PaperlayerView exact revision on arXiv

Source-linked AI summary

Socio-Technical Grounded Theory for Software Engineering

Rashina Hoda

arXiv:2103.14235v3cs.SE

TL;DR

SE researchers face challenges understanding and applying traditional Grounded Theory guidelines in socio-technical research. The paper presents STGT, an iterative, incremental method with contextualized guidelines for conducting and evaluating such research.

  • Problem

    Software engineering researchers face challenges understanding and applying traditional Grounded Theory guidelines, while socio-technical research calls for updated guidance.

    Source: §4 · p. 7

  • Method

    STGT is an iterative, incremental method for socio-technical research, using traditional and modern techniques with interleaved data collection and analysis.

    Source: §4 · p. 7

  • Results

    The paper presents a definition of socio-technical research context and contextualized guidelines for conducting and evaluating STGT, including nuanced evaluation guidance for different outcomes.

    Source: §1 · p. 2

  • Takeaways & Limitations

    STGT is intended to ease Grounded Theory application and improve quality outcomes in software engineering.

    Source: §1 · p. 2

    Without the remaining STGT steps and advanced data analysis and theory development procedures, outcomes should not be claimed as mature theories.

    Source: §7.2 · pp. 20–21

Abstract

from arXiv · show

Grounded Theory (GT), a sociological research method designed to study social phenomena, is increasingly being used to investigate the human and social aspects of software engineering (SE). However, being written by and for sociologists, GT is often challenging for a majority of SE researchers to understand and apply. Additionally, SE researchers attempting ad hoc adaptations of traditional GT guidelines for modern socio-technical (ST) contexts often struggle in the absence of clear and relevant guidelines to do so, resulting in poor quality studies. To overcome these research community challenges and leverage modern research opportunities, this paper presents Socio-Technical Grounded Theory (STGT) designed to ease application and achieve quality outcomes. It defines what exactly is meant by an ST research context and presents the STGT guidelines that expand GT's philosophical foundations, provide increased clarity and flexibility in its methodological steps and procedures, define possible scope and contexts of application, encourage frequent reporting of a variety of interim, preliminary, and mature outcomes, and introduce nuanced evaluation guidelines for different outcomes. It is hoped that the SE research community and related ST disciplines such as computer science, data science, artificial intelligence, information systems, human computer/robot/AI interaction, human-centered emerging technologies (and increasingly other disciplines being transformed by rapid digitalisation and AI-based augmentation), will benefit from applying STGT to conduct quality research studies and systematically produce rich findings and mature theories with confidence.

1 INTRODUCTION

SE researchers have applied Grounded Theory (GT) to human and social software-engineering phenomena, but traditional guidance is difficult to access and poorly suited to interwoven socio-technical contexts. This paper presents Socio-Technical Grounded Theory (STGT) as contextualized guidance intended to improve accessibility and research quality.

1 INTRODUCTION

  • GT is a systematic research method for theory development, especially suited to understudied phenomena and complex how-and-why questions.Its three versions are Classic, Strauss-Corbinian, and Constructivist; its practices include iterative data collection and analysis, open coding, and constant comparison.Source: §1 · p. 1
  • A 20-year review found major quality concerns in SE GT studies, and only 15% provided detailed accounts of their procedures.Concerns included failing to identify the GT version, combining guidelines without rationale, and applying practices à la carte; fewer studies made theoretical contributions.Source: §1 · pp. 1–2
  • Because SE phenomena interweave social and technical aspects, studying either alone can leave the research incomplete; GT therefore needs adaptation for these contexts.Examples include requirements elicitation, pair programming, multi-team coordination, and DevOps.Source: §1 · p. 2
  • Modern social-media and open-source data, repository mining, natural language processing, and AI tools create opportunities that traditional GT guidelines do not address.The paper argues that contemporary guidance is needed to help researchers adapt these tools systematically.Source: §1 · p. 2
  • STGT defines the socio-technical research context and provides contextualized guidance, required skills, and expanded philosophical foundations for conducting and evaluating research.The guidelines aim to ease access and improve quality outcomes in SE.Source: §1 · p. 2
  • STGT permits full-method or data-analysis-only use, later selection of emergent or structured theory development, frequent reporting from interim to mature outcomes, and evaluation tailored to application and outcome.It also offers guidance on harnessing modern data types, sources, collection tools, and analysis techniques.Source: §1 · p. 2

2 GROUNDED THEORY – SOCIAL TRADITIONS

Grounded Theory (GT) is a rigorous method for developing evidence-based theory from social phenomena, and it has evolved into distinct methodological traditions. Its procedures combine iterative data collection and analysis with theory development grounded in real-world evidence.

2.1 The Origins

  • GT challenged the dominance of quantitative research and skepticism toward qualitative rigor, promoting theory development; Glaser and Strauss formulated it through studies of dying patients and nursing care.Their procedures were documented in The Discovery of Grounded Theory – strategies for qualitative research (Glaser and Strauss; 1967).Source: §2.1 · pp. 2–3

2.2 The Method

  • GT is a complete, rigorous method for developing theory from evidence collected from real-world practice and investigating social phenomena through a social lens.Its work covers data collection, analysis, and advanced theory development.Source: §2.2 · p. 3
  • GT uses iterative data collection and analysis, coding, theoretical sensitivity, memoing, theoretical sampling and sorting, and theoretical saturation to develop concepts, categories, and theory.Researchers use emerging theoretical gaps to guide later data collection; sorting memos can help structures emerge, and coding with theory templates is optional. Saturation is reached when further data collection yields diminishing returns.Source: §2.2 · p. 3
  • GT produces mid-range theories limited to the studied contexts but open to modification as further evidence is gathered.Source: §2.2 · p. 3

2.3 The Evolution

  • Strauss and Corbin’s GT introduced structured axial coding and analytical templates, while Glaser argued this approach forced theory rather than allowing it to emerge (Glaser; 1992).The structured approach can produce earlier theory structure and later deductive verification; some researchers prefer it for this structure (Masood et al.; 2020a,b).Source: §2.3 · p. 3
  • Charmaz’s Constructivist GT treats reality as subjectively interpreted and constructed by the researcher, unlike the objectivist stance; she characterized Glaserian GT as primarily positivist and Strauss-Corbinian GT as positivist with pragmatic influences.Constructivist GT subsequently became a prominent version.Source: §2.3 · pp. 3–4

3 GT IN SOFTWARE ENGINEERING

GT can support theory development in software engineering, but traditional sociological guidance has proven difficult to apply consistently and well. The section describes its promising contexts, growing use, quality challenges, and the need for socio-technical guidance.

3.1 The Promise

  • GT can build theoretical foundations for SE while combining research rigour with practical relevance and informing recommendations, guidelines, and tools.The paper illustrates this potential with AI ethics research that could explain developers’ perceptions, barriers, and enabling strategies.Source: §3.1 · p. 4
  • GT best fits human and social, practice-based topics needing theory development and deep how-or-why inquiry.It is especially suited to areas with research gaps or theories that do not resonate with practice.Source: §3.1 · p. 4

3.2 The Progress

  • GT gained steady acceptance in SE, particularly over the last decade, propelled by theory-development efforts and growing attention to human and social aspects and agile development.The section notes that GT has been used to study varied human and social aspects, and points to exemplar studies and SE guidance on traditional GT.Source: §3.2 · pp. 4–5

3.3 The Challenges

  • Among 98 reviewed journal papers, half claimed GT, only a third detailed its application, and five were comprehensive exemplars (Stol et al.; 2016).The exemplars were Coleman and O’Connor (2007), Hoda et al. (2011), Adolph et al. (2012), Hoda et al. (2012b), and Jantunen and Gause (2014).Source: §3.3 · p. 5
  • Researchers face barriers understanding sociological guidance, choosing among GT versions, distinguishing full GT from coding techniques, and providing credible evidence without breaching ethics.Raw interview and observation data often cannot be shared because of confidentiality and anonymity requirements, while withholding evidence can raise credibility concerns.Source: §3.3 · pp. 5–6
  • Other challenges include rushed or shallow claimed applications, poor presentation, unguided adaptations, and reviewers who are overly trusting or harsh.The authors caution that quality issues do not always reflect intentional misuse; genuine attempts can also fail amid researcher apprehension, misunderstanding, and poor presentation.Source: §3.3 · pp. 5–6

3.4 Misalignment and Need for Evolution

  • Traditional sociological GT is misaligned with SE’s socio-technical context, motivating flexible, modern guidance; earlier method translations show contextualization can help, and STGT first defines context then method.The proposed guidance is intended to address modern data types, sources, collection, and analysis techniques.Source: §3.4 · p. 6

4 SOCIO-TECHNICAL GROUNDED THEORY

STGT is an iterative, incremental method for socio-technical research that combines traditional and modern techniques to develop theories. Its guidelines define the research context, researcher capabilities, data and methods, and philosophical foundations needed for its application.

4 SOCIO-TECHNICAL GROUNDED THEORY

  • STGT interleaves data collection and analysis, developing theories through emergent or structured approaches and primarily inductive, but also deductive and abductive, reasoning.Its cycles progressively advance the research toward theoretical outcomes, and its guidelines also cover philosophical foundations, application, reporting, and evaluation.Source: §4 · p. 7

4.1 Socio-Technical Research

  • The socio-technical research framework defines context through four dimensions: phenomenon, domain and actors, researcher, and data, tools, and techniques.Together, these dimensions describe a research landscape distinct from the contexts for which traditional GT was originally designed.Source: §4.1 · p. 7

Socio-Technical Phenomenon

  • Socio-technical phenomena couple human and technological interactions, so studying either alone can leave an incomplete understanding; programming and other SE practices exemplify this coupling.Potential topics also include human aspects and values, social dynamics, technology’s role in varied settings, and socio-technical aspects of SE practices.Source: Socio-Technical Phenomenon · p. 7

Socio-Technical Domain and Actors

  • SE, computer science, HCI, AI, and information systems are inherently socio-technical domains, and relevant actors may include technology creators, users, human participants, virtual avatars, or AI agents.Software engineers are both producers and users of technology, interacting across physical and digital work settings.Source: Socio-Technical Domain and Actors · pp. 7–8

Socio-Technical Researcher

  • STGT researchers or interdisciplinary teams should bring together social-science, philosophical, qualitative, and theory-development capabilities with technical and domain knowledge.The repertoire includes philosophical foundations, qualitative research, theoretical and socio-technical sensitivity, and domain knowledge; the researchers need not be one person or experts in every area.Source: Socio-Technical Researcher · p. 8

Socio-Technical Data Tools and Techniques

  • STGT can combine traditional interviews, observations, and field trips with modern data, tools, and techniques, including software repositories, data mining, natural language processing, and sentiment analysis.Large public datasets can help scale theories to wider contexts when researchers derive patterns through systematic, rigorous STGT procedures rather than letting data access drive trivial questions.Source: Socio-Technical Data Tools and… · p. 9

4.2 STGT – Philosophical Foundations

  • STGT presents accessible foundations in reasoning, ontology, epistemology, and research paradigms, and compares its philosophical foundations with traditional GT approaches.The paper treats understanding these concepts as critical to conducting and evaluating quality research, especially STGT studies.Source: §4.2 · p. 9

Types of Reasoning

  • Grounded theory generates theories inductively, but theory development is not purely inductive: inductive, deductive, and abductive reasoning all apply.Abduction helps researchers form explanatory conjectures for surprising findings and return to data to examine or check them.Source: Types of Reasoning · p. 9

ONTOLOGY, EPISTEMOLOGY, & RESEARCH PARADIGMS

STGT accommodates multiple socio-technical realities and research worldviews, asking researchers to choose and declare philosophical stances suited to the study context. Its theory development is predominantly inductive, with deduction and abduction used in specified procedures.

ONTOLOGY, EPISTEMOLOGY, & RESEARCH PARADIGMS

  • STGT is predominantly inductive through coding and constant comparison, while deduction informs theoretical sampling and abduction informs memoing, theoretical structuring, and integration.Abductive hypotheses are iteratively examined through data collection, analysis, and memoing, with socio-technical context needed to develop robust hypotheses.Source: ONTOLOGY, EPISTEMOLOGY, & RESEARCH… · p. 10

Ontology

  • STGT enables research across physical, virtual, combined, and simulated realities that characterize socio-technical software engineering contexts.These realities may intermix contexts, symbols, languages, norms, and guidelines from physical and virtual worlds.Source: ONTOLOGY, EPISTEMOLOGY, & RESEARCH… · p. 10, Ontology · p. 11

Epistemology

  • STGT epistemological stance depends on both the reality studied and researcher preferences, allowing objective metrics, subjective interpretation, or a mix of both.For example, game analytics can support an objectivist study of virtual worlds, while combined development-team contexts can mix metrics with subjectively interpreted observations.Source: Epistemology · p. 11

Research Paradigm

  • Research paradigms combine ontological and epistemological stances, and declaring them helps align study conduct, presentation, and evaluation criteria.A positivist evaluation may wrongly expect generalisability or reproducibility from constructivist work, for which novelty, credibility, and rigour may be more appropriate criteria.Source: Research Paradigm · pp. 11, 13

STGT Research Paradigms

  • STGT is open to different research worldviews, but requires careful paradigm selection suited to the study context rather than paradigm agnosticism.Researchers should make their stance clear so findings can be understood and evaluated in commensurate ways.Source: STGT Research Paradigms · p. 13
  • STGT stance choices can vary by study focus, and researchers may combine epistemologies with theories or perspectives to pursue specific aims.For example, test-case selection can be studied objectively through tool performance or constructivistically through testers’ motivations and behavior; a constructivist-feminist lens can examine women contributors’ experiences.Source: STGT Research Paradigms · p. 13

5 STGT METHOD – OVERVIEW AND BASIC STAGE · 5.1 STGT – An Overview

STGT combines procedures from three traditional grounded theory variants in a basic data collection and analysis stage followed by flexible advanced theory development. It allows researchers to choose a research paradigm suited to the study context and distinguishes partial from mature findings.

5 STGT METHOD – OVERVIEW AND BASIC STAGE

  • STGT adapts procedures from Glaserian, Strauss-Corbinian, and Constructivist grounded theory into a method with basic and advanced stages.The overview introduces the method and points to further details in a forthcoming book, with examples of specific procedures referenced throughout.Source: §5.1 · p. 13, §5 · p. 13

5.1 STGT – An Overview

  • Researchers can choose the ontology, epistemology, and research paradigm best suited to the study context, including a constructivist paradigm.Source: §5.1 · p. 13
  • STGT balances being informed by literature with avoiding undue influence, progressively focuses data collection and analysis on emerging categories, and distinguishes method evaluation from outcome evaluation.It also recognizes partial findings and evaluates them using criteria distinct from those for mature findings.Source: §5.1 · p. 13
  • The basic data collection and analysis stage combines a lean literature review, study preparation and piloting, and iterative basic data collection and analysis.Initial concepts and categories emerge during piloting; a few strong categories and preliminary hypotheses mark the stage’s end.Source: §5.1 · p. 14
  • In emergent theory development, iterative targeted data collection and analysis continues to theoretical saturation, producing an integrated mature theory that can be further structured.Source: §5.1 · p. 14
  • In structured theory development, structured data collection and analysis continues to theoretical saturation, producing a structured mature theory that can be further integrated.Source: §5.1 · p. 14
  • Emerging categories and hypotheses can be reported as interim findings, while mature theory can be reported as final findings.Source: §5.1 · p. 14

5.2 Literature Review

STGT uses an early lean review to identify gaps and motivate a study, followed by targeted reviews that compare emerging findings with existing work. Literature reviews can also build sensitivity and domain knowledge, advance theory, or—under specified conditions—include systematic or grounded-theory reviews.

5.2 Literature Review

  • An early, high-level lean literature review identifies gaps and motivates a study, with existing theories guiding whether to pursue original theorizing, change the topic, or involve experienced theorists.If robust theories already exist, changing or refining the topic is suggested; if the topic is extensively studied but lacks resonant theories, a GT study should be attempted by experienced theorists.Source: §5.2 · p. 14
  • Periodic, in-depth targeted literature reviews compare emerging categories and hypotheses with existing work and situate original findings in the broader research landscape.They are conducted during the advanced stage and help fill research gaps.Source: §5.2 · p. 14
  • Learning about theories and theory development beyond the study domain can improve theoretical sensitivity, particularly for SE researchers without sociology or theorizing training.Suggested fields include sociology, psychology, and nursing.Source: §5.2 · p. 15
  • Practitioner and research literature in the study domain build domain knowledge and socio-technical sensitivity, supporting effective data collection, analysis, and reporting.This sensitivity helps researchers, including newcomers to a domain, understand its terminology, concepts, and jargon.Source: §5.2 · p. 15
  • Experienced researchers can use relevant literature knowledge to refine or advance theory, while guarding against its influence on original theorizing; becoming agile illustrates this practice (Hoda and Noble; 2017).This advanced practice is suited to researchers with study-domain and theorizing experience.Source: §5.2 · p. 15
  • Systematic reviews and mapping studies are not recommended for GT novices; experienced researchers using them with STGT should explain their paradigmatic interplay and relationship to original findings (Kitchenham; 2004).Researchers should explain how review findings inform theory development or how bias from those findings is managed.Source: §5.2 · p. 15
  • Grounded Theory Literature Review is a possible approach for future STGT studies in SE and may be better suited to accompany STGT than a systematic literature review (Wolfswinkel et al. (2013)).Source: §5.2 · p. 15

5.3 Study Preparation and Piloting

STGT study preparation involves selecting a topic and team, preparing and ethically assessing protocols, and piloting data collection and analysis before refining the study. Pilot data may be retained unless the pilot context differs from the study or the protocols require a major redesign.

5.3 Study Preparation and Piloting

  • Preparation includes choosing a topic and research team, with an experienced theorist recommended but not required, and developing study protocols.Protocols may cover recruitment, interviews, surveys or questionnaires, observation, or modern data mining techniques; ethics approval should be obtained when applicable.Source: §5.3 · p. 15
  • Piloting helps researchers refine protocols and become familiar with data collection and analysis, with one or two protocol-testing pilots and two or three familiarization pilots suggested for new researchers.Experienced researchers can also benefit because each topic brings new contexts and challenges.Source: §5.3 · p. 15
  • Pilot data can be used in the study unless the pilot context differs from the target context or the protocols require complete redesign, potentially including a topic change.Examples include piloting a survey with research students when the study concerns industry practitioners, or rarely needing to change the research topic.Source: §5.3 · p. 15

5.4 Basic Data Collection and Analysis · (Theoretical) Sampling · Basic Data Collection

STGT frames data collection and analysis as iterative, using emerging findings to guide theoretical sampling and combining traditional and modern data sources with attention to credibility and ethics.

5.4 Basic Data Collection and Analysis

  • Basic data collection and analysis iterates through sampling, data collection, analysis, and memoing.Source: §5.4 · p. 15

(Theoretical) Sampling

  • Researchers can begin with context-appropriate sampling, then use ongoing, theory-driven sampling to target data that develops emerging concepts and fills theoretical gaps.Theoretical sampling is not predetermined; its dynamic process interleaves data collection and analysis and may also use snowballing.Source: (Theoretical) Sampling · pp. 15–16
  • Theoretical sampling moves from high-level findings to specific data targets and can refine research protocols as concepts emerge.The paper describes this as deductive reasoning and gives refining interview questions or repository-mining keywords as examples.Source: (Theoretical) Sampling · p. 16
  • Researchers should use the socio-technical context to select relevant sources, such as participants or documentation, and suitable collection techniques.Techniques can include interviews, observations, and automated or manual mining of repositories and online forums.Source: (Theoretical) Sampling · p. 16

Basic Data Collection

  • STGT accepts traditional sources alongside modern public and domain-specific sources and immersive collection approaches.Examples include interviews, practitioner blogs, YouTube talks, Twitter, Trello, GitHub, Slack, and virtual game worlds.Source: Basic Data Collection · p. 16
  • Interviews and observations provide rich, contextualized data but are labor-intensive, while mining public data can scale collection but requires selection, filtering, and cleaning.Source: Basic Data Collection · p. 16
  • Combining traditional and modern approaches can use interview-derived concepts as seed terms for large-scale collection, supporting relevant, robust theory across wider contexts.The paper recommends this combination where applicable: traditional collection helps identify and filter public datasets, while modern sources and techniques help scale theoretical outputs.Source: Basic Data Collection · p. 16
  • STGT qualifies “all is data” by requiring credible modern data and care about ethical considerations in collection and use.Source: Basic Data Collection · p. 16

Basic Data Analysis

STGT data analysis combines open coding and constant comparison while examining technical and social aspects together to capture the full socio-technical meaning of the data.

Basic Data Analysis

  • Socio-technical coding considers technical and social aspects together, rather than ignoring technical elements or treating them as decorative black boxes.Coding condenses textual raw data into formats that capture its essence and meaning.Source: Basic Data Analysis · p. 16
  • Open coding stays receptive to emerging codes, while constant comparison identifies patterns across sources and raises codes through increasingly abstract conceptual and category levels.Open coding can generate large amounts of code; constant comparison makes the inductive progression from specific instances toward general patterns.Source: Basic Data Analysis · p. 16
  • STGT analyzes complete socio-technical contexts, turning activities such as requirements brainstorming and architecture discussion into richer concepts relevant to practitioners.Unlike a traditional sociological analysis focused on social concepts, STGT can distinguish debates about low-fidelity prototypes from debates about design patterns, including different developer responses.Source: Basic Data Analysis · pp. 16–17
  • Quality software engineering GT studies have often used socio-technical analysis implicitly, producing theories of reconciling agile and software architecture (Waterman et al.; 2015), becoming agile (Hoda and Noble; 2017), and scrum practice variation (Masood et al.; 2020b).The passage also points to further examples in Hoda and Noble; 2017; Hoda et al.; 2011, 2012b; Masood et al.; 2020a,b; Shastri et al.; 2021a,b.Source: Basic Data Analysis · p. 17
  • Explicitly acknowledging data’s socio-technical nature can support quality STGT analysis, while NLP, sentiment analysis, and AI-based techniques may augment and scale manual analysis.The paper presents these techniques as promising future means of easing analysis and enabling improvements in qualitative data analysis and theory development.Source: Basic Data Analysis · p. 17

Basic Memoing

Basic memoing systematically records researchers’ reflections and conjectures to support theory development. Through constant comparison and memoing, initial codes are condensed into concepts and categories until emerging findings can be presented for peer review and publication.

Basic Memoing

  • Memoing systematically documents researcher reflection and supports theory development, distinguishing STGT from other qualitative research methods.Memos capture thoughts, ideas, reflections, and evidence-based conjectures about emerging concepts, categories, and possible links; they may be written notes, sketches, annotated images, or transcribed recordings.Source: Basic Memoing · p. 17
  • Constant comparison and memoing progressively condense initial open codes into concepts and categories, narrowing the focus of analysis.As key concepts, categories, and preliminary hypotheses emerge, analysis is ready to move into advanced theory development.Source: Basic Memoing · p. 17
  • When key categories are sufficiently strong and detailed, they can be presented for peer review, workshops, and conference or journal publication.Source: Basic Memoing · p. 17

6 STGT METHOD – ADVANCED THEORY DEVEL-

STGT postpones the choice of theory-development approach until researchers have experience with basic analysis, then offers emergent and structured modes. Both modes use theoretical sampling, advanced memoing, and theoretical saturation to develop mature theories.

OPMENT

  • STGT offers emergent and structured theory-development modes, delaying the choice until researchers can understand their findings and select an appropriate approach.Both modes also include targeted literature review, theoretical sampling, advanced memoing, and theoretical saturation.Source: OPMENT · p. 17

Advanced Memoing

  • Advanced memoing iteratively reassesses and compares memos to confirm, reject, or establish relationships between categories through abductive reasoning grounded in the full socio-technical context.This replaces one-time theoretical sorting; evidence can strengthen proposed explanations as it is collected.Source: Advanced Memoing · p. 17

Theoretical Saturation

  • Theoretical saturation occurs when further data collection yields no new or significant additions, leaving strongly supported, dense categories and hypotheses.The last few attempts, such as the final 3–4 interviews, may serve to verify the theory.Source: Theoretical Saturation · p. 18

6.1 Emergent Mode of Theory Development (Option 1)

  • The emergent mode iterates targeted data collection and analysis until saturation, allowing theoretical structure to develop progressively before optional structuring.It suits broad topics with emerging strong categories but no clear structure; sampling targets theoretical gaps, while targeted coding focuses on significant concepts and categories but can still yield new ones.Source: §6.1 · p. 18, Context of Use · p. 18, Targeted Data Collection · p. 18

Theoretical Structuring

  • Theoretical structuring is optional but highly recommended: researchers identify a fitting theory genre and may map the emergent structure to a suitable existing template.Examples include presenting a process theory without a predefined template or adopting the six C’s template when it fits.Source: Theoretical Structuring · p. 18

6.2 Structured Mode of Theory Development (Option 2)

  • The structured mode iterates data collection and analysis under an existing theoretical structure, ending in saturation and theoretical integration.It suits a narrow, well-defined topic with a relatively clear structure after the basic stage; researchers may adopt a fitting template early or use their own emerging structure to guide further work. Collection fits, strengthens, and saturates existing categories, while structured coding identifies and strengthens relationships.Source: §6.2 · pp. 18–19, Structured Data Collection · p. 19, Structured Data Analysis · p. 19

Theoretical integration

  • Theoretical integration completes a mature, structured theory by connecting its categories to an overall storyline that explains the phenomenon.Integration can also add a theoretical layer, as in a classification of Scrum variations into standard, necessary, contextual, and clear-deviation types.Source: Theoretical integration · p. 19

7 STGT - APPLICATION, OUTCOMES, REPORTING

STGT can be applied as a full method to develop mature theories or in a limited data-analysis capacity to produce descriptive findings and preliminary hypotheses. Researchers should report the application type and claim only outcomes appropriate to it.

7.1 Full STGT Method

  • Full STGT extends beyond basic data collection and analysis through theory development, producing mature theories; it can be standalone or combined with other methods or R&D projects.Combined studies may use STGT to establish theoretical foundations and explain a problem before other methods validate findings or develop solutions.Source: §7.1 · p. 19

Standalone STGT Studies

  • Standalone STGT studies can be small and yield one mature theory or large and yield multiple related theories, with scale shaping duration, team, and reporting needs.Small studies may report preliminary results along the way; for large studies, articles should explain how their theories relate, with the thesis detailing how they hang together.Source: Standalone STGT Studies · p. 20

Combined STGT Studies

  • Combined STGT studies can investigate a problem space and provide theoretical foundations or taxonomies that inform later validation and solution development.Solutions may include tools, techniques, or platforms developed through other methods or an R&D project's development component.Source: Combined STGT Studies · p. 20

7.2 STGT for Data Analysis

  • Basic open coding, constant comparison, and memoing within another method yield descriptive categories and preliminary hypotheses, not mature theories without full STGT's advanced procedures.When both basic and advanced analysis are used within another method, researchers are strongly encouraged to consider full STGT because theoretical sampling and iterative, interleaved data collection and analysis underpin its theory claims.Source: §7.2 · pp. 20–21

7.3 Ethical Concerns

  • STGT studies should assess ethics early, obtain applicable committee approval, protect participant confidentiality, and seek consent or obscure identities when using public data where possible.The need to obscure identities is especially important when reporting could cause harm.Source: §7.3 · p. 21

7.4 Reporting STGT

  • Mature theories should be reported in formats that allow sufficient detail, while partial findings should be clearly identified and shared for research and practitioner feedback.For multiple papers from one study, researchers should explain the larger study and how each paper relates; feedback can reveal theoretical gaps and guide further development.Source: §7.4 · p. 21

8 STGT – EVALUATION GUIDELINES

STGT evaluation considers both method application and outcomes, using nuanced guidance rather than a one-size-fits-all checklist. It distinguishes preliminary or partial outcomes from mature theories.

8 STGT – EVALUATION GUIDELINES

  • Evaluation guidance covers both STGT method application and outcomes, distinguishing preliminary or partial results from mature theories rather than prescribing a checklist.Authors can use the criteria to strengthen manuscripts, and reviewers can use them to evaluate studies.Source: §8 · p. 21

8.1 Evaluating STGT Method Application

  • Method application is assessed for credibility and rigour, with practical evidence of how guidelines were applied and adequate evidence of data, analysis, and research artefacts.Examples include showing basic coding and constant comparison, embedding sanitised raw-data excerpts in findings, and supplying evidence of theory development for mature theories.Source: §8.1 · p. 22

8.2 Evaluating STGT Outcomes

  • Partial outcomes and preliminary theories are evaluated for originality, relevance, and density, including contextual fit and category depth supported by concepts, properties, and data examples.Originality may require explaining how researcher perspectives or biases were addressed; relevance can be validated through practitioner feedback, while findings should be more than intuitive or trivial.Source: §8.2 · pp. 22–23
  • Mature theories are held to higher standards: novelty, usefulness, parsimony, and modifiability, including actionable implications, compact explanation, and capacity to adapt to new evidence and contexts.A theory’s compact account should capture key categories and relationships, while its fuller description retains rich, nuanced findings; its generalisability can improve as it evolves across contexts.Source: §8.2 · p. 23

9 VISION AND NEW FRONTIERS

The article introduces Socio-Technical Grounded Theory (STGT) to identify software engineering’s socio-technical research context and provide clearer guidance for applying, reporting, and evaluating grounded theory studies. It also outlines ways STGT and the socio-technical framework may extend across disciplines, methods, and uses.

9 VISION AND NEW FRONTIERS

  • Traditional sociological grounded theory may be more applicable when a study concerns a primarily social phenomenon or its social nature rather than technical aspects.Source: §9 · p. 23
  • STGT defines socio-technical research context and offers philosophical, methodological, application, reporting, and evaluation guidelines to support researchers applying and assessing grounded theory.The article says these guidelines address challenges such as choosing among traditional GT versions and reporting preliminary, partial, and mature findings.Source: §9 · p. 23
  • STGT may support theory development in qualitative-data contexts where socio-technical factors matter, including computing and interaction research, and potentially medicine, education, and social sciences.Source: §9 · p. 23
  • The article’s socio-technical research framework may also inform methods beyond STGT, including socio-technical case studies, surveys, and ethnography.Source: §9 · p. 24
  • STGT can be used for full theory-development studies or data analysis alone, independently or alongside other methods in R&D studies and larger programs.Source: §9 · p. 24
  • Modern data sources and collection and analysis methods may expand grounded theories’ relevance and generalisability across contexts while retaining contextual nuance.Examples include GitHub, StackOverflow, Twitter, software-repository mining, natural language processing, and sentiment analysis.Source: §9 · p. 24
  • Advances in technology and artificial intelligence may supplement, augment, or automate parts of qualitative analysis, easing human effort and improving theory-development quality and scale.Source: §9 · p. 24
  • STGT’s systematic pattern-finding and constant-comparison procedures may also guide machine and deep learning in AI-based systems.Source: §9 · p. 24

Related papers

Loading 2103.14235v3…