Abstract
The use of documents as a source of extant data is relatively common in grounded theory (GT) research. While GT promotes the dictum “all is data,” finding consistent commentary on how to use documents as data is difficult, especially among seminal works. The need to be aware of the context of extant data is a vital step prior to commencing analysis, especially in view of the lack of physical interaction between the researcher and that data. Contextual positioning is proposed as a tool that can be used to prepare extant data for analysis. Contextual positioning enhances the interactivity of the data collection process and positions the researcher before the document in a more reflexive manner. A model of contextual positioning is presented in this article to assist researchers in positioning extant data (such as documents) more reflexively. A concrete example of the use of this method is outlined to promote understanding of the value of this process.
Introduction
The sheer number and variety of documents available offers the grounded theorist an abundance of data that can aid in building a grounded theory (GT). Despite the wealth of potential data sources to be found in documents, the positioning of such data in GT research has not been explicated in great detail. In this article, we explore the position of documents in a GT study and propose the process of contextual positioning. Contextual positioning enables the researcher to position extant data in their study with greater reflexivity through an enhanced awareness of the context from which the source of data is sourced and the one in which it exists.
Documents as Data in GT Research
Defining Data Sources in GT Research
Grounded theory methodology (GTM) is characterized by the systematic application of essential methods that guide the researcher through processes of theory building in the context of their adopted philosophical viewpoint (Birks & Mills, 2011). This methodology can use both quantitative and qualitative data to find out what is really going on (Glaser & Strauss, 1967) in respect of the studied phenomenon. Historically, however, qualitative studies are more prevalent than quantitative studies in GT research. Whether the grounded theorist subscribes to traditional, evolved, constructivist, or other schools of thought (Mills & Birks, 2014), the dictum “all is data”—first mentioned in The Discovery of Grounded Theory (Glaser & Strauss, 1967)—acts as a guiding principle for those who use GTM. Under this banner, the grounded theorist has a universe of potential data sources to use in the development of a GT. In addition to the common data sources of interviews, focus groups and field observations, a number of other potential sources of data are available. Table 1 provides examples of such sources.
Possible Sources of Data for the Grounded Theorist.
Although many forms of data are available to the grounded theorist, researchers positioned in the qualitative paradigm—inclusive of many grounded theorists—have shown a preference to utilizing elicited data such as interviews and focus groups (Silverman, 1998). Restricting the scope of research data is problematic as it can deemphasize the value of other sources of information. Silverman’s (1998) survey of qualitative research articles published in Sociology and Qualitative Health Research demonstrates that interviews dominate as the single most preferred method in qualitative research.
Silverman (1998) expressed concern over the methodological impact of such trends and queried whether a preference for interviews was associated with the use of interviewing as a nursing tool, given the high proportion of nursing research published in these journals. To ascertain current trends, we replicated Silverman’s survey of the same journals. The results of both Silverman’s original study of publications from 1991 to 1996 and our subsequent survey of articles from 2010 to 2012 are presented in Table 2.
Qualitative Research Articles Published in Sociology and QHR.
Note. QHR = Qualitative Health Research.
The evidence presented in Table 2 demonstrates the preference for interviews as a primary data source in qualitative research, despite the availability of many types of data. In examining how these data relate to GT specifically, studies professing to use GT were isolated, and these results are presented in Table 3. The data in this table confirm a continued and increasing preference for the interview method in qualitative GT studies.
Grounded Theory Research Articles (Sociology and Qualitative Health Research, 2010-2012).
Reasons for the favored status of interviews indicated in Tables 2 and 3 are not immediately clear. It is apparent, however, that when grounded theorists move beyond the interview as a data source, inconsistencies emerge, particularly in relation to using documents. When referring to the use of documents as data, for example, the language is uncertain and inconsistent terminology is used. Some of the terms to describe documents include the following:
Caches of documents (Glaser and Strauss, 1967),
Textual data (Burnard, 1996),
Inert text (Prior, 2003),
Extant text (Charmaz, 2006),
Technical literature (Corbin & Strauss, 2008),
Naturally occurring material/written texts (Denzin & Lincoln, 2011), and
Literature as data (Birks & Mills, 2011).
Inconsistencies in nomenclature suggest a lack of consensus among grounded theorists in respect of dealing with the technical aspects of varying data sources and in establishing a clear approach to positioning all types of data sources in a GT study. This lack of consensus about the concept of documents as data is reinforced in the broader literature. Initially, Glaser and Strauss (1967) argue that documents can “hardly be used as a chief source of data” (p. 168), whereas Charmaz (2006) and Birks and Mills (2011) attest that documents can be used as primary or secondary sources of data.
Interacting With Data Sources in GT Research
Focusing on textual data is an essential part of developing a working knowledge of GT; if all is data, then all data become text at one stage or another. As an example, the interview is an interaction between researcher and participant. The researcher is actively involved in a process of producing data that is ultimately transcribed into text form. While the minutes of a meeting record a similar human interaction, the researcher has minimal (if any) control of data production in such a situation. The researcher is thus positioned very differently in respect of text produced from a meeting in contrast with that produced from an interview. Charmaz (2006) uses the term “extant text” to indicate data sources that the researcher had no hand in shaping (p. 35). Charmaz, therefore, distinguishes extant text from that which is “elicited” via research participants for a specific purpose or project.
In this article, we use Charmaz’s terms, “extant data” and “elicited data,” as they indicate a cognizance of the data source rather than referring to it simply as text. By using this distinction, we are keen to assert that data gathering is not merely a dichotomous process of elicitation. Figure 1 displays the spectrum of data source on which extant data are posited at one end, distant from elicited data at the other. Elicited data always involve an interpersonal interaction between the researcher and participant/s in an interview or focus group, or actors in a scene being observed. Extant data may take the form of existing text relevant to the study yet produced for purposes other than the research undertaking, such as data gathered from blogs or web forums. Other forms of extant data include questionnaires, surveys, or journals solicited by the researcher.
Spectrum of extant data and elicited data.
A human interaction is marked by a myriad of sensorial experiences, whereby the researcher has a broad spectrum of sensorial awareness about the data source before them. Even before the researcher engages in the process of analysis, they are influenced by sensorial experiences (engagement of senses) with the research material. The physical presence of extant data offers less to the researcher in the way of contextualizing data analysis, as the researcher is engaged in a lower level of sensorial experience than the one in which data are elicited. While sensorial experience is one means of contextualizing the data, the symbols present in interpersonal interaction are far stronger than those used in an interaction with extant data.
Although extant and elicited data may be at opposite ends of a spectrum of researcher–material interaction, it is clear that overlap exists. For instance, a best-selling novel or a blockbuster movie (interactive media) is an extant source of data that offers a more limited sensorial experience for the researcher to interpret than an interview where data are elicited. In other words, novels and movies may evoke a strong emotional reaction despite the viewer’s relatively passive position. Interviewing is a spectrally different sensorial environment as the sensorial experience for the researcher and participant through human–human interaction is far greater.
The Researcher and Extant Data
Despite the need to be cognizant of the distinctions of different data types and the requirement to approach data reflexively, grounded theorists are encouraged to treat extant data as they would any other data source relevant to a developing GT (Birks & Mills, 2011; Charmaz, 2006). To encourage as much is problematic, given that the moment words are transformed into text, a gap is created between the author and the data (Irvine & Gaffikin, 2006). To illustrate, the United Nations Translation Service uses professional translators to render written words from one language into another and interpreters to do the same with spoken words (United Nations, 2009).
The difference between translation and interpretation is clear; a word in one language is translated into an equivalent word in another language and, as such, denotes a unidirectional process. Interpretation, however, renders the spoken word of one language into an equivalent linguistic context. The interpreter is dealing with the speech of a dynamic, fast-flowing, inflective, emotive, instructive, directive person and attempting to derive an accurate interpretation of what is being said and placing it in a context foreign to where it belongs. This process is not unidirectional but omnidirectional in nature because the interpretation is heavily influenced by a multiplicity of human (including sensorial) factors.
In the same vein of translation versus interpretation, the position of the grounded theorist determines the extent of the interpretive spectrum in dealing with the data before them. When interviewing a participant, the grounded theorist is interacting in a dynamic, fast-flowing, process. Transcribing what is spoken is just that—a transcription; however, the undeniable influences present in any human interaction see this as a process of data generation (Birks & Mills, 2011) as opposed to that of data collection, which describes gathering something that is already there. The grounded theorist does not collect something non-existent, nor do they generate what is not present. Whether collecting or generating data, the grounded theorist must acknowledge their influence on the data source and recognize their position in the process of development of a theory that is grounded in that data.
The information elicited in the dynamic of researcher–participant interactions is a key difference between data generation and data collection. During data generation, the researcher as an interviewer can develop a deeper understanding of how to position the generated data for analysis as many questions are answered (either explicitly or implicitly) such as the “who, what, when, where, why and how” of context. For instance, the tones, inflections, gestures, and emotions of communication can inform how the researcher approaches data analysis.
Conversely, the researcher is often bereft of the context needed to optimally position extant data for analysis. In GT, data collection should not be a simple process of gathering artifacts, rather it should be a systematic and reflexive process aimed at collecting the data source and its concomitant information to optimally position that data for analysis. Data collection should not be an objectifying process, but rather a considered, reflexive undertaking that places data sources such as documents in a continuum rich with purpose, intent, interpretation, and context. It ought to be the researcher’s intent to find out “what is going on” (Glaser & Strauss, 1967) to establish an optimal entry point to analysis.
In summary, data generation and analysis occurs very differently to data collection and analysis, especially in the context of positioning documents. The tells given by the participant during interview are symbols that the interviewer interacts with and responds to in that context. The interviewer responds to not just what is said but how it is said. Their theoretical sensitivity to data from the participant is heightened in such instances, comparative to documents. Documents say what without illuminating the context in which it is said. The need to position documents is therefore necessary to restore the balance, not to turn the document into an open-ended interview. It is important to position the document to be theoretically sensitive to its possibilities as a data source, in its own unique context.
Positioning Extant Data in a GT Study
From the preceding discussion we can see that, in both the collection and generation of data, the position of the researcher and their interaction with the data source determine a context for analysis. Only in the case of data generation, however, is the researcher able to derive such knowledge from implicit and explicit means. Conversely, a researcher engaging in data collection must prepare the text for analysis by using a process of contextual positioning. Contextual positioning requires approaching the extant data to establish the important “who, what, when, where, why and how” of context. Contextual positioning is thus achieved through targeted questioning. This process is quite distinct from the analytical questioning of data used in approaches such as discourse analysis. Contextual positioning uses targeted questioning for the purpose of positioning data for analysis but is not intended as an analytical tool per se. When using documents as data, targeted questioning compensates for the decreased sensory involvement and symbolic interactions occurring between the researcher and extant data. No longer is the researcher privy to the moods, expressions, gestures, and tones of the interviewee or focus group. Extant data present a different challenge requiring a different approach. Targeted questioning acknowledges the differences in the nature and level of involvement and interactions that occur between the researcher and sources of elicited and extant data.
This process makes it possible to establish a three-dimensional context centered around the positionality and reflexivity of the researcher toward the data and its source, GT methods, and the research process collectively. Contextual positioning is vital to the development of a GT as it enables the researcher to situate the data in relation to the research study, thus facilitating a contextually relevant analysis of that data. All researchers instinctively assess data to some extent, but the use of a structured approach promotes most effective positioning of data that may otherwise be regarded as static. Table 4 proposes sample questions that the GT researcher can use for this purpose.
Sample Questions for Establishing Contextual Positioning.
While it may be argued that targeted questioning would arrive at answers that load the researcher with a priori assumptions before the data are analyzed, we believe that it is a contemporaneous, a posteriori process. The researcher may question the source or other associated information to arrive at a contextualized understanding of the data. This approach is no more likely to impose preconceptions upon the researcher than does establishing a relationship with an interview participant; in essence, targeted questioning allows the researcher to establish rapport with the extant data. Figure 2 illustrates the process by which collected extant data can be optimally prepared for use in GT research through contextual positioning.
Closing the gap through contextual positioning.
To illustrate the application and relevance of contextual questions, Table 5 presents examples of responses that may be derived from this process using a GT study of nursing education accreditation in Australia undertaken by the authors. Questions posed of the extant text (in this case—feedback provided to the nursing and midwifery regulatory authority in Australia about the national accreditation service) and the corresponding responses are presented to demonstrate the mechanism by which extant data can be placed on equal footing to the oft-preferred interview transcript. Contextual positioning through interrogation provides a three-dimensional context to support analysis, namely, the position of the extant data, the position of the researcher, and the position of the extant data to both the researcher and the study. Contextual positioning is a simple and effective method to scope the context in which the extant data and the researcher are situated.
An Example of Establishing Contextual Positioning.
While the sample answers are by no means definitive, they serve to elucidate the nature of the extant data and enable the researcher to approach extant data in the study with a greater level of awareness and reflexivity. By positioning the extant data through the use of contextual positioning, the researcher identifies the key actors relevant to the data, defines its scope, chronicles its position in time, locates it in an appropriate context, rationalizes and justifies its position, and explains its role as a thing and as a source of data. Through the use of contextual positioning, the researcher can take a thorough approach to locating and understanding the context and usefulness of the data. The researcher can be certain of the “groundedness” and cogency of data that contextual positioning provides to the research study. Such questions qualify the position of the extant data in the same vein that qualifying questions position the statement of an interviewee in a more appropriate context for analysis.
Conclusion
Having an awareness of the context of extant data in a GT study requires a concerted scholarly approach to establish consensus on the matter. The paucity of extant data in GT studies is of concern as is the dearth of literature on methods of preparing extant data for analysis in GT studies. We locate extant data through the use of contextual positioning as we ascribe to a view that context is inherent to analysis. Contextual positioning enhances the interactivity of the data collection process. No longer is the extant data source a static collection of letters, words, sentences, and paragraphs, rather, it presents as an enlivened thing, suitably contextualized, and ready to contribute to the development of a theory grounded in data in the hands of an informed researcher.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research and/or authorship of this article.