Literature review and methodology in doctoral journalism research

1. Introduction

A doctoral dissertation is a public demonstration that its author can conduct research independently. Two components are particularly important in demonstrating this independence: the literature review and the methodology, including the research methods. The literature review shows whether the candidate understands the conversation into which the dissertation intervenes, and the methodology shows whether the evidence produced by the candidate is sufficiently reliable. At the Academy of Journalism and Communication, the regulation on doctoral training, as amended in 2026, gives both components formal weight. It makes the review essay a compulsory part of the programme, requires candidates to analyse and evaluate domestic and international works directly related to their topic, and sets a dissertation length of 150 to 180 pages of main text, followed by a defence that includes mandatory independent review [1].

However, formal requirements alone do not necessarily ensure research competence. Through teaching and supervising doctoral candidates, the author has observed that the relevant knowledge is scattered across Vietnamese methodology textbooks written for the research process in general, international handbooks that were not written with Vietnamese journalism in mind, and the experience of individual supervisors. The Vietnamese context matters here, since the 2025 Press Law defines Vietnamese journalism as revolutionary journalism, whose agencies serve as the voice of Party and State bodies, among other organisations, and as a forum of the people [2]. The problem has also acquired new urgency, because generative artificial intelligence (AI) is now used to retrieve and summarise sources and even to carry out literature reviews [3, pp. 3 and 8], while producing citations that sound plausible but are wrong [3, p. 7].

Against this backdrop, the paper asks how the literature conceptualises the literature review and the methodology of a doctoral dissertation, what tensions arise when these conceptions meet journalism research in the digital and AI era, and what these tensions imply for doctoral training in journalism studies in Vietnam. Its contribution is to bring review writing, research design and research integrity into one frame and to propose a framework of five competences for doctoral courses and their assessment.

2. Methodology

This paper adopts a critical narrative review design that incorporates selected procedural elements of systematic reviewing. The choice follows from the nature of the research question. The question of how scholarship conceives the review and the methodology of a dissertation is conceptual and normative, and the relevant documents are heterogeneous, ranging from monographs and peer-reviewed articles to guidelines issued by scholarly organisations and regulatory texts. Such heterogeneity precludes quantitative synthesis through meta-analysis, but it still calls for explicit selection criteria so that readers can check the argument. The paper therefore retains the interpretive and evaluative work of a narrative review, while making its sampling, screening criteria and analytical organisation transparent in the spirit of systematic reviewing.

The documents used in this paper were collected through purposive sampling, which suits the aim of a critical review to analyse representative conceptions without enumerating every published work. In 2025 and 2026, the author defined the scope of collection around five problem areas, namely review writing, review procedure, research design and methods, research integrity in the context of artificial intelligence, and the Vietnamese institutional framework for doctoral training and for journalism education at the Academy of Journalism and Communication. For each area, the author gathered methodological monographs and textbooks, journal articles, guidelines from scholarly organisations and regulatory texts issued by competent authorities.

The author retained only documents that met two conditions. Regarding the content, a document had to address reviewing, research design, analysis, academic writing or research integrity at the postgraduate level, or be a journalism study illustrating a specific method, or be a regulatory text currently governing doctoral training, scholarly publication or journalism in Vietnam. Regarding verification, the author had to hold the full text and to have checked its bibliographic details against the original. The author excluded every document that could not be traced to its source, including documents that appeared in the reference lists of other works. This condition matters all the more now that generative artificial intelligence can produce plausible but false citations, a problem examined in Section 3.4 on generative AI and research integrity.

Following screening, 131 documents were retained for analysis. By type, the corpus comprises 51 monographs and textbooks, 46 journal articles, 4 book chapters, 21 guidelines, reports and publications of scholarly organisations, and 9 Vietnamese regulatory texts. By language, it contains 108 English-language and 23 Vietnamese-language documents. By date, 92 documents were published in 2011 or later, 32 of them in 2021 or later, while the remaining 39 are earlier classic works, mainly on journalism theory and research methods.

The author reviewed the documents in full, grouped them by the problems they address and compared the groups to identify points of agreement, points of tension and open questions. The four lenses emerged from this comparison and structured the following section. Owing to space constraints, the discussion cites only a small number of representative documents for each lens. The paper has two limitations. First, the corpus includes only English-language and Vietnamese-language sources. Second, because of purposive sampling, it leans towards methodological guidance and regulatory texts rather than empirical studies of dissertations. The findings therefore offer a structured interpretation of the field, not an exhaustive inventory.

3. Results and discussion

3.1 The literature review as argument

A literature review is primarily an argument rather than an inventory of studies, and this is the view that methodological guides reiterate most often. Harris [4, pp. 5 and 13] treats the literature as the formal record of scholarly debates, to which new research is an entry, and defines a gap as something either not yet described or explained or explained poorly. In this sense, identifying a research gap requires more than establishing that a topic has not previously been studied in Vietnam. A gap becomes visible only within a larger coherent account that the researcher is trying to understand and develop [4, p. 5].

Writing guides translate this stance into textual operations. Cantero [5] separates summary, which restates the information of a source, from synthesis, which reorganises that information to reveal relations among sources, and from evaluation, which judges the value of sources within the synthesis itself. A review that places summaries side by side remains, in substance, an annotated bibliography [5]. This distinction explains a familiar weakness of Vietnamese dissertations, in which the review is divided into domestic and foreign studies and each half becomes a sequence of summaries. Cantero [5] proposes chronological, thematic and methodological organisation. A thematic or methodological structure, in particular, obliges the writer to compare sources rather than summarise them one by one.

The argumentative view also reshapes the place of theory. Harris [4, p. 6] describes what may be called a literature review trap, in which researchers beginning their first independent project follow the stories that others tell without a clear sense of their own purpose and get lost following first one theory and then another. The remedy is not to read less theory but to read it against the candidate's own question, asking which theories answer it, which contradict it and which bear on it only indirectly [4, p. 6]. A theoretical framework then emerges from an evaluation of the theories that prior studies have used, neglected or misapplied, rather than being announced before the purpose of the research is clear. Organisation matters for the same reason. When the review is built around problems instead of around the origin of sources, every section is forced to compare, and the reader can follow how the writer moved from the literature to the question.

3.2 The procedural turn in reviewing

A second approach shifts attention from the rhetoric of the review to its procedure. The PRISMA 2020 statement [6, pp. 1 and 3] updates the widely endorsed PRISMA 2009 reporting standard for systematic reviews. It provides a 27-item checklist together with an abstract checklist and revised flow diagrams. Its authors stress that it is a reporting guideline, which neither prescribes how to conduct a review nor assesses methodological quality. This procedural approach has also become increasingly evident in journalism studies. Danzon-Chambaud [7, pp. 3–4 and 12] combined several equivalent terms for automated journalism into a single search string, searched five databases for the period 2005 to 2020, screened close to 500 results and acknowledged that the restriction to English-language work limits the picture of the field. He also published a PRISMA checklist and flow diagram, identified institutional theory and Bourdieu's field theory as well-established lenses and recommended field theory for future research [7, p. 12]. This shows how a transparent review can lead directly to a reasoned theoretical choice.

For doctoral dissertations, which are rarely full systematic reviews, the value of this approach lies in the habits it cultivates. Recording every database, search string, filter and date of search turns the search into an object that supervisors and examiners can inspect and repeat. A tension nonetheless remains between transparency and coverage. The databases that make searches reproducible index Vietnamese-language scholarship poorly, so a seemingly rigorous procedure may nevertheless exclude relevant domestic literature that a Vietnamese dissertation most needs to engage. Procedural transparency therefore has to be complemented by deliberate searches of domestic journals and repositories, reported with the same care.

This approach suggests two practical habits for Vietnamese doctoral candidates. The first is to treat domestic sources as part of the protocol rather than as an afterthought, recording which Vietnamese journals, university repositories and dissertation collections were consulted, with the same detail as an international database. The second is to log screening decisions with their reasons. When a dissertation reports how many records were found, how many were read in full and why the others were set aside, examiners can question each step of the selection. They no longer have to judge the review by its overall impression. The candidate can also update the search quickly when the defence takes place long after the first round of reading.

The reporting items of PRISMA 2020 [6, p. 5] make these habits concrete. Authors are asked to name every database, register and website searched together with the date on which each was last consulted, and to present the full search strategy for each source, including any filters and limits. Doctoral candidates can adopt the same items in an appendix without claiming that their review is fully systematic. The review by Danzon-Chambaud [7, p. 4] adds a technique that is particularly useful at the doctoral level. To identify the theoretical foundations of research on automated journalism, he recorded every bibliographic reference cited more than five times across the selected studies. A candidate can do the same after the first few dozen entries of a review matrix, counting the authors and works that recur in the reference lists of the selected studies and then searching the publication records of those core authors for work that the original search missed.

3.3 Paradigm, design and context

The third lens concerns the coherence between philosophical assumptions and research design. For Creswell and Creswell [8], what defines mixed methods research is the integration of quantitative and qualitative data, grounded in a pragmatism that selects methods by their fit with the research purpose. The approach suits journalism well, since news products can be examined both as quantifiable data and as texts requiring interpretation. Yet integration, not mere co-presence, defines the design. A dissertation that reports a survey in one chapter and interviews in another, without connecting them, has used two methods but has not thereby conducted mixed methods research.

Creswell and Creswell [8] distinguish three core mixed methods designs, which suit different questions in journalism research. A convergent design collects quantitative and qualitative data in parallel and merges them in the interpretation, as when a survey of journalists is compared with in-depth interviews with newsroom managers. An explanatory sequential design begins with quantitative results and uses qualitative data to explain them, for instance when interviews help to interpret patterns found in a content analysis. An exploratory sequential design begins with qualitative exploration and builds on it a quantitative instrument, which is often the most realistic path for a new concept that has no validated measure in Vietnamese. Whichever design is chosen, the research proposal should state where and how the two strands meet, and the empirical chapter should show that point of integration, because it is what examiners will ask the candidate to defend.

Coherence also concerns the fit between concepts and context. Hanitzsch and Vos [9, pp. 146–148] note that the classic premise of Siebert and colleagues still finds traction, namely that the press takes on the form of the social and political structures within which it operates. They also criticise role research for having been framed largely within a Western model oriented to the media's contribution to democracy. For Vietnamese candidates, the implication is methodological as much as theoretical. Survey instruments and coding schemes borrowed from studies of Western press systems carry assumptions that may not hold for a press system defined by law as revolutionary journalism. A coherent design therefore runs from paradigm to question, from question to method, and from method back to the context that gives the concepts their meaning.

The same article shows how a review can widen the conceptual repertoire available to researchers. Hanitzsch and Vos [9, p. 146] conceptualise journalistic roles as discursive constructions of journalism's identity and place in society, and they identify 18 roles in the domain of political life that address six essential needs, alongside roles in everyday life related to consumption, identity and emotion. For a Vietnamese dissertation, such a repertoire is valuable not because every role applies, but because it allows the candidate to state explicitly which roles are relevant to Vietnamese revolutionary journalism, which need to be redefined and which have no counterpart, instead of importing a single democratic template without discussion. In the Vietnamese dissertation, this choice belongs to the chapter on the theoretical framework, in which the candidate operationalises the concept of role into indicators for the empirical chapter to measure and analyse.

3.4 Generative AI and research integrity

The fourth lens addresses the newest challenge to research integrity, namely generative artificial intelligence. The European Commission [3, pp. 7–8] identifies two forms of error that bear directly on reviews, namely citations that sound plausible but are wrong, and accurate citations accompanied by inaccurate summaries of the cited work. The same guidance [3, p. 8] counts carrying out a literature review or identifying research gaps among uses of AI that may have a substantial impact on the research. Researchers are asked to note such use in the methods section when it meaningfully shapes the results. They are also advised not to upload unpublished work of their own or of others to external AI systems without safeguards. Such errors can have particularly significant consequences in the literature review chapter, because a fabricated or misread source can enter both the reference list and the argument for the research gap.

Vietnamese guidance now mirrors these principles. The 2026 guidance of the Ministry of Science and Technology sets out a model institutional code for the use of AI. Under this code, AI is treated as a support tool and final responsibility for AI-generated content rests with the user. The tool, its version and the scope of its use are to be disclosed, and the use of AI to create non-existent references is prohibited [10, Appendix, Art. 16]. The practical conclusion for doctoral candidates is that AI may assist the search and the editing, but the reading, the evaluation and the argument must remain demonstrably their own. A simple rule follows, which this paper has applied to its own sources, namely that no source is cited unless its original is in hand.

In practice, the rule can be applied through a short verification routine for every source that enters the review. The first step confirms that the source exists, by locating it in a bibliographic database or on the publisher's site and checking that its identifier leads to the right document. The second step obtains the original and takes the year, edition and pages from the document itself rather than from a citation elsewhere. The third step checks that the cited passage actually supports the statement attached to it. The fourth step checks timing, so that a source is not used to explain a policy or an event that it predates. Recording the result of these steps in a dedicated column of the review matrix allows both the candidate and the supervisor to see at a glance which citations are ready for submission.

Disclosure becomes much easier when it is prepared from the beginning. A simple AI-use log can record the date, the tool and version, the task, the prompt, the part of the output that was kept and the check that was performed on it. Maintaining such a log requires relatively little additional effort during the research process and allows the candidate to write precisely the disclosure required when presenting methods in the research proposal and the introduction, rather than reconstructing months later which tool contributed to which step.

3.5 The Vietnamese institutional frame

These perspectives must also be considered within the Vietnamese institutional and regulatory context. The Academy's regulation, as amended in 2026, sets the training period at three years for master's degree holders and four years for bachelor's degree holders [1]. It requires the dissertation to cite at least five sources from the Academy's journal. It also requires the candidate to be the main author of qualifying publications that total at least 2.0 points under the scale of the State Council for Professorship and contribute substantially to the dissertation's findings. The defence is organised in three stages. The first is a unit-level assessment, which a dissertation passes only with the approval of five of the seven members present, after which the candidate has no more than 90 days to revise the text. The second is a mandatory independent review by two reviewers, and the third is a final council of seven members, at least three of whom come from outside the Academy [1]. Dissertations must also engage with the legal framework of the press, which since 1 July 2026 is the 2025 Press Law, replacing the 2016 law and defining Vietnamese journalism as revolutionary journalism that operates in a professional, humane and modern direction [2]. These requirements give the literature review and the methodology a gatekeeping function, yet they do not teach the reasoning that their thresholds presuppose. Doctoral training must therefore address the gap between formal regulatory requirements and the development of research competence.

The structure of the dissertation also shapes how the review and the methodology appear in the text. Unlike the model with a separate methodology chapter that international handbooks usually assume, a journalism dissertation at the Academy, in the author's experience, typically consists of an introduction and four chapters, namely the literature review, the theoretical framework, the empirical situation, and emerging issues and solutions. Research methods take shape mainly in the research proposal, which an Academy advisory council approves when the topic and supervisor are assigned [1], and the introduction then summarises them. Methodology in the broader sense runs through the whole dissertation. The review chapter establishes the gap and the research questions. The theoretical chapter selects theories, defines and operationalises concepts and builds a set of indicators. The empirical chapter collects and processes data according to those same indicators, and the final chapter derives emerging issues and solutions from the empirical findings. Because no single chapter gathers the methods in one place, the links between chapters become the main evidence of methodological quality, and examiners can judge that quality only when every link is visible in the text.

3.6 Towards an integrated framework of competences

The four lenses above suggest that the literature review and methodology constitute two interconnected stages in a coherent sequence of research decisions, running from the search log and the review matrix through the research proposal, the theoretical framework and its indicators to the empirical data, the proposed solutions and the defence. On this basis, the paper proposes five competences for doctoral candidates in journalism studies. The first is argumentative positioning, the ability to place the dissertation within an ongoing scholarly conversation and to derive a gap from a synthesis of prior work. The second is procedural transparency, the ability to document searches, criteria and decisions so that others can inspect and repeat them. The third is paradigmatic coherence, the ability to align assumptions, questions, design and analysis, from the operationalisation of concepts to the processing of data, including the integration of data in mixed designs. The fourth is verification and integrity, the ability to confirm the existence and content of every source and to use and disclose AI responsibly. The fifth is contextual and institutional literacy, the ability to adapt concepts and instruments to Vietnamese revolutionary journalism and to meet current regulations. Each competence can be assessed through an artefact, namely a search protocol table, a review matrix with a verification column, a map linking each question to its concepts, indicators, data and analysis, and an AI-use log.

Within a doctoral programme, methods courses can arrange the five competences in the order in which candidates meet them while preparing a dissertation proposal. Positioning and transparency belong to the early sessions, when candidates define their questions and build their first search protocol. Coherence belongs to the sessions on design and methods, when each research question is matched with data and analysis. Integrity runs through the whole course, while contextual literacy is best taught through the candidates' own topics, since the adaptation of concepts to Vietnamese journalism differs from one field of inquiry to another. Assessment can then follow the same sequence, so that supervisors see the chain of decisions behind a proposal and not merely its final text.

4. Conclusion

This paper has examined how scholarship conceives these questions in order to clarify how doctoral training in journalism studies in Vietnam might treat the literature review and the methodology of a dissertation. Scholarship consistently conceptualises the literature review as an argument, in which gaps emerge from synthesis rather than from the absence of local studies. Journalism scholarship has embraced a procedural turn towards transparent searching, although the databases that support transparency underrepresent Vietnamese-language work. Coherent design requires an alignment from paradigm to method and back to the context of revolutionary journalism. In the Vietnamese dissertation, this alignment does not sit in a separate methodology chapter but runs from the research proposal through all four chapters. Generative AI has opened new integrity risks, to which international and Vietnamese guidance respond with consistent principles of human responsibility, disclosure and verification.

The main contribution of the paper is a framework of five competences, each linked to an assessable artefact, offered as a basis for course design, supervision and examination rather than as a finished standard. Its limits follow from a purposive sample that leans towards methodological guidance and regulatory texts and is restricted to English and Vietnamese sources. Future research could test the framework by analysing review chapters of recent journalism dissertations or by surveying supervisors and candidates, to strengthen the independence of scholarly reasoning and the verifiability of citations in future doctoral dissertations.

In preparing this article, the author used Claude (Anthropic, Claude Opus 5.5) to review the source materials on which the article draws, propose a structure, draft and revise the English text, translate it into Vietnamese, edit language and format references in the Vancouver style, and check each citation against the original source. The author defined the research problem and the argument, selected all sources from the project's document repository, reviewed and revised the text and finalised it. The AI tool generated no data and no references, and the author entered no confidential or unpublished data into it. The author takes full responsibility for the content of the article.