Research designs that yield data that are to be analyzed using statistics or other mathematical calculations, typically referred to as quantitative research, are preferred by many human science researchers over qualitative research designs, which yield data generally analyzed through processes that rely on researchers’ direct and reflexive engagement. This preference might be due to the belief that data processing is quicker for quantitative designs, which is often true, or the perception that findings from quantitative designs are more useful, which is not necessarily true. Qualitative inquiry that is conducted in a systematic, transparent, and rigorous way potentially yields findings that are both enlightening and credible. Along with this contribution, qualitative inquiry is particularly useful for understanding many of the research questions and interests that inspire gender researchers due to its key attributes that include respect for and appreciation of participants and their realities, the ability to explore not just questions of what but of how and why, and the availability of multiple approaches that range from being highly prescribed to extremely flexible in application.

My aim with the present paper is not to expand on the comparison between quantitative and qualitative research methods but rather to encourage gender researchers to design, carry out, and report findings from useful and thoughtful qualitative studies in a way that allows readers to engage with, have trust in, and derive meaning from research articles. The overall purpose of my paper is to provide general guidelines to help authors frame qualitative research papers for publication in Sex Roles. My paper is not meant to substitute for Hesse-Biber’s (2016) informative editorial about foundations of qualitative and mixed methods designs, nor is it meant to provide in-depth guidance to plan, carry out, and report findings from any specific type of qualitative study. Instead, my paper is directed toward authors who are in the process of synthesizing and reporting findings from research conducted within any of the various approaches that compose qualitative inquiry. To satisfy this broad goal, while acknowledging that variety and creative expression are frequent attributes of qualitative research reports, I provide recommendations and considerations more often than explicit directives. I also offer the disclaimer that adherence to these suggestions is not a guarantee of acceptance; myriad factors influence whether a given work is published in Sex Roles.

The content in this paper is divided into three sections. In the first section, I define qualitative inquiry and offer general recommendations to enhance quality and readability of qualitative research reports. In the second section, I classify multiple qualitative approaches and provide recommendations to help authors best describe their research for the interdisciplinary audience that comprises Sex Roles readers. In the third section, I identify and critically assess common practices in qualitative research reporting.

Reporting Qualitative Research

The distinguishing terms qualitative and quantitative are used to refer to the following in combination: purpose or question, type of data that are gathered, and methods of analysis. This definition from Saldaña (2011, pp. 3–4) acknowledges the first two:

Qualitative research is an umbrella term for a wide variety of approaches to and methods for the study of natural social life. . . data collected and analyzed is primarily (but not exclusively) nonquantitative in character, consisting of textual materials. . . or visual materials. . . that document human experiences about others and/or one’s self in social action and reflexive states.

There are a variety of ways that qualitative data analysis is performed. Coding, a systematic process of selecting excerpts of text and assigning a summarizing word or phrase to that text, is very common but is just one option for analysis. Frequency estimates often play a role in how qualitative findings are developed and reported, despite a more natural association with quantitative methods. By frequency estimates, I do not mean that findings are presented as counts or proportions, rather I acknowledge that prevalence and recurrence of similar content tend to drive findings. Regardless of variation in analysis methods, findings from qualitative research, consistent with Saldaña’s characterization of qualitative data, are generally presented as descriptive text, visual expressions, or as other nonnumerical expressions.

Essential Elements

For the most part, the essential standard elements of qualitative research papers are the same as the essential elements of any research report, and the essential elements of qualitative gender research reports resemble those of any qualitative research report. These essential elements include: (a) a clear statement of research purpose or question; (b) a rationale for conducting the research; (c) justification for research design decisions; (d) thorough, but not minutely detailed, descriptions of methods, including sampling decisions and data collection, data management, and data analysis processes; (e) a summary of findings that resulted from the study; (f) a discussion section including argument for the contribution and implications of the research findings, and (g) honest confessions as to the shortcomings or compromises involved in carrying out the research. Authors who make a good faith effort to describe these elements in a clear and revealing way have made a good start toward producing a quality research report.

The Role of Checklists

There are many available checklists or rubrics that help students, reviewers, or others identify the presence of attributes and develop a quality rating for research reports. Such a checklist might also serve authors both as a guide that provides directions and as a task manager that tracks progress toward completion of a report. Among many available, I describe two in the following that are frequently used and readily available for authors preparing articles for Sex Roles.

The Consolidated Criteria for Reporting Qualitative Research (COREQ; Tong et al. 2007) includes 32 items that are recommended for inclusion in qualitative research reports, such as “Where was the research conducted?” Critical Skills Appraisal Programme (CASP; Critical Skills Appraisal Programme, 2017) consists of nine items, such as “Was the research design appropriate to address the aims of the research?” scored as yes, no, or “can’t tell,” with one additional question of judgment, that is, “Is the research valuable?” (http://www.casp-uk.net/casp-tools-checklists).

The lengthy list of items on the COREQ guides authors toward development of a comprehensive report, although the wording of items does not allow one to calculate a relative grade or score for a given research project. Not all COREQ items are direct indicators of quality (e.g., “What was [the researcher’s] occupation at the time of the study?”) and some do not apply to all designs (e.g., the practice of returning transcripts to participants for review). As an alternative, CASP provides a reasonable starting point, and the shorter but broader list of CASP items is likely to apply to nearly any qualitative study. Although the developers do not endorse a rating system, it is possible to compare the proportion of yes to no responses to determine to what extent the report addresses the criteria.

Use of a checklist does not guarantee that the research report meets any particular quality standards. Authors of a report might include checklist-recommended information, such as an explicit rationale for the use of a given qualitative approach, and then go on to describe a process that is inconsistent or incoherent with the chosen approach. Additionally, quality assessment tools often disproportionately focus on methods. This focus might encourage authors to compose a lengthy Method section that far exceeds what is needed to establish credibility and results in a tedious experience for readers who are often more interested in findings than process. Another limitation of qualitative checklists is that it is not possible on any single form to address every quality standard that applies to every approach, or combination of approaches, possible within qualitative research, so for many studies, checklists contain irrelevant items and/or lack relevant ones. Despite these limitations, I have found checklists helpful when used as part of a reflexive approach to writing a qualitative research report. Checklist items remind authors not only to include critical information but also to communicate and defend decisions that were made in the field in response to changing contexts or other emergent issues. Given the advantages and cautions I expressed here, I advise authors using a checklist to prepare an article for submission to Sex Roles to strive to balance quality of research conduct with comprehensive, yet concise reporting. Note that in Sex Roles it is possible to make use of an optional online supplement with your paper that is available to readers electronically to report methodological details that may be of interest only to select readers and that may be valuable to report to increase credibility and for full disclosure.

Engaging Readers

The Publication Manual of the American Psychological Association (American Psychological Association, 2010) provides comprehensive guidance to authors to encourage accurate but concise reporting. Although clarity, precision, and concision all contribute to readability, I also recommend that authors not overlook the need to engage readers. Qualitative researchers are uniquely positioned to write reports that include vivid, evocative, and moving language, often just through thoughtful use of (relevant) quotes from participants. This does not mean qualitative research reports should showcase bias, slanted or opinionated reporting, or language that is critical or inflammatory, unless this is representative of the findings or otherwise essential to the authors’ purpose. As always, knowledge claims must be supported by empirical evidence.

The process of qualitative inquiry establishes connections between researchers and participants that can lead to deeper understandings of human concerns and experiences. Thoughtful use of language encourages connections between authors and their readers, and it helps readers better understand the nuance that underlies findings presented as summarizing themes, processes, experiences, or via other composite or aggregate forms. Authors’ use of engaging language can expand the appeal of research reports, and Sex Roles, as an interdisciplinary journal, benefits when authors craft reports that have appeal across disciplinary lines.

In particular, qualitative researchers might consider exercising a little creativity in titles of categories, themes, or other elements derived through data analysis. In my experience, authors of qualitative research reports in fields dominated by quantitative research, such as STEM fields, tend to be very practical and literal in assigning names to the themes they develop and report in order to summarize findings for their research reports. Authors might consider how to use a theme to convey not just the practical but also the personal or emotional aspects they wish to represent. To this end, use of metaphor, alliteration, cliché, or meme, or attempts to draw suitable parallels with recognizable aspects of folklore, fable, or fiction, are strategies authors might consider as they report their findings—to the extent these practices enhance rather than detract from readers’ ability to appreciate the authors’ interpretive expressions. One reminder is that Sex Roles is an international, as well as multi-disciplinary, journal so readers benefit from bracketed explanations when authors refer to specific contexts or aspects of popular culture (e.g., Beyoncé [a popular U.S. singer]).

Authors frequently incorporate quotations from participants within the titles of works; this practice is an extension of “in vivo” coding (using participants’ own words as codes) and might also yield category or theme names that are both compelling and authentic. For interested readers, I recommend as examples of engaging and credible research papers Svenningsen et al. (2016) exploration of intense and realistic, but delusional memories reported by intensive care unit survivors, and Kings et al. (2017) investigation of men’s nebulous expectations regarding first-time fatherhood.

Role of the Researcher

Researcher role is addressed in various ways on both checklists I referenced, although I believe it warrants further discussion. The most frequent criticism I have heard over time about qualitative research in general is that of bias that results from the researcher’s direct interaction with participants and resultant data. I offer in response that there is no value-free human subjects research of which I am aware: Researcher preference (i.e., bias) informs every aspect from identifying the question or area of interest, to deciding whether to use software and which program(s) to use, to determining which, if any, data might be “transformed,” discarded, or estimated (in the case of missing survey items), all the way up to deciding whether to admit and how to describe any or all of the limitations that apply to the work.

In contrast with quantitative approaches, qualitative researchers are more often inclined or encouraged to acknowledge their biases. I agree with the recommendations, found in both checklists I cited, which suggest that authors should identify their relationship with both their subject matter and their sources of data, which usually include human participants and sometimes include their creations. Additionally, although many, maybe most, qualitative researchers align themselves with a belief system that acknowledges variations in perception resulting from individual or contextual differences, the degree to which this guides a given research project might vary. My suggestion is that authors should disclose their general worldview and other meaningful information that potentially influenced aspects of the specific study. I also suggest this disclosure be brief and relevant and not be used as an opportunity to show how well read the author is in the philosophical history that underlies a certain qualitative approach; additional detail is more likely to be appreciated on a research blog, in an online supplement, or by motivated attendees of a research seminar.

Variation within Qualitative Inquiry

Categorizing Qualitative Research

One of the greatest challenges associated with qualitative inquiry is that it has many sub-types. What distinguishes one of these from another might be study design or type of research question, although in other instances, the characteristics that make a research study an example of this rather than that have more to do with author decisions made during analysis or reporting. In the present paper, for the sake of consistency and familiarity, I use the term approach to refer to the sub-types within qualitative research; these sub-types are alternately referred to as methodologies, methods, or by Saldaña (2011), as genres—a term I believe emphasizes the interpretive nature of qualitative inquiry. It is not my aim to suggest one approach is superior to another or to provide an exhaustive list of qualitative research approaches. Rather my goal is to provide some idea of the range of labels used by authors to describe qualitative inquiry and to consider what Sex Roles’ readers are likely to expect when they see a study described as an example of a certain approach.

I have combined approaches into four numbered clusters, shown in the following, and I ordered them based on my interpretation of the least to the most flexible in terms of practices and processes. I provide additional detail on some frequently used approaches and a necessarily limited list of primary sources associated with many approaches; these are not the only sources but they are those with which I am most familiar. Although there are many notable survey texts, I emphasize approach-specific sources because these provide more detail and are often held in higher regard by qualitative researchers, instructors, and peer reviewers.

Less to more Flexible Approaches

Cluster 1

Grounded theory, phenomenology, and narrative inquiry are examples of well-developed qualitative approaches that are each associated with certain types of purposes or questions and a limited list of analysis schemes. When readers who are familiar with qualitative inquiry see these labels, they are likely to have some idea of how the research was carried out and will have expectations about the research report before they begin to read it. The subtle variations within these approaches largely, but not exclusively, relate to how data are processed or analyzed.

Grounded theory as an approach is characterized by a desire to develop a theory to address a concern or question. These questions often are articulated as a question of process (e.g., How do transgender individuals navigate challenges to regular participation in physical activity?; What is the process for women in male-dominated academic disciplines to become assimilated into a Ph.D. cohort?). Researchers usually gather data through individual interviews and use relatively large samples that represent the range of those best able to address aspects of the theory. The sampling strategy is driven by theory development so that as new findings are identified through analysis, researchers might decide that additional views are needed to further refine the theory.

Grounded theory is associated with conventions for analysis that include multiple stages of coding. The criterion “theoretical saturation” (Glaser & Strauss, 1967, p. 61), or the idea that no novel relevant information appears after analysis of multiple interviews, originated with grounded theory. Theoretical saturation is a result of analysis so that it is not a standard that can be used a priori to determine sample size. Although grounded theory research should be characterized by purpose and intent, and not applied after the fact as a framework for data collected absent those qualifications, a comprehensive grounded theory study might be proposed and carried out to incorporate multiple data types such as secondary analysis of data collected for another (related) study. Commonly used resources for grounded theory research include Glaser and Strauss (1967), Corbin and Strauss (2015), and Charmez (2014).

Phenomenology has subtypes including existential, transcendental, and interpretative phenomenological analysis (IPA). Phenomenology is characterized by questions or purposes that have to do with how people experience a thing or phenomena of interest (e.g., What is the lived experience of having sex reassignment surgery?). Researchers using phenomenological methods conduct in-depth analysis of individual interview data from a small number of participants. Research designs including fewer than ten participants are common, although multiple interviews per participant might be used. Analysis does not rely on coding but instead follows fairly complex processes of researcher-initiated reflexive deconstruction and reconstruction of each interview transcript; see Moustakas (1994) for detailed examples. Findings from some types of phenomenological methods are shared with participants to assess accuracy of depiction of the experience. Research reports might contain a thematic analysis and a composite description of participants’ lived experience of the phenomena of interest. Sources for general phenomenological qualitative inquiry include Moustakas (1994), Giorgi (2009), and Van Manen (2016). The seminal work on IPA is Smith et al. (2009).

Narrative inquiry, according to Clandinin (2013, p. 71) is “an approach to the study of human lives conceived as a way of honoring lived experience as a source of important knowledge and understanding.” Clandinin’s definition of narrative differs from the phenomenological approaches because it does not focus on one narrowly defined experience but rather considers multiple experiences, probably unified by some context, such as high school or the time spent in a certain job, but often extending over a period of time. Reissman (2008) provided an alternative approach to narrative methods, broadly describing narrative in terms of its relationship with storytelling while acknowledging that the scope of what is considered narrative varies by discipline and researcher. Both authors focus on the text, or created narrative, as a developed source of data that facilitates in-depth analysis. Texts might comprise or combine participant-provided data, such as interview transcripts and visual or artistic expressions or researcher-initiated data, such as field notes, journal entries, and reflexive responses to participant-initiated data. Reissman offered multiple alternatives for analysis and described steps for each, whereas Clandinin described and demonstrated stages in a broad process she called “unpacking” (p.81) that facilitates transformation of the original data through a series of stages into a final product that is suitable for dissemination, such as a report researchers might prepare for Sex Roles.

Cluster 2

Some other approaches are associated with specific types of questions and defined entities, as well as use multiple types of data, but they have more flexibility in analysis than those I included in Cluster 1. These include ethnography and its younger sibling autoethnography, case study, and descriptive qualitative research. The latter is largely distinguished from other approaches by researchers’ focus on descriptive rather than on interpretive results. As with Cluster 1 items, readers are likely to have certain expectations regarding research described by these labels. However, when compared to Cluster 1 items, there is potential for wider variation in units of analysis, data processing, and data analysis within the Cluster 2 approaches so that it is challenging to describe standard conventions for sampling, data type, data analysis, and presentation of a research report.

Ethnography and autoethnography traditionally have to do with examination of culture, but culture can be broadly (or loosely) defined and, as a result of technology, might refer to a virtual rather than physical or location-associated group. Several of the seminal ethnographic works provide examples, not directions, because these are presented as episodic stories rather than methods texts. These include Whyte’s (1967) description of Italian-American young men living in Boston, Stack’s (1974) woman-focused exploration of Black family structure in the U.S. Midwest, and immersion in drug and crime culture in Philadelphia described by Goffman (2014). Methods sources include Spradley (1980/2016; 1979/2016) and Fetterman (2010) for ethnography. Ellis (2013) wrote an autoethnography of teaching a graduate class as a resource for autoethnographic research, which can be distinguished from classic ethnographic research by being an insider’s exploration of one’s native culture.

Case study designs often have the purpose of identifying successful or less successful practices, although this is not a requirement. A particular case study research project might range in scope from a small group, such as the volunteer staff at a community service organization, or a larger system, such as a health system or school district, or might involve comparison of many cases. Well-known case study methodologists are Stack (1995) and Yin (2013).

Descriptive research might be considered the default qualitative method. Descriptive designs reflect a variety of questions of researcher interest and often represent an exploratory research direction. To my knowledge there is not an entire book that describes descriptive practices, although a clear and helpful article authored by Sandelowski (2000) provides a good introduction to descriptive research. Additionally a general qualitative data analysis text such as Saldaña (2016), Gibbs (2007), or Bazeley (2013) will often contain a good deal of information about managing descriptive data.

Data for any of these approaches might include transcripts from individual or group interviews; observation field notes; journals or reflective writing (especially ethnography or autoethnography); visual data such as drawings or photographs; the text of letters, government documents, or news reports; or anything else that is relevant and can be incorporated into analysis. Ethnographic and case study designs might integrate qualitative and quantitative data. (I discuss mixed methods designs in more detail in the following.) The default form of analysis for the approaches in Cluster 2 tends to be coding; increasingly this is coding done via qualitative data analysis software (QDAS) because of data management capabilities that include ability to manage multi-media data and monitor consistency of analysis practices in collaborative projects.

Cluster 3

Researchers might describe their research using these terms: (qualitative) content analysis, discourse analysis, arts-based or arts-informed research, and video/photo elicitation (e.g., photovoice). These approaches are inconsistently employed so that the typical reader is less likely to know what to expect from a research report. Thus, authors using these approaches should take particular care to describe in their report details related to design or framing elements, data source(s), units of analysis, and data processing and analysis practices. To further enhance credibility of research reports, authors should explain and justify why their choices were made, with particular focus on how these choices support the purpose of the research.

Content analysis might be considered qualitative or quantitative—or a mixed methods design. Please refer to Neuendorf’s (2011) Sex Roles paper for an excellent guide to quantitative approaches; the author also includes some discussion about use of qualitative approaches for “clearly latent constructs in messages” (p. 282). For authors pursuing qualitative content analysis (QCA), I recommend Schreier (2012), who provided guidance on appropriate use of QCA and differentiated between QCA and descriptive coding-centered data analysis. Although discourse analysis is a highly developed approach and could be as easily included in Cluster 1, I have included it here because it is flexible and, as such, it is often applied as a secondary analysis technique as opposed to an approach like grounded theory that tends to inform the project from start-to-finish. A broad guide to discourse analysis is presented in Gee (2014). Critical discourse analysis, an extension of discourse analysis, is associated with van Dijk, Wodak, and Fairclough, who each contributed chapters to a text edited by Wodak and Meyer (2015).

Highly regarded authors on arts-based or arts-informed research methods include Barone and Eisner (2011), Butler-Kisber (2010), and Leavy (2015). I have used Banks (2007) as a guide for visual data analysis; additionally Saldaña (2016) provided considerations for visual analysis and described additional resources in a section of his comprehensive coding methods book.

Cluster 4

Authors sometimes describe their research by identifying the method used to collect data, (e.g., interview research; focus groups) or by taking a label from an outcome of analysis (e.g., thematic analysis; conversation analysis) or by identifying an interpretive lens (e.g., critical research; feminist research) or a philosophical framework (e.g., Bourdiesian; Foucauldian). In a given research report, one of these descriptors might operate as the defining approach or it might be a sort of demi- or co-approach. For example, a lens or philosophical framework might only be an expression of authors’ position rather than a description that tells how a specific study was carried out, so it is used in combination with one or more of the more explicitly defined approaches reviewed previously, such as critical qualitative interview research or feminist ethnography, to tell readers about both motivation and design.

Descriptions based on method, analysis or interpretive/philosophical lenses, whether offered alone or in bundles, often offer readers only a vague notion of how the research was designed and carried out. This is not to suggest these are not legitimate ways to do qualitative research, but rather that authors should consider that the fewer conventions or authoritative resources readers automatically associate with an approach label, the more detail is necessary to allow readers the opportunity to assess the quality of conduct and credibility of results. For instance, when authors state their research reflects a critical approach (or design, or methodology), I recommend that they tell readers what they think this means, perhaps by offering an authoritative definition (authoritative meaning the source the authors consider to be an authority), and describe how it impacted the essential elements of the research (e.g., purpose, data analysis, etc.). Likewise, if data were analyzed to look for excerpts that suggest “habitus” (Bourdieu, 1977, p. 82), I suggest that it is helpful to tell readers about the factors underlying this decision and how it influenced data processing, rather than vaguely offering: “We used a Bourdieusian approach to this research.”

For these highly flexible approaches, authors have many alternatives for expert sources. I find that original works are essential sources when research is framed within a specific philosophical lens. For authors who are modifying an existing method or developing a novel one, readers probably deserve to know the reasoning for this innovation, such as the reasons that conventional approaches are insufficient.

In general, for work done within a less structured or a bundle of approaches, authors should offer authoritative citations for their decisions regarding the essential elements of research listed in the first section of my paper. For example, if authors chose to use coding methods taken from Glaser and Strauss (1967), readers deserve to know why this was selected in lieu of other available approaches to data analysis.

A Couple of Leftovers

There are two additional approaches of note that, in my view, do not neatly fit in any of the clusters I identified here. The first is mixed methods research, and the second includes the range of labels associated with qualitative meta-studies (e.g., meta-summary, meta-synthesis, qualitative meta-analysis, meta-ethnography.). There are many books on mixed methods (e.g., Hesse-Bieber, 2010; Morse & Niehaus, 2009; Teddlie and Tashakkori, 2009) but no single set of agreed-upon conventions regarding appropriate questions or purposes, acceptable data types, optimal sample sizes, preferred methods for data collection and analysis or how and when data streams are integrated, the best way to present findings, or much of anything else. Generally speaking, the expert source used to plan a mixed methods study is also the best guide to use when writing the research report.

Aggregate reports of previous research, or qualitative meta-studies, present unique challenges for authors. Although the process of a systematic search is essentially the same for any meta-study, there is a great deal of variety in how authors report their findings. This means that any attempt to pool results across studies is not nearly as straightforward a process as it is for a quantitative meta-analysis. For those who want to report (or plan) meta-studies, I recommend you refer to one of two sources. Sandelowski and Barroso (2007) categorized qualitative meta-studies as meta-summary or meta-synthesis and provided process guidelines for both. Paterson et al. (2001) delineated three levels of integration—meta-summary, meta-method, and meta-theory—in their comprehensive guide to qualitative meta-studies. Although Sandelowski and Barroso as well as Paterson et al. represent health professions, their provided guidelines are applicable to other social research topics including gender research. For authors who would like to take the further step of developing a confidence rating for findings from meta-studies, the GRADE-CERQual group (Lewin et al. 2015) provided a basic overview of the process and its aims. It is also possible to reach out directly to the working group (http://www.cerqual.org/) to request personalized support.

Mine is not an exhaustive list of sources; there are many other high quality sources available that might be in book or article form. As I noted previously, I have deliberately not included survey texts and, as a general rule, I do not recommend that these be the sole source for design and methods decisions because these texts tend to provide as much or more information about distinguishing among methods rather than guidance on working within a method.

Other Considerations

In this next section, I periodically cross the divide I described back in the first section that exists between making recommendations and providing directives. Below I list and describe some potential challenges authors might encounter when writing up qualitative research reports, grouped into five broader categories of coherence, data analysis, use of software, use of quantitative research criteria, and others.

Coherence within Approaches

The practice of combining elements from one approach with those of another is not universally accepted or rejected within qualitative inquiry, although there are challenges on two levels: methodological (how the research was carried out) and theoretical (foundational or framing aspects of the research). Regarding the former, authors who are not forthcoming about their decision process should expect to be questioned when mixing aspects of design, data processing, analysis, and reporting that are associated with unique methodologies. Sandelowski (2000) suggested that work within the descriptive approach is often characterized by borrowing techniques from the more structured approaches so provides a citable source for those working with descriptive designs. However, any blending of approaches benefits from explanation, rationale, and citation of supportive or example sources. Ideally, a blending of approaches reflects an informed decision on the part of the researcher to fit together methods in order to best explore a unique or challenging question or to work in a unique or challenging context. Otherwise, authors usually should be faithful to the processes associated with the approach that was chosen for a particular study and provide the expected outcome (e.g., grounded theory studies should result in a theory).

Lack of theoretical coherence can be (a) an error of omission, as when authors who assert to be informed by a specific theoretical approach seem to have forgotten about this immediately after it was described in the introductory section in the paper, or (b) an error of commission, as when authors of qualitative research reports present evidence to support inappropriate standards such as reproducibility or generalizability of findings. (My personal impression is that these evidences of credibility are often added after the fact at the request of co-authors or peer reviewers who are less confident about the inherent strengths of qualitative research.) As with methodological coherence, I suggest that compromises to or blending of aspects from multiple theoretical foundations are made in a thoughtful way and reported in a forthright way.

Performing and Describing Data Analysis

Coding

As I noted previously, coding is just one choice for data analysis. Although I do not recommend that authors should provide a textbook definition of coding, there are variations in what units are coded (words, lines, phrases), and how codes are derived that should be acknowledged. Authors should briefly describe the basic elements of their coding process and provide appropriate citations.

Thematic Analysis

This is often how authors describe a coding process in which many codes are clustered into fewer, summary codes. The phrase thematic analysis itself is not consistently used to any extent that readers know what authors did; in other words, it is not necessarily a sufficient description for either an approach or the data analysis process. In their widely cited work, Braun and Clarke (2006, p. 77) referred to thematic analysis as an “analytic method” and provided guidelines (and so offer a useful citation) for data analysis. To provide comprehensive information for readers, authors who describe their research as thematic analysis not only should acknowledge their source for analysis but also should describe design elements, including relevant theoretical foundations, and provide appropriate citations.

Use of Stock Phrases or Techniques

There are several stock phrases associated with qualitative inquiry that are at times viewed as universal measures of quality, although many are actually associated with a particular approach. Some examples of these include: theoretical saturation, line-by-line coding, the method of constant comparison (all originating from the grounded theory of Glaser and Strauss; see their 1967 work), member checking (associated with some types of phenomenology and other interview-driven projects; see Moustakas, 1994, for discussion of how participants, in responding to their own interviews, become active agents in co-constructing findings), interrater or intercoder reliability (usually associated with a pre-determined list of codes rather than open coding processes; reasons for or for not engaging in this practice are discussed in Saldaña, 2016). These descriptions might be quality elements, depending on the approach, or might be misplaced practices, again depending on the approach. Authors should consider the applicability of these quality criteria in association with a given research report, especially if it is not typically associated with the primary approach, and they should consult and provide appropriate citations.

One additional stock phrase, themes emerged, is not associated with any particular approach but is very commonly used; indeed, I have used it myself. However, it is imprecise and perhaps misleading, suggesting an uncontrolled or even a magical process. Use of this description in a research report also allows a researcher to avoid taking credit (or blame) for his or her analysis decisions. I suspect it reflects another effort to make the analysis sound more systematic and less biased. Instead, I recommend that authors acknowledge the active part they play in analysis: “I/We developed these themes by….”

Presenting Themes

When themes are developed and presented in the results section, whether via thematic analysis, cycles of coding, or as a convention within another qualitative approach (e.g., Interpretative Phenomenological Analysis), authors submitting to Sex Roles are advised to incorporate a themes table into their research report. This table should, at minimum, list each named theme and its description. It is desirable to also include a prototypical quote (although authors should not repeat the same quotes in the text of the report). Inclusion of participant quotes as data to support a theme also provides evidence of validity for readers. Additionally, authors are encouraged to associate code numbers or pseudonyms with themes, or provide a count of participants who contributed to each theme. Authors might also include interrater or intercoder reliability, when applicable, such as in content analysis designs. This table should first be mentioned early in the Results section and its organization should parallel that of the Results section; thus the table will serve readers both as a roadmap to the results and as a summary of them. Please refer to Berkovitch and Manor (2018, Table 2) and Morgan and Davis-Delano (2016, Table 1) for examples of themes tables in papers published in Sex Roles.

Use of Qualitative Data Analysis Software (QDAS)

QDAS in General

Qualitative data analysis software (QDAS) programs continue to grow in popularity. In essence, QDASs facilitate creation of a database from imported, entered, or selected text or multi-media excerpts. Most QDASs were developed to perform primarily coding-based analyses and require thought and perhaps workarounds to do anything different. QDASs are clearly helpful for some types of studies, such as meta-studies, and are beneficial in some situations, such as when there is a large, geographically spread-out team that is responsible for managing a great deal of data.

The Role of QDASs in Research Conduct and Design

The availability of a QDAS is not on its own justification for selecting a purpose or approach, or for combining elements of different approaches. QDAS is a tool the researcher might opt to use to help with analysis, not the driver of design decisions.

The Role of QDASs in Research Reports

“We used program X to conduct data analysis” is not an ample description. Most of the same information is needed for QDAS-facilitated analysis as is needed for any qualitative data analysis, that is, what units are coded, what type of coding is used, how codes were aggregated or clustered, etc. QDAS guides and instructional materials are often extremely helpful and frequently were authored by qualitative researchers, although I would not recommend a QDAS manual as a credible enough source on its own to support research design, selection of approach, and analysis decisions.

Quantitative Practices Applied to Qualitative Research

Using Descriptive Statistics to Report Qualitative Findings

Authors should avoid this practice because it creates a false impression in readers that the analysis is entirely frequency-based. Although qualitative analysis involves a search for patterns, and patterns consist of repeated information, the patterns tend to reflect researcher-initiated abstraction or are derived from data reduction, aggregation, interpretation, and other types of transformation. This is not the same as a finding after compiling survey results (e.g., “X% of people selected option c”). To associate weight with a finding, the alternatives are to rely on conventional but imprecise descriptions (e.g., most, many, a few, or some) or to provide the number of participants who support the finding. For Sex Roles, it is preferred that authors describe the number of participants who support a finding; this is especially important when authors argue for the relevance of a category, trend, or theme that is represented by few participants. In a couple of instances, source or participant counts are conventions: for some measures associated with qualitative meta-studies, it is necessary to specify the precise number of times a finding is represented across a sample or the number of findings a given source contributed; additionally, theme validation criteria recommended by Smith (2011) in association with interpretative phenomenological analysis is based on the proportion of total participants who reflect a theme.

Subgroup Analysis

This is an error that is probably exacerbated by some reviewers’ fascination with demographic information along with some authors’ enthusiasm for creating tables to show a row per participant and sometimes imaginative or exotic pseudonyms. Qualitative samples are rarely large enough or balanced enough to consider subgroup analysis and purposively recruited participants are unlikely to demonstrate the characteristics of a random sample that are required for statistical subgroup analyses. Additionally, purposes of qualitative inquiry tend to center around understanding detail and nuance, as opposed to identifying causal attributions or test associative relationships. In my view, subgroup analysis, like use of descriptive statistics to describe findings, is an example of incongruity between research design and outcomes of interest. I recommend authors who hope to explore potential associative relationships craft interview guides with that goal in mind, and, after analysis of qualitative data has contributed helpful information, plan follow up survey research studies to more appropriately test the hypothesized associations.

Authors who have previously submitted or published qualitative research articles might have been asked by reviewers to provide a great deal of detail about participants; in some instances I suspect this is because they are trying in their own minds to identify associative relationships. In general for Sex Roles, authors who want to provide relevant demographic information (e.g., gender, race/ethnicity, occupation) about participants quoted in their text should consider including a participants table rather than cluttering their text with this information for every quote. This table should be organized by pseudonym (sorted alphabetically) or other ID in the leftmost column and include an appropriate level of information to distinguish, humanize, and situate participants without risking revealing actual identities. First referring to this table early in the Results section ensures placement of the table near the beginning of the results where it will be most helpful to readers who might want to consult it to learn more about an individual who is quoted. It is important then to identity, by ID or pseudonym, each speaker quoted in the ensuing text.

Note that this participants table does not substitute for describing the aggregate sample in the Participants subsection of the Method. Additionally in both table and text, authors should only provide the essential details readers need to interpret the research, and they should always err on the side of preserving participants’ confidentiality. For gender research, essential details might encompass more or different elements than would be necessary to explore other research questions. Often “a participant who identifies as transgender said” is as meaningful in a given study as “Jean, a 37-year old participant who identifies as transgender and has brown eyes and a bachelor’s degree in biology said….” I refer interested readers to Morse (2008, p. 299) for further discussion about what she termed “irrelevant demographics.” Examples of participant tables in papers published in Sex Roles can be found in Berkovitch and Manor (2018, Table 1) and Miller et al. (2018, Table 1).

Theory Testing by Coding Data into Theoretical Constructs

This is potentially a misuse of a qualitative design that results in essentially underpowered survey research conducted using an untested instrument. For researchers who want to assess participants’ scores on a construct, attribute, or trait, a better approach is to seek out already created structured interview items (surveys) with established psychometric properties, and perform power analysis calculations to determine sample size before recruiting a random (or random-ish) sample. If there is not a standardized instrument, I offer that qualitative inquiry is likely better suited to help authors develop questions than to assess scores on a construct.

Recruiting Participants Based on Quantitative Standards

Even a large qualitative sample is likely still too small to facilitate presentation of credible descriptive statistics, subgroup analysis, or theory testing, as I noted previously. Striving for stratified samples is often another way of trying to get at subgroup analysis. Depending on the approach and associated analysis, authors might end up drowning in data if the sample size is too large. Additionally, there are references that support smaller (e.g., 10–12) sample sizes for many interview designs (e.g., Guest et al. 2006). One approach that is exceptional in this regard is grounded theory because it is often desirable to stratify the sample based on where individuals are positioned within the process related to the research question.

Others

Piecemeal Publication

This has particular significance in qualitative inquiry because an interview (or group interview) can in itself be considered a type of intervention during which participants (and researchers) are prone to have emotional experiences, develop arguments or rationales, change their way of thinking, acquire new knowledge, and in these or other ways exit an interview with different thoughts than those they carried in. Additionally, anyone who has participated in interview research, in any role, would probably agree that participants’ responses to items are not necessarily orderly or in order. Trying to split responses to write multiple articles is not good research practice, and “piecemeal publication” represents an ethical violation described by the American Psychological Association (2010, pp. 13–15). There are reasonable exceptions, such as development of a paper exploring an emergent element in greater detail. For example, I agree with Holland’s (2009) decision to perform an additional analysis and write up a separate research report about three participants within a larger study of 15 in which she explored the unique barriers to the use of pole dancing for exercise described by these three when compared to the remainder of the sample. Authors who believe they have a compelling reason to present a portion of their data as an additional article, as Holland did, should do their best to convince editors, reviewers, and readers of the legitimacy of this practice when they initially submit their paper. Additionally, authors are required to be honest both during the submission process and in the text of the paper about alternate presentations of data from the same study.

Asserting Lack of Generalizability before Presenting Findings as Universal Truths

I think as qualitative researchers we believe our findings might be extended beyond one specific person in one certain context; otherwise, we would not undertake any research. As an alternative to using the familiar phrase, “qualitative results are never generalizable,” or employing the other extreme of projecting the findings from a single small study onto the universe of similar settings, authors might seek the reasonable middle-ground appropriate to their research. For a single case, authors might only be confident enough to suggest that their analysis revealed potentially testable circumstances. For a larger, perhaps geographically spread sample, authors might disclose why they conclude their findings indicate a trend or a common process. This information, presented in the discussion section of the paper, represents a good time for authors to use evocative, suggestive, or conditional rather than confident and authoritative language. See Morse (1999) for additional discussion including more confident applications of generalizability than I have offered here.

In Closing

As practitioners of qualitative inquiry, we are preoccupied with human and social issues, and we thrive and grow when we conduct our scholarly activities in social ways and incorporate dialogue, discussion, (polite and informed) critique, and, in general, work together to improve the practice and reporting of qualitative inquiry. I encourage researchers who are not actively conducting qualitative inquiry to consider the value of these methods to encourage comprehensive understanding of their research interests. For those who are already qualitative practitioners, I hope the information in my paper inspires or reinforces a thoughtful, reflexive approach to reporting your research.