Abstract
Paradigms are often presented as a way of distinguishing various qualitative and experimental research approaches. But carefully tracing their adoption through the work of Egon Guba, we can see that the model of researcher-defined paradigms used within qualitative research also arose from the replacement of a more open conception of naturalistic inquiry (N/I); the anthropomorphizing of ideal research types; the inclusion of the subject matter in the characterization of different types of researchers; that it is the inclusion of the subject matter in Guba’s conception of a naturalistic inquirer that necessitates his appeal to philosophy (i.e., ontology and epistemology) as the basis for selecting methodology; and that by doing so Guba violates his own concerns about researchers choosing their methodology before considering their subject matter, something that he referred to as the law of the hammer. The adoption of researcher-defined paradigms also rejects the position that the appropriateness of a methodological approach, including N/I and qualitative approaches, is primarily determined by the subject matter and researcher’s objectives, something Patton has referred to as the paradigm of choices. This review of the origins of researcher-defined paradigms problematizes and defamiliarizes this core concept within some models of qualitative research. Given that Guba’s model and its appeal to philosophy as the basis for selecting methodology still underlies a fundamental division within conceptions of qualitative inquiry, reconsidering its development and potential alternatives will allow current researchers to better appreciate the model of qualitative research they choose to work under.
Keywords
Introduction
A dominant model of qualitative research is underpinned by the idea that all inquirers have their own researcher-defined paradigms, including their own epistemological and ontological positions, which are presented as determining their methodological approach (Guba, 1978, 1990; Lincoln & Guba, 1985; Denzin and Lincoln, 2017; Blaikie and Priest, 2017). I will call this model the Gubaian or researcher-defined paradigm model of qualitative research. I refer to it as Gubaian in honor of Egon Guba, who, with Yvonne Lincoln, greatly advanced this model and the practice of qualitative research more generally. The Gubaian model was developed over nearly 50 years of thinking and reflection (Preissle, 2008; Stufflebeam, 2008). It is made up of many threads and has been often revised. Given its long history, we can easily miss some of the thinking behind its adoption. Yet this model remains so ingrained in how some people see qualitative research that they equate the two, claiming that the qualitative research process should start with the definition of a researcher’s paradigm (Creswell & Poth, 2017; Denzin and Lincoln, 2017). This view can blind us to the fact that the paradigm model is only one way of conducting qualitative research; and may limit critical reflection on whether this model continues to be the best choice for qualitative researchers to work under.
In this article, I examine how the Gabaian model arose within qualitative research by following its development through the work first of Guba, and later, Guba and Lincoln (Preissle, 2008). It provides a detail description of the moves Guba makes, the rationale he gives for them, and the alternatives that he rejects to establish his paradigm model. This review will serve as a useful introduction to students trying to grasp the use of paradigms within qualitative research. Closely examining how the individual pieces of Guba’s conception of a paradigm came together will provide a better understanding of his whole position. For those who currently work under this model, some of the arguments and debates around its features occurred decades ago. Often what we are left with are the remnants and conclusions, having lost the perspective of what was really being argued against. Looking back to see the factors influencing the birth of this model will likely be surprising to some researchers who are committed to it. For those who think qualitative researchers should move beyond researcher-defined paradigms, fully understanding why the model developed the way it did provides a useful vantage point for making meaningful suggestions for moving the practice of qualitative inquiry forward. My main thesis is that when we look closely at the development of the paradigm conception used within qualitative research, we will see that it is both not as appealing and less fundamental to qualitative inquiry than it is often presented as being.
The Gubaian model has its roots not in the distinction between quantitative and qualitative research, but rather in the use of more naturalistic approaches to inquiry, particularly in education (Guba, 1965; Schwandt, 2013). The article begins by reviewing this rise of naturalistic inquiry (N/I) for evaluations in education. The next section details how the researcher-defined paradigm model arose from Guba’s more restrictive model of N/I. It then reviews his claim that different types of researchers hold different and often conflicting sets of assumptions, and that it is these assumptions which determine the type of research methodology that they pursue. I conclude by comparing two options for qualitative researchers - the researcher-defined paradigm model and the paradigm of choices model - as they relate to how inquirers make and defend their methodological choices, arguing for the preferability of the choices model.
The Rise of Naturalistic Inquiry in Education Research
In the 1960s and 1970s, an experimental approach was dominant for conducting most evaluations in education. Modeled on the natural sciences, this approach to evaluation used experimental designs, often focused on an educational intervention’s impact on a small set of discrete variables. An example would be if we sent a set of educational resources to some schools and not to others, and then examined differences in the average grades between the two schools to determine how effective the resources are. For the earliest part of his career, Guba, whose graduate work was in statistics and measurement, worked comfortably within this experimental approach (Schwandt, 2013).
But Guba and others (Denzin, 1970; 1971; Parlett & Hamilton, 1976; Patton, 1975; Willems and Raush, 1969) began to recognize the limits of the experimental approach, particularly for evaluating educational interventions. Some of the issues these authors saw were the inability of large experiments to capture the unique experiences of distinct or minority populations; their inability to appreciate facilitators or barriers within specific contexts, e.g., what was influencing the use or impact of the intervention within specific classrooms; the focus of the experimental approach on a small set of variables, which do not adequately reflect the complexity of most learning environments; the difficulties for experimental designs to be altered to incorporate new information or early findings; and perhaps most importantly given the amount of public funds being spent, was the limited uptake and usefulness of results from experimental evaluations for frontline educators. These reasons still hold for why experimental research designs are not appropriate, or on their own are insufficient, to adequately evaluate most educational interventions (Christ, 2014).
By 1965, Guba held that for the evaluation of educational programs, “there are two general strategies available to an investigator who seeks to inquire into a given set of phenomena, which I shall characterize as essentially experimental, manipulative, and interventionist, on the one hand, and aexperimental, observational, and laissez faire, on the other” (Guba, 1965, p. 1). One way to conceive of these different approaches is in terms of the different levels of constraints they put on data collection. For the experimental approach, the investigator places a high degree of structure and constraints on the antecedent variables in a study. For example, in a psychological lab experiment, the investigator may control who comes into the study, the place where the study is conducted, the conditions under which participants are given the intervention, etc. For the aexperimental approach, which Guba and others later refer to as N/I, the investigator places a low level of antecedent constraints, focusing on observing participants in their natural environment. While mostly absent from education research at the time, this naturalistic approach was already being used in fields like anthropology and naturalistic biology.
Guba presents these two strategies as a dichotomous choice between the naturalistic and experimental polls. In practice, studies and evaluations likely exist across a continuum between the two polls. For example, focus groups using an interview guide place more constraints on the antecedent conditions of the study than a purely observational design. Guba (1965) says that the investigator or evaluator “can identify a crucial phenomenon and pursue it with whatever techniques seem appropriate” (Guba, 1965, p. 1). At this point for Guba, it is the intentionality of the inquirer when considering their subject matter that directs their methodological choice. This allows for the same researcher to use different strategies if they choose to study different topics or have different study objectives. But Guba argues that the aexperimental strategy is preferrable for most change and evaluation research, particularly for change processes in education. In fact, he sees the aexperimental strategy as being preferable for a wide range of social science research.
Further developing his conception of the naturalistic aexperimental approach, Guba (1978) says that he borrows from the analysis of Willems and Raush, 1969.
1
Willems positions inquiry across “a two-dimensional descriptive space,” based not on the nature of the phenomena, but by the choices made by the investigator in conducting a study (1969, p. 46). The first dimension is the level of manipulation of the antecedent conditions made by the investigator. The second dimension that Willems adds is “the degree to which units are imposed by the investigator upon the behavior studied” (1969, p. 46). The data instruments and the study design will impact the type of data collected from study participants. This could range from audio and video recording of participants, which put few constraints on outputs, to the collection of discrete quantitative measures. Although he does not discuss it directly, this imposition of the investigator on the outcome units could presumably also occur during data analysis. Figure 1 replicates Willems’ two-dimensional descriptive space for describing research activities (Willems and Raush, 1969, p. 47). “A space for describing research activities” recreated from Willems & Rauch (1969, p. 47).
For Willems and Raush, 1969, “each of these functions [i.e., dimensions] can vary on a continuum” (p. 47), with studies being located across this entire space. For example, a medical observational study based on the secondary use of patient data would put a low level of constraints on the antecedent conditions, as the data would not have even been collected within the context of a study, but may put a high level of constraints on the outcomes that are abstracted from the patient charts (i.e., it would be located near the Low-High area). More naturalistic studies, like some participant observation studies that do not place restrictions on outputs, would be nearer the Low-Low area of Willems’ descriptive space. Although he does not discuss it directly, it seems that qualitative methods can potentially be used anywhere across this space, although they may be less likely the more constraints the investigator places on the outcome units. One of the advantages of defining naturalism in terms of the study choices of the investigators is that it avoids the “polemics and petulant, empty argumentation” that often occurs over words like ‘natural,’ ‘scientific’ or ‘true-to-life’ (Willems and Raush, 1969, p. 44). It also presents a framework that allows for the location of studies which is more judgement neutral. Depending on the topic and the intentions of the investigator, some studies are better pursued with different levels of constraints. It is up to the investigator to determine where in this space they will design their study. That choice should be made consciously and transparently. This approach to defining naturalism is also flexible enough to locate most behavioral studies and is not restricted to human studies. Willems and Raush (1969) includes articles outlining the use of N/I to study topics as varied as the behaviors of children in classrooms, primates, and even captive white-footed mice. Willems and Raush, 1969 warns that “the controversies [over naturalistic and experimental research] usually focus on the most extreme cases, and upon cases that represent only a small proportion of the possibilities” (p. 49). His conception of naturalistic research, which is defined by the methodological choices of the investigator and is presented in his space for describing research activities, avoids this focus on extremes.
Guba’s Researcher-Defined Paradigm Model
The researcher-defined paradigm model has its roots in the transition from Willems’ model (Figure 1) to the one presented by Guba (1978). While he says that he is still aiming to locate N/I and that he based his table on Willems’ model to do it, carefully comparing the two tables, we can see that Guba make several fundamental though unacknowledged changes that greatly impact the nature of his model (Figure 2). First, Guba reduces the two continuous dimensions to two segments: low and high (Guba, 1978, p. 9). This segmentation moves the continuous space of inquiry presented by Willems into a categorical type of thinking. Guba then adds solid lines around his boxes, which reinforces the categorical and oppositional nature of his model. It is no longer a space or a continuum describing the choices that investigators can make when studying a phenomenon. The space is reduced to four boxes. The solid lines imply that it is difficult for reconciliation or transferring back and forth across the boxes to occur. Rather than describing it as “a space for locating research activities,” Guba refers to the quadrants as distinct “modes of inquiry.” Finally, rather than providing markers for locating individual studies (e.g., Low-Low, Med-Med), Guba emphasizes the extremes of the ‘ideal’ naturalistic inquiry and the ‘ideal’ experiment. He does not directly describe examples of these ideal studies. Based on the dimensions in the table, an example of an ‘ideal’ naturalistic inquiry would be a participant observation study in which participants do not even know that they are being observed and the data collected is unaltered audio and video of the natural interactions. An example of an ideal experiment would be either a clinical lab experiment which restricts participants’ responses to discrete choices or some experiments in the natural sciences, e.g., Galileo’s experiment of dropping balls off the Leaning Tower of Pisa to measure the speed at which objects fall to the earth (Guba, 1967). This focus on extreme cases is exactly what Willems warns against. Representation of inquiry space or domain of inquiry recreated from Guba (1978, p. 10).
Yet Guba restricts his conception of N/I even beyond the changes made to Figure 2. Willems’ naturalism is flexible enough to locate most studies, including non-human studies (Willems and Raush, 1969). Guba says that his conception of N/I is primarily concerned with “empirical inquiries that require interaction with human subjects” (Guba, 1978, p. 11). But it is not just any human studies to which Guba restricts his conception. Guba quotes Bogdan and Taylor saying “the phenomenologist is concerned with understanding human behavior from the actor’s own frame of reference…. [he or she] examines how the world is experienced” (Guba, 1978, p. 12). At the end of this section, Guba says that “the naturalistic investigator is a phenomenologist.” By 1978, for Guba, N/I is no longer presented as the preferable approach for studying many human behaviors and interactions, including educational evaluations. On Guba’s new conception, the main subject for the naturalistic inquirer appears fixed on understanding human perspectives and experiences. While setting out this subject for a program of research is fine, incorporating this subject matter into his conception of N/I creates a problem for his position. If he incorporates its subject matter into his conception of what N/I is, he cannot then also use the subject matter as the basis for choosing between the experimental, naturalistic and other research approaches. As he himself recognizes, his very conception of naturalism creates the problem for him of having “no basis for speculating about why an investigator would prefer N/I” (Guba, 1978, p. 10), when that is not a problem for those who hold less restrictive conceptions.
To address this issue, Guba extends his conception, proposing that N/I “also differs from other modes with respect to its basic assumptions and characteristics” (Guba, 1978, p. 3). These assumptions are presented as not just the implications of conducting a study within a particular mode or quadrant of research. While previous definitions use ‘naturalistic’ to qualify the type of methods used, Guba (1978) now proposes that there are different types of inquirers associated with each of the quadrants he constructed, i.e., that there are naturalistic inquirers who hold certain assumptions. The focus has moved from types of research projects to types of researchers. The freedom of researchers to move around Willems’s descriptive space is presumably lost. Guba does not provide detailed descriptions of the assumptions that he sees inquirers in all four quadrants holding, focusing only on the extremes of the naturalistic and experimental polls. He also does not discuss how these descriptions are meant to relate to inquirers in the real world. In a footnote, Guba says that “related analyses may be found in Patton (1975), Parlett and Hamilton (1976) and Shrock (1977)” (Guba, 1978, p.11). As with his adjustments to Willems’ model, Guba again makes fundamental changes and leaves out important information from the work that he claims supports his position without indicating or arguing for these changes.
A paradigm can be seen as a set of assumptions that underlie a practice of research for a scientific community (Kuhn, 1970). Patton (1975) and Parlett and Hamilton (1976)
2
recognize the dominant paradigm of evaluation at the time as being modeled on the scientific method (e.g., the use of experimental designs, the privileging of quantitative measurement, variables, and statistical methods, etc.). They both present the case for an alternative paradigm. As Patton says, this “alternative paradigm stresses understanding that focuses on the meaning of human behavior, the context of social interaction, an emphatic understanding of subjective (mental, not nonobjective) states, and the connection between subjective states and behavior” (Patton, 1975, p. 13). This alternative paradigm is in keeping with the previous arguments for the greater use of N/I in evaluations. To present this alternative paradigm as clearly as possible, Patton (1975) says he sharpen the “contrasts [between the scientific and alternative paradigms], to bring them into high relief, to make them appear as opposites. Such an analysis, based on non-existing ideal-types, will clearly overstate the case. Tacit understandings about flexible parameters will here appear as absolute rules of procedures. Areas of mutuality, common concern, and similarity of commitments will be largely ignored” (p. 15, emphasis in the original).
Patton uses a dramatic explanatory approach of contrasting non-existing, ideal, and extreme research types. His rationale for doing so was that the scientific paradigm was so pervasive within the education evaluation community at the time that this level of contrast was required for others to recognize the assumptions underlying their own research. Although his exposition presents these two approaches as conflicting, Patton soon says that “the debate and competition between paradigms is being replaced by a new paradigm - a paradigm of choices… [which] recognizes that different methods are appropriate for different situations” (Patton, 1980, p. 20). For Patton, there are clear implications in selecting a N/I approach that impact on how the study should be conducted and evaluated. But it is up to the researcher to decide which methodology is best suited for a research project. Presented in this way, it is easy for scholars trained in experimental design to both recognize this alternative approach and advocate for its use in appropriate situations (Patton, 1980, p. 19).
Patton’s position is in keeping with that taken by Guba in his earlier work (e.g., Guba, 1967). But by 1978, Guba adopts a more constrained conception of N/I. While Guba (1978) says that he derives his conception from Willems, his naturalism is actually more in line with the adoption and extension of the alternative paradigm outlined by Patton (1975), while only fitting its exposition into a table similar to the one presented by Willems. We can see this break with Willems by examining the assumptions Guba attributes N/I.
But Guba also breaks with Patton in significant ways. Patton is comparing two broad research approaches. Guba focuses on types of inquirers. Patton (1975) is clear that his characterizations are artificial ideals, meant only to contrast the scientific and alternative paradigms. Guba omits any description of his characterizations being an explanatory approach, nor does he characterize his depictions as either extreme or ideal. In fact, Guba starts to anthropomorphize his characterizations, seeming to relate them more closely with the actions of real people. For example, Guba says that his different inquirers “tend to see the world from quite different perspectives” (Guba, 1978, p. 11). Although these different ways that researchers tend to see the world are supposed to be profound enough to determine their methodological choices, there is no analysis of how someone gets their perspective and whether they can change it. There is also little analysis provided to support Guba’s unique claims about there being two different types of inquirers who hold fixed sets of assumptions; nor does he describe where researchers who do not fit into either extreme quadrant fit into his model.
Guba’s Move into Philosophy
Potential Types of Concerns Related to A Research Project.
aWillems and Raush, 1969, p. 2-4.
For Guba, the issue of appropriateness of a research approach is settled not at the level of methodology nor at the level of the philosophy of science, but at the level of metaphysics, in particular questions regarding the nature of reality or ontology. Guba extends his conception of naturalism by claiming that evaluators in education research differ in terms of their “philosophical base,” and that it is their philosophical positions which determine their methodological approach. He does not provide any argument for his radical move to say that a research approach is determined by a researcher’s metaphysical position, but again it seems motivated by the lack of other basis for deciding on a research approach given his conception of naturalism. Guba holds that the “naturalistic investigator is a phenomenologist while the conventional inquirer is a logical positivist” (Guba, 1978, p. 11). Guba introduces these philosophical terms without much discussion of how and the extent to which they are meant to reflect different approaches in evaluating educational programs. In academic philosophy, phenomenology refers to “the study of structures of experience or consciousness” (Smith, 2018). It is a tradition associated with writers like Husserl, Heidegger, Merleau-Ponty, Sartre, and others. Although it is harder to define logical positivism (or logical empiricism as it was later called), it is the movement loosely focused on the empirical verification of meaning, anti-metaphysical positions, a rejection of realism, and a value placed on scientific methodologies (Ayer, 1952: p. 40-41; Creath, 2021; Liston, 2023; Isaacson, 2007: p. 460–462). It is a movement associated with writers like A.J. Ayer, Rudolf Carnap, early Wittgenstein and others. Both phenomenology and logical empiricism are complex philosophical traditions, with many disputes about how they are best defined and approached. Guba does not reference any of these writers nor does he provide analysis of their positions and their potential relationship to educational evaluations (Guba, 1978). It seems clear that Guba is not implying that evaluators of educational programs have considered in detailed and accepted one of these philosophical traditions, both of which can be seen as fading philosophical forces by the second half of the 20th century. It is also clear that he is not using these terms in the way they are used in academic philosophy.
In his section on philosophical claims distinguishing N/I and experimental approaches, Guba also contrasts the views of Max Weber and Emile Durkhiem. It is not clear whether Guba is trying to assign these writers to phenomenological and logical positivist traditions. The positions of both are complex and again it would be interesting to consider in depth how their work would have implications for the evaluation of educational interventions in a modern American classroom setting. But again Guba does not provide any analysis of either author to identify specifically what is being rejected by proponents of his conception of naturalism when they reject Durkhiem and what impact it would have on educational evaluations. Given the briefness of the analysis Guba gives, one has the sense that Weber and Durkhiem are again just being used here as metaphors for the experimental and naturalistic positions. Yet on the Gubaian model, these philosophical positions are supposed to be clear enough to allow evaluators to determine their methodological approach.
Does philosophy select methodology?
A third element that Guba raises in his short section on the philosophical assumptions which are meant to preselect a N/I methodology is that positivists are concerned with strong generalizations in relation to psychology. It is the case that the inquirers who use an experimental approach are likely to focus on generalizations. Yet the directionality of how this works is important and reflects the differences between whether the selection of a research approach is made at the methodological or philosophical level (as claimed by Guba) (Figure 3). On the choices model, the inquirer first examines their topic and determines which methodological approach is appropriate. If it is a naturalistic approach, they design their study to meet the implications of conducting a N/I study. If they think an experimental approach is appropriate, they will employ an experimental design which will likely include a focus on variables and aim at generalizable conclusions. The processes for selecting a methodology through the paradigm of choices and the researcher-defined paradigm models.
Guba claims researchers first have an ontology, i.e., a view of the nature of reality; and an epistemology, i.e., a view of the type of knowledge claims that can be made based on that ontology. It is these assumptions that decide their methodological approach, without considering the subject matter. That the researcher’s assumptions select the methodology maybe initially appealing to researchers who see themselves only as naturalistic inquirers, i.e., who would reject any use of an experimental approach. For the supporter of the researcher-defined paradigm model, the differences are philosophical, not in the sense that they are derived from the study of philosophy, but rather in the sense that they are foundational to how they see themselves, and importantly, other inquirers. Rather than just seeing it as a powerful methodological approach, they go all in, totally identifying themselves as researchers who will only use naturalistic approaches; and view all other researchers similarly as only accepting a single methodological approach that fits with their preexisting suppositions.
Ironically, Guba himself warned against researchers having a prior preference of a research approach, which is the element he inadvertently builds into the foundation of his researcher-defined paradigm model. Willems and Raush, 1969 says that there are two often overlooked reasons why discussions about N/I and experimental designs “have a peculiar tendency to deteriorate into polemics and petulant, empty argumentation” (p. 44). One reason is the “empty disputation” about what it means for methods to be more natural or more real. The second reason “is what Professor Egon Guba once called ‘the law of the hammer.’ The law says: If you give a child a hammer, things to be pounded become the most important things around. In the society of research investigators, the analogue to this law points to the fact that we often not only allow our favorite methods and techniques to dictate the choice and merits of our own research problems, but we often let our own favorite methods dictate and evaluate the merits of someone else’s research problems. Believing in the inherent, a prior correctness and scientific efficacy of certain methods, we prescribe and proscribe, praise and blame” (p. 45).
By claiming that inquirers hold different sets of assumptions that lead them to select what is the appropriate methodology for them, Guba is arguing that methodological choices are made a priori and not based on the consideration of a subject matter. The model he produces (Figure 2) enshrines an oppositional stance towards other researchers. By claiming these differences are philosophical and personal, he moves the choice of methodology beyond considerations of how best to advance the study of a subject, but rather it is pre-selected based on who you are as a researcher. In short, the shadow of Guba’s ‘law of the hammer’ hangs heavy over his conception of researcher-defined paradigms.
Does ontology really select methodology?
Guba claims that it is differences in “the inquirer’s view of reality” (Guba, 1978, p. 14) which lead to different methodological approaches being selected. This claim leads to qualitative researchers preoccupation with ontology (Creswell & Poth, 2017; Denzin & Lincoln, 2017; Guba & Lincoln, 1994). But does one’s ontological position really select a methodology for a study? Guba’s analysis focuses on ideal research types. But once we consider more realistic and diverse examples of studies, we can see that the view that there are profound differences in how inquirers view reality, and that these differences pre-determine the selection of a methodology, seems hard to justify.
Let us consider two possible studies. In the first study, imagine that we observe children in a school playground. To make sure that their behavior is not impacted by our study, we do not let the children know that they are being observed and place no conditions on their participation. We video all the interactions, not putting any restrictions on the outputs. According to Willems’ definition, this would be a N/I. But suppose we are not concerned with the children’s experiences of the situation, rather aim only to determine which piece of playground equipment was used the most. Would it make sense to say that we chose the naturalistic approach because of our view that reality is actually multiple realities? Given Guba’s restriction on the subject matter for N/I, he may want to reject that it is a N/I under his definition. But if it is not a N/I, what type of methodology would it be best classified as? And, presuming that a view of reality or ontology still determines methodology, what view of reality would help us select it?
For the second study, imagine we are testing a new cognitive behavioral technique. During the study, data is collected both before the intervention and after to test its impact. The selection of participants can be conducted using random selection and based on a sample size calculation. We could put restrictions on the context in which data is collected and try to control for factors which we think may impact the results. In many ways, the study would be an experimental design. But what if the inquirer recognized that everyone’s experiences of the intervention are different, so that the data for the study are in-depth interviews to understand each person’s unique reality while using the intervention. Although the study is experimental in design, given its focus on understanding the experiences of individual people, i.e., exploring their multiple realities, is it reasonable to claim that its methodology is chosen based on a different view of reality than Guba’s N/I? It is hard to see that it does. In other words, both N/I and experimental studies of experience would not differ in their view of the reality of the nature of their subject matter.
Guba’s underlying claim that it is ontology that determines the selection of a study design is very unique. In terms of other naturalistic inquirers, this appeal to ontology is not found in his own earlier work, in the articles by Patton (1975) or Parlett and Hamilton (1976) that he says supports his position, nor in Willems and Raush’s work from which he says his position is derived. Ontology is not discussed in the work of other early supporters of qualitative methods or N/I (Glaser & Strauss, 1967; Denzin, 1970, 1971). Considering it from the perspective of academic philosophy, no philosopher of science takes the position that investigators need to have a clear ontological position before selecting between a N/I or experimental approach. In Effective Evaluation (1985), Guba and Lincoln describe their conception of N/I as being a paradigm. The modern concept of a paradigm arose within the work of the American historian of science, Thomas Kuhn (Kuhn, 1970). It is illustrative that Kuhn’s survey of the history of the natural sciences and his conception of a paradigm do not refer to or include ontology or epistemology (Chafe, 2023). The terms and their roots are only briefly mentioned in The Structure of Scientific Revolutions. Kuhn’s conception also allows for scientists to work without having or knowing that a paradigm exists (Kuhn, 1970; Masterman, 1970). None of the other major paradigms that Guba and Lincoln (1985) identify as being on par with their conception of N/I – Logical, Scientific, Judgmental, Adversarial, Modus operandi, Demographic – begin with an appeal to ontology. It is unclear then why such a dramatic reconfiguring of our conception of research and the embedding of philosophical suppositions into the research process are required in order to use N/I in the evaluation of American school programs (Guba, 1978).
Given the role of ontology in the researcher-defined paradigm model, it is surprising how little analysis is done by supporters of this model to determine and debate different ontological positions. For example, Guba’s original presentation of the role of different views of reality in selecting a methodology is less than one page in length (Guba, 1978, p. 15). This section also does not reference the work of a single philosopher. There are clear mistakes in his analysis, for example, saying that logical positivists are realists, which is at odds with their radical empiricism (Ayer, 1952; Liston, 2023). There is no analysis of how researchers adopt an ontological position, even given the implication that it is ontology which determines the type of researcher that you are and the methodology that you will use. If ontology was serving the role that Guba’s researcher-defined paradigm model says it does, you would expect it to be the key focus of attention. I suspect the reason it is not included is that one’s view of reality does not actually determine methodology, even for Guba; and that it is mainly that he inadvertently includes his subject matter in his conception of the naturalistic inquirer which leads him to begin to posit that it does.
Guba’s Paradigm and Qualitative Research
By 1985, Guba and Lincoln combine many of the elements described above into their conception of the naturalistic paradigm. These elements include the implications of conducting naturalistic inquiry (i.e., the importance of context, lack of concern with generalizability), the characterization of ideal types of inquirers, the anthropomorphizing of these conceptions, the idea that philosophical assumptions select methodology, and the focus of N/I on studying human experience. Given all of that, it is perhaps surprising that they also want to hold on to a choices model. They say “the choice between paradigms in any inquiry or evaluation ought to be made on the basis of the best fit between the assumptions and postures of the paradigm and the phenomenon being studied or evaluated” (Guba & Lincoln, 1985, p. 56). But given everything that they have loaded into their naturalistic paradigm and that they only offer their idealized characterization of the experimental paradigm as an alternative, it seems contrived now to allow the subject matter to determine the appropriate paradigm. Nor do they seem to appreciate that their use of philosophical assumptions as the basis for selecting a methodology ultimately excludes the choices model.
Even at this point, Guba and Lincoln do not equate their naturalistic paradigm with qualitative research. They say that “it would surely be a gross error to equate quantitative methods with the scientific paradigm and qualitative methods with the naturalistic paradigm” (Guba & Lincoln, 1985, p. 64–65). Yet by at least the publication of the first authoritative Handbook of Qualitative Research (Denzin and Lincoln, 1993), the Gubaian paradigm model is welded into what is presented as qualitative research. One indication of its influence is that Guba and Lincoln’s presentation of researcher-defined paradigms in the first Sage Handbook has been cited by other authors over 33,000 times (Guba & Lincoln, 1994; Google Scholar, 2023). Since then, they have expanded the range of available paradigms that can be adopted, but the underlying model remains mostly unchanged (Lincoln, Lynham and Guba, 2017). The Gubaian paradigm model is presented still in the sixth edition of the Handbook, due out later this year (Lincoln, Lynham and Guba, 2023). For many people who identify as qualitative researchers, defining their paradigm and just as importantly, demanding that other people define theirs, has become fundamental to how they think research should be conducted. For many younger researchers, the questionable foundations of the Gubaian model and the viable alternative to it have been mostly lost to history.
Conclusion
This article provides a review of how Guba’s researcher-defined paradigms arose and the alternatives that it replaces. Given the added complexity and the lack of thorough analysis that underpins the early development of the researcher-defined paradigm model, it is interesting that the model has become as dominant as it has, that it has not been questioned more. One reason is that careful analysis examining how the different aspects of this model were developed is difficult. Many of the new insights in this article came from not only analyzing Guba’s original articles and conference papers, but also carefully reviewing the work he said he took his ideas from. One limitation is that some of these early articles and conference papers were unpublished and could not be located to include in this analysis. Others would have been lost without the Education Resources Information Center (ERIC) database. For many people learning qualitative research, researcher-defined paradigms are presented to them as a combined whole and is not presented with other viable models which could be palatable for researchers to adopt. Most current qualitative researchers learned to talk about their research in terms of paradigms. Rather than seeing them as being a constraint, they may feel at home talking about defining their paradigms and claiming that the selection of their research methodology is philosophically based.
A concern that drove both Guba and Patton to adopt more N/I approaches was the recognition that there is a wider range of stakeholders who could potentially use their evaluations. Qualitative research has the potential to make greater contributions to many fields, but restricting it only to what accords with the current qualitative paradigm greatly reduces the impact it could have. Many research problems in fields like health, business and education would not fit within the current framework of researcher-defined paradigms. As Glaser (2004) points out, the characteristics of Lincoln and Guba’s naturalistic paradigm “restrict the flexibility and autonomy of the researcher” (p. 4). The choices model is better suited for those researchers who want the freedom to explore all of Willems’ space for inquiry, rather than be confined within Guba’s boxes. Given the added burdens associated with researcher-defined paradigms, especially in the context of multidisciplinary or multi-stakeholder teams or mixed methods research, we should seriously consider whether the discussion of paradigms still make sense for most qualitative research projects.
Guba concluded one of his first papers in which he argued for the greater use of N/I by saying “if I have not convinced you of the utility of my particular suggestions, I trust that I have at least given you cause for concern regarding conventional strategies. If I have planted a festering doubt, more than half of my purpose will have been accomplished.” (Guba, 1965, p. 38). In this article, I traced the early development of his conception of researcher-defined paradigms and present an alternative to it: a choices model. By clarifying how the Gubaian model developed and showing that this model of qualitative inquiry arose as much from a rejection of other conceptions of N/I as it did from the attempt to distinguish N/I from the experimental approach, I hope to make us pause to reconsider its value. To recognize that the original suppositions that lead us down the path of tying researcher-defined paradigms to qualitative research are not as strong as some may think. It is an attempt to create a space for those qualitative inquirers who do not want to work under the Gubaian model. If I have made you even consider that possibility, more than half of my purpose will have been accomplished.
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, authorship, and/or publication of this article.
