Skip to Content
You are currently on the new version of our website. Access the old version .
SocietiesSocieties
  • Article
  • Open Access

5 June 2025

Generative AI as a Teaching Tool for Social Research Methodology: Addressing Challenges in Higher Education

Department of Sociology and Social Research, University of Milano Bicocca, 20126 Milan, Italy
This article belongs to the Special Issue Digital Learning, Ethics and Pedagogies

Abstract

Teaching social research methodology in university courses, whether qualitative or quantitative, presents significant challenges for both instructors and students. These challenges include the availability of empirical datasets, the illustration of data analysis techniques, the simulation of research report writing, and the facilitation of scenario-based learning. Emerging AI tools, such as ChatGPT-4, offer potential support in higher education, though their effectiveness depends on the context and their integration with traditional didactic methods. This article explores the potential of AI in teaching social research methodology, with a focus on its benefits, limits and ethical considerations. Furthermore, the paper presents a case study of AI application in teaching qualitative research techniques, specifically in the analysis of solicited documents. Generative AI shows the potential to improve the teaching of social research methodology by providing students with opportunities to engage in hands-on learning, interact with realistic datasets and refine their analytical and communication skills. The integration of AI in education should, however, be approached with a critical mindset, ensuring that AI tools serve as a means to sharpen (not replace) traditional methods of learning.

1. Introduction

In higher education, teaching social research methodology requires a balance between theory and practical application. Students need to understand how to design research projects, collect data, and conduct empirical analysis. They also need to learn how to apply these skills in real research contexts and familiarize themselves with practical applications. However, several challenges arise in applied sessions, such as limited access to datasets suitable for specific needs and difficulties in mastering data analysis techniques.
One of the main challenges in teaching social research methodology is providing students with appropriate empirical datasets for practice. While quantitative datasets are often accessible through statistical repositories, qualitative data, such as interview transcripts, ethnographic notes, or focus group discussions, are far less available. These materials should not only be accessible but also relevant to students’ fields of study, possess appropriate levels of complexity for educational purposes, and comply with privacy and ethical guidelines.
Students also encounter challenges when transitioning from theoretical knowledge to the application of quantitative and qualitative data analysis. This requires proficiency with statistical and coding software. Another difficulty for students is becoming familiar with analytical outputs even before being able to conduct analyses independently. Having access to analytical results allows students to become accustomed to reading and interpreting them. They need to learn to link these results to theoretical frameworks and report conclusions in research documents to share with the scientific community.
Another key pedagogical aspect is the opportunity for students to work on topics that interest them. This approach helps bridge the gap between theoretical knowledge and real-world applications, preparing students for both academic and professional settings. It is especially valuable when exploring emerging topics or addressing issues related to hidden populations, ensuring that students develop a broader perspective.
Traditional teaching methods often fall short of fully addressing the practical and ethical complexities of teaching social research methodology in higher education, particularly when it comes to student engagement and applied learning [1,2]. This article addresses these didactic challenges by focusing on university-level courses in social disciplines such as sociology, social work, tourism, media, and marketing. It explores the potential of new technologies, especially generative artificial intelligence tools like ChatGPT, to support teaching practices while encouraging critical thinking and ethical awareness. While the focus is on showcasing the opportunities that AI offers, this is not intended as an uncritical endorsement of these tools. Instead, the discussion aims to contribute to a broader debate by critically reflecting on AI’s role, acknowledging both its benefits and limitations, and highlighting some issues, particularly in relation to methodological rigor, ethical concerns, and its potential impact on students’ critical thinking skills.

2. Generative AI and Its Potential in Education

In recent years, generative artificial intelligence has gained increasing attention in various fields, including education. AI-powered tools are being explored as potential resources for improving learning experiences, enhancing student engagement, and supporting academic research (for an overall background on AI in education, see [3,4]).
Generative AI can provide real-time tutoring and guidance, offering students additional support during study sessions. AI-driven tutoring systems can enhance learning by offering personalized clarifications and interactive engagement that adapt to students’ needs [5]. Additionally, AI can play a crucial role in research and writing assistance, aiding students in crafting research proposals, analyzing data, and simulating research scenarios (for an example, see [6]). Another important contribution of AI is the generation of customized learning materials. AI can create tailored exercises, summaries, and explanations designed to align with individual learning needs [7].
Beyond traditional academic settings, AI also supports autodidactic learning and open education. AI can foster independence and autonomy among self-directed learners by offering guidance and adaptive learning [8]. However, concerns remain regarding the extent to which AI should be integrated into education. Scholars caution that AI should not replace human educators but rather complement them by automating repetitive tasks and allowing instructors to focus on higher-order teaching activities [1,9]. While AI enhances efficiency and accessibility, human oversight remains essential to ensure critical thinking, ethical considerations, and the development of analytical skills among students [2].
In the context of social research methodology courses, AI offers new ways to overcome traditional challenges, such as the lack of real-world data and the difficulty of applying theoretical concepts to practical exercises.
Generative AI can be used to create synthetic datasets for classroom exercises, though these should be carefully evaluated to ensure they accurately reflect real-world complexity. In the domain of quantitative research, AI-generated data, including demographic datasets, survey responses, and time series, can offer students the opportunity to engage in statistical analysis without concerns related to data privacy or availability. AI-generated datasets can be adjusted to different levels of complexity, making them suitable for learners with varying expertise. In qualitative research, AI-generated interview transcripts, focus group discussions, ethnographic narratives, and documents could provide students with opportunities to practice qualitative coding, thematic analysis, and data interpretation. This flexibility ensures that students can work on emerging research topics and real-world scenarios without the challenges of data collection. Moreover, instructors can customize AI-generated datasets to reflect specific fields such as urban studies, labor sociology, or migration research, ensuring relevance to different academic disciplines. By tailoring the level of complexity, AI-generated content can become an adaptable resource, enabling the integration of real-world case studies into the curriculum.
AI can be integrated into practical exercises to expose students to different research approaches. However, it is crucial to encourage personal analysis and critical evaluation to prevent over-reliance on AI-generated insights. Through interactive engagement with AI-generated data, students can develop their ability to formulate research questions, analyze responses, and compare different analytical approaches. In quantitative analysis, students can work with AI-generated datasets to practice calculating simple descriptive measures such as means, standard deviations, and frequency distributions. More advanced learners can explore multivariate statistical techniques, including regression analysis, clustering, and factor analysis. In qualitative research, AI-assisted coding exercises enable students to identify themes, categorize textual data, and refine their ability to conduct qualitative research. AI-generated analyses also offer an opportunity for students to compare their manual work with AI-assisted outputs, encouraging critical reflection on methodological approaches and the reliability of AI-driven insights.
A further application of AI in teaching social research methodology involves its role in training students to interpret analytical outputs and utilize them in research communication. By working with AI-generated outputs, students can learn to construct research reports integrating figures and visualizations. AI tools can help students connect their analytical results to theoretical frameworks and hypotheses, reinforcing the relationship between empirical research and conceptual models.

3. An Example of Application of Generative AI in Teaching Social Research Methodology

The objective of this article is to explore how generative AI, specifically ChatGPT, can support the creation of educational content for teaching qualitative methods in social research. To do this, I adopted an exploratory case study approach, which can be used in qualitative research to investigate emerging tools and practices in context [10]. This approach is particularly suitable for evaluating novel educational technologies in the early phases of implementation.
The case study presented here is a teaching-focused simulation in which the instructor (the Author) interacts with ChatGPT to generate datasets, conduct analysis, and produce teaching materials for use in university-level social research methodology courses. The prompts submitted to the AI will be presented along with its responses; for the sake of brevity, in some cases, only excerpts of the responses will be shown, accompanied by a summary.
To strengthen contextual relevance, prompts were constructed to simulate realistic classroom scenarios and data needs. Each phase was carefully documented: definition of pedagogical objectives and expected student learning outcomes; construction and refinement of prompts to generate realistic qualitative data; thematic analysis of AI-generated texts and comparison with academic standards; evaluation of output utility and limitations based on didactic criteria. A critical reflection is presented in the Discussion Section. This structure aims to ensure transparency and replicability, allowing instructors or researchers to adapt the approach to their own teaching or methodological goals.
I have chosen to use an example of qualitative research, specifically a study that employs solicited diaries (for a review see [11]), to provide students with realistic, ethically sound textual data for thematic analysis practice. While in traditional documentary analysis, documents are spontaneously produced by individuals or groups for their own purposes, in research using solicited documents, the materials are created at the request of the research team for investigative purposes, also online [12]. Solicited documents occupy an intermediate space between documentary analysis and qualitative interviews. This makes them an excellent teaching tool, as they closely resemble both interview transcripts and naturally occurring texts, providing students with a versatile and practical approach to qualitative data analysis.
A key consideration when using generative AI is the formulation of effective instructions, i.e., prompts. A well-structured prompt should be precise, unambiguous, and provide relevant context. Moreover, a strong prompt should specify the role ChatGPT is expected to take on and the objective to be achieved (in this case, acting as an instructor preparing course materials) [13].
As a first step, I asked ChatGPT to generate the empirical dataset. Specifically, I requested the creation of ten diary entries, imagining that they were written by ten tourists who had been assigned this task by researchers before their departure. The diaries should recount the first day of travel, allowing for an exploration of personal expectations and concerns. I also asked ChatGPT to differentiate within the imagined sample. This dataset serves as a foundation for students to practice data familiarization, thematic coding, and interpretation. ChatGPT responded by generating ten diary entries describing a hypothetical trip to Barcelona. Each entry represents a different traveler, varying in age, gender, educational background, and family status (see Table 1).
Table 1. ChatGPT generates the empirical dataset.
Then, I asked ChatGPT to provide a summary of the participants’ profiles. The response includes a description of each individual, outlining their reasons for traveling, expectations, and any concerns (Table 2). This output models the process of data reduction and abstraction, a crucial step in qualitative analysis training.
Table 2. ChatGPT provides a summary of the participants’ profiles.
Since I imagined I intended to use the material for more advanced students, I asked ChatGPT to generate longer and more detailed diary entries, incorporating multiple themes in each. I requested different levels of complexity and the inclusion of specific topics. ChatGPT quickly adjusted all ten diary entries according to my specifications (Table 3). This version is intended to allow for deeper engagement with layered narratives and support higher-order analytical skills.
Table 3. ChatGPT generates longer and more detailed diary entries.
Then, I asked ChatGPT to perform a qualitative thematic analysis on the ten diary entries (using their more extended version). ChatGPT identified six recurring themes and provided summaries. This material can be useful to illustrate to students the expected output of qualitative thematic analysis, specifically the identification of emerging themes. Additionally, it can serve as a comparison for their independent work in theme identification. A key consideration is that in the second part of the response, ChatGPT seems to lean towards frequency analysis and link themes to individual characteristics. This requires careful attention, as qualitative analysis should not move in the direction of establishing associations or, even more so, making generalizations. This is a strong limit of the response to be considered (Table 4).
Table 4. ChatGPT performs a qualitative thematic analysis of the ten diary entries.
I tried to adapt the empirical dataset for a different degree program. I asked ChatGPT to rewrite the diaries, adjusting them to new scenarios. ChatGPT rewrote all ten diaries while maintaining diversity among subjects and perspectives. As an example, I requested an adaptation for a social work course, suggesting the idea that the diary entries represent the first day of admission to a care facility. ChatGPT generated the material, ensuring the absence of ethical or privacy issues as there would be in a real-world setting (Table 5).
Table 5. ChatGPT adapts the empirical dataset for a different scenario.
I returned to the travel diaries and asked ChatGPT to present the themes it had previously identified, accompanied by one or more illustrative excerpts from each diary. This step reinforces students’ understanding of linking codes to evidence and enhances transparency in qualitative reporting. Moreover, this approach provides material to show students the first steps in organizing qualitative data (Table 6).
Table 6. ChatGPT presents the themes it had previously identified with illustrative excerpts from diaries.
Next, I asked ChatGPT to prepare a draft of the results report. I requested that, after a brief introduction, each identified theme be presented with a short explanation accompanied by excerpts from the diaries (Table 7). This function enables learners to explore academic conventions in presenting thematic findings, including how to integrate data excerpts.
Table 7. ChatGPT prepares a draft of the results report.
Next, I asked ChatGPT to connect the empirical results with theoretical concepts. ChatGPT added bibliographic references. When I inquired whether the citations were real or fabricated, ChatGPT admitted that citations were a mix of both and offered to include only real citations if I preferred. In the end, it provided a list of the cited references (Table 8). While some references were fabricated, the process can serve as a pedagogical tool to highlight the importance of source validation.
Table 8. ChatGPT tries to connect the empirical results with theoretical concepts.
Finally, I asked ChatGPT for a summary table of the results. ChatGPT provided a structured table summarizing the identified themes, their descriptions, illustrative excerpts from the diaries, and theoretical references (Table 9). This format models how results can be synthesized for academic presentations and student assignments.
Table 9. ChatGPT provides a summary table of the results.
Until now, I was asking ChatGPT to produce and analyze texts, but research practice increasingly incorporates images and other multimedia formats. Then, I asked ChatGPT to generate images for each diary entry, assuming that researchers have requested participants to attach a representative image of their day, expanding the exercise into the domain of visual analysis. This encourages multimodal thinking and allows for the introduction of semiotic or compositional methods. ChatGPT provides me with varied images that align with the different personalities of the subjects. I report an example (Table 10). Students can be introduced to visual analysis methods, such as semiotic analysis [14], compositional analysis [15], content analysis [16], or visual ethnography [17].
Table 10. ChatGPT generates images for each diary entry.
I was returning to written diaries and asking ChatGPT to suggest an original way to present the results of a qualitative thematic analysis. It provided me with ten different suggestions, with a good level of originality (Table 11). The responses included innovative formats such as podcasts, story maps, and digital exhibits, which can inspire students to think beyond conventional reporting.
Table 11. ChatGPT proposes original ways to present the results of the qualitative thematic analysis.
To continue testing ChatGPT’s ability to generate ideas, I asked for an original way to analyze diaries. ChatGPT suggested ten different possibilities, each accompanied by a reflection on its strengths and uniqueness. Additionally, it provided recommendations on tools that can be used for that type of analysis and included a brief example (Table 12). This exercise helps illustrate how digital tools can stimulate methodological innovation in teaching contexts.
Table 12. ChatGPT proposes an original way to analyze the diaries.

4. Discussion

The integration of generative AI, particularly through tools like ChatGPT, represents a transformative opportunity to address some of the challenges faced in teaching social research methodology. In the recent literature, various studies have explored the use of generative AI in higher education (for instance, [1,2,7]). Compared to these contributions, the present study focuses more directly on qualitative methodology instruction and emphasizes the structured use of AI to replicate specific phases of the research process (data collection, analysis, and presentation) within a simulated teaching environment.
In the work session presented in this article, I used ChatGPT-4 to generate, analyze, and present data according to standard (and basic) disciplinary methods. ChatGpt has provided coherent responses generally aligned with methodological literature. When provided with structured and specific instructions, it produced outputs that were consistent with methodological expectations and educational goals. These results confirm that generative AI can function as a useful support for instructors in creating content such as qualitative datasets and analysis examples, while maintaining ethical neutrality.
Prompts played a crucial role in guiding ChatGPT toward the creation of realistic diary entries, the identification of thematic patterns, and the generation of analytical reports. Refining these prompts followed an iterative process aimed at increasing specificity, reducing ambiguities, and ensuring alignment with pedagogical objectives. For instance, when asking ChatGPT to generate solicited diaries, the prompt explicitly included information about the travel scenario, the demographic diversity of participants, and the diversity of themes to be embedded in the narratives.
Despite these refinements, some limitations persisted, highlighting the challenges of using generative AI in qualitative research teaching. For instance, certain diary entries exhibited repetitive phrasing or overly structured narratives. The reliance of the AI model on existing data patterns occasionally led to subtle biases, particularly Eurocentric representations, and assumptions tied to middle-class travel experiences. Limitations also include the fact that the AI partially generated fictitious bibliographic references and, at times, leaned towards a quantitative analysis of qualitative data. While not problematic in a controlled classroom setting, where the role of instructors is crucial, these limitations call for critical review and contextual framing when used in diverse educational environments.
The integration of AI into educational content production also introduces critical considerations related to reliability, creativity, ethics, and authorship [18]. In terms of reliability, although the materials generated by ChatGPT were generally coherent and pedagogically functional, they require instructor oversight to ensure factual consistency and conceptual accuracy. The creative dimension of AI outputs remains shaped by the limitations of the underlying language model. This calls into question the extent to which AI-generated materials can offer true novelty or methodological depth.
Ethical implications are equally central. As noted by some authors [19], the use of generative AI in academic contexts must be transparent, particularly when it comes to acknowledging machine-generated content, evaluating embedded biases, and avoiding the uncritical acceptance of outputs. These concerns are not only technical but pedagogical: students should be encouraged to question the processes that produce data and to understand the limits of automated analysis.
Finally, the question of authorship becomes increasingly complex when AI contributes to the structure, phrasing, or interpretation of content [20]. Although the creative direction and evaluation remain under human control, the boundaries between human and machine contributions are blurred. This opens a broader debate on academic integrity and intellectual property in AI-assisted education and invites educators to cultivate reflective discussions around these issues.
Based on the findings of this case, a set of best practices can be proposed. These include refining prompts to enhance output quality, critically evaluating AI-generated material, comparing it with real-world examples, and fostering student awareness of the limitations and biases inherent in such technologies. When used transparently and reflectively, generative AI can enrich teaching practices by enabling the creation of adaptable learning materials and stimulating methodological discussion. Nonetheless, this integration must be guided by critical engagement with ethical, epistemological, and pedagogical concerns.

5. Conclusions

This article explored how generative AI, and ChatGPT in particular, can be used to create content for teaching research methodology at the university level. Through an applied example involving solicited diaries and qualitative thematic analysis, the article examined both the capabilities and limitations of AI-generated educational material.
The main contribution lies in presenting a replicable instructional approach that combines prompt engineering, simulated datasets, and critical evaluation exercises. This approach enables instructors to overcome common teaching challenges, such as the scarcity of usable qualitative data, to offer students the opportunity to engage with realistic and ethically safe research materials.
Based on the findings and reflection developed in the discussion, several recommendations can be offered. AI-generated content should always be accompanied by instructor-led contextualization to ensure conceptual rigor and avoid misinterpretation. Prompts must be iteratively refined to align with specific learning goals and to reduce generic or biased outputs. Exercises that involve comparing human and AI analyses can support students in developing critical methodological skills. Ethical considerations, including transparency, source evaluation, and authorship awareness, should be explicitly integrated into the teaching design.
While generative AI shows promise as a content-creation tool, its effective use in education depends on critical guidance, human oversight, and pedagogical integration. Future research could expand this approach by involving students directly in AI-assisted exercises or by comparing learning outcomes between traditional and AI-supported teaching environments.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Data Availability Statement

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Acknowledgments

During the preparation of this manuscript, the author used ChatGPT-4 (OpenAI, 2024) for the purposes of generating simulated qualitative data, drafting thematic analyses, and exploring pedagogical applications. The author has reviewed and edited the output and takes full responsibility for the content of this publication.

Conflicts of Interest

The author declares no conflicts of interest.

References

  1. Dempere, J.; Modugu, K.; Allam, H.; Ramasamy, L. The impact of ChatGPT on higher education. Front. Educ. 2023, 8, 1206936. [Google Scholar] [CrossRef]
  2. Nguyen, K.V. The Use of Generative AI Tools in Higher Education: Ethical and Pedagogical Principles. J. Acad. Ethics 2025. [Google Scholar] [CrossRef]
  3. Davar, N.F.; Dewan, M.A.A.; Zhang, X. AI Chatbots in Education: Challenges and Opportunities. Information 2025, 16, 235. [Google Scholar] [CrossRef]
  4. Pham, S.T.; Sampson, P.M. The development of artificial intelligence in education: A review in context. J. Comput. Assist. Learn. 2022, 38, 1408–1421. [Google Scholar] [CrossRef]
  5. Srinivasa, K.; Kurni, M.; Saritha, K. Learning, Teaching, and Assessment Methods for Contemporary Learners, Pedagogy for the Digital Generation; Springer: Berlin/Heidelberg, Germany, 2022. [Google Scholar] [CrossRef]
  6. Arosio, L. From Diaries to Digital: The Role of AI in Web-Mediated Documentary Analysis. Sociol. Study 2024, 14, 213–227. [Google Scholar] [CrossRef]
  7. Owoseni, A.; Kolade, O.; Egbetokun, A. Generative AI and Its Implications for Higher Education Students and Educators. In Generative AI in Higher Education; Palgrave Macmillan: Cham, Switzerland, 2024. [Google Scholar] [CrossRef]
  8. Firat, M. How Chat GPT Can Transform Autodidactic Experiences and Open Education? preprint 2023. [Google Scholar] [CrossRef]
  9. Felix, C. The role of the teacher and AI in education. In International Perspectives on the Role of Technology in Humanizing Higher Education Innovations in Higher Education Teaching and Learning; Sengupta, E., Blessinger, P., Makhanya, M.S., Eds.; Emerald Publishing Limited: Bingley, UK, 2020; Volume 33, pp. 33–48. [Google Scholar] [CrossRef]
  10. Yin, R.K. Case Study Research and Applications: Design and Methods; Sage: Thousand Oaks, CA, USA, 2018. [Google Scholar]
  11. Akinreni, T.; Donald, M.; Reñosa, C.; Putri, A.Z.; Scott, K.; McMahon, S.A. A methodological review of solicited diaries as a qualitative tool in health research in low- and middle-income countries. SSM—Qual. Res. Health 2024, 6, 100492. [Google Scholar] [CrossRef]
  12. Arosio, L. Personal documents on the Internet: What’s new and what’s old. J. Comp. Res. Anthropol. Sociol. 2010, 1, 23–38. [Google Scholar]
  13. Open AI. Best Practices for Prompt Engineering with the OpenAI API. Available online: https://help.openai.com/en/articles/6654000-best-practices-for-prompt-engineering-with-the-openai-api (accessed on 3 June 2025).
  14. Barthes, R. Camera Lucida: Reflections on Photography; Hill and Wang: New York, NY, USA, 1981. [Google Scholar]
  15. Rose, G. Visual Methodologies: An Introduction to Researching with Visual Materials; Sage: Thousand Oaks, CA, USA, 2016. [Google Scholar]
  16. Larsen, S.; Brun, W.; Øgaard, T. What tourists worry about—Construction of a scale measuring tourist worries. Tour. Manag. 2009, 28, 260–265. [Google Scholar] [CrossRef]
  17. Pink, S. Doing Visual Ethnography; Sage: Thousand Oaks, CA, USA, 2013. [Google Scholar]
  18. Shahzad, M.F.; Xu, S.; Zahid, H. Exploring the impact of generative AI-based technologies on learning performance Through self-efficacy, fairness & ethics, creativity, and trust in higher education. Educ. Inf. Technol. 2025, 30, 3691–3716. [Google Scholar] [CrossRef]
  19. Dwivedi, Y.K.; Kshetri, N.; Hughes, L.; Slade, E.L.; Jeyaraj, A.; Kar, A.K.; Baabdullah, A.; Koohang, A.; Raghavan, V.; Ahuja, M.; et al. “So what if ChatGPT wrote it?” multidisciplinary perspectives on opportunities, challenges, and implications of generative conversational AI for research, practice, and policy. Int. J. Inf. Manag. 2023, 71, 102642. [Google Scholar] [CrossRef]
  20. Watson, S.; Brezovec, E.; Romic, J. The role of generative AI in academic and scientific authorship: An autopoietic perspective. AI Soc. 2025. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.