulate meaningful hypotheses. This was particularly true about African American fathers, Latino fathers, and gay fathers.

Because one does not know enough to formulate meaningful hypotheses, the participants are used as a source of knowledge. The researcher asks generalized questions, and develops hypotheses based on the answers to their questions. Thus the hypotheses are developed after the data is collected.

In hypothesis-testing research, the goal is to use random sampling from as large and as diverse a population as possible, in order to assure that the research sample is representative and therefore that the findings will be generalizable. From this perspective, a small, self-selected sample, such as many of the groups of fathers we interviewed, would not be representative, and therefore no useful conclusions can be drawn from it.

However, representativeness is not the appropriate criterion for evaluating a qualitative study. The aim of qualitative research is to develop transferable theories, rather thangeneralizable hypotheses. This means that the abstract patterns described by a study’s theoretical constructs will be applicable to other research samples, even though the specific content of the pattern may not be.

Moreover, random sampling is theoretically impossible in studies of culturally diverse and/or marginalized groups. Members of these subcultures are not as likely to be selected by a random sampling process as are members of majority or mainstream groups. Rather, a subculture reflects an elaborate social network that researchers cannot enter at random. They must begin by contacting respected members of the culture, gain their trust, and build contacts from there. This is particularly true of groups like the gay fathers or divorced fathers, many of whom prefer to remain invisible. The reality of having to deal with the complications of outsider/insider status effectively rules out random sampling.

There are many qualitative alternatives to random sampling. The one that we used was a combination of convenience and snowball sampling. Convenience sampling means recruiting individuals that are accessible, for example members of parenting support groups. Snowball sampling means starting with a convenience sample of a few research participants and asking them to suggest others. These, in turn, are asked to suggest more participants. In this manner, the research sample grows from the first few researchparticipants, the way a snowball enlarges while rolling down a hill.

In hypothesis-testing research, the size of your sample is determined in advance by using statistical power analysis. In hypothesis-generating research, in contrast, the size of your sample cannot be determined in advance. It is determined instead by a procedure calledtheoretical saturation. After the first study, each new research sample is used to refine your theoretical constructs, provide new information and new insights into your theory. After several studies with new samples, additional research samples no longer add to your understanding, but simply confirm what you already know. This point, when additional samples are not providing any new information, is defined as theoretical saturation.

Thus, both the quantitative and qualitative paradigms acknowledge that more research is needed before the results of any single study can be considered accurate. However, a quantitative approach would conclude that, if the sample is small and self-selected, the participants are not a representative sample of fathers, and therefore, no useful knowledge can be drawn from this sample. The qualitative paradigm, in contrast, values the knowledge that emerges from even very small, self-selected samples because hypotheses can be generated from these studies.

Our qualitative approach has resulted in many journal articles, several book chapters, numerous presentations, including three Continuing Education Workshops at American Psychological Conventions, and most recently our “How-to” book. We have more articles forthcoming, and plan to do another workshop at APA in Hawaii this summer. If you would like to learn more about qualitative research contact us at LBSilverst@aol.ocm or Carlauer@aol.com.

©An introduction to coding and analyzing data in qualitative research.

----

Nguồn: http://blogs.yu.edu/news/2003/11/01/2850/ (truy cập 27/10/2012)