Quantitative, Qualitative, and Mixed Methodologies in Digital Social Sciences

📜🔗LINK to the source

1. ANALYTICAL SUMMARY

Context and issues — Social research in the digital age

Social life and data production are migrating online: platforms, APIs, sensors, vast textual corpora and networks. This shift opens up unprecedented possibilities (scale, hard-to-reach populations) but raises challenges of quality, representativeness, privacy and technical competence (p.1-9). The work is aimed at doctoral students, young researchers and applied researchers in sociology, psychology, education, public policy, communication and management. Its central thesis: contemporary problems require a "methodological bilingualism", capable of mobilising statistical rigor as well as interpretative depth (p.7-8).

Operational contributions — An integrated framework, from design to rigor

The volume unfolds a complete journey: principles of the quantitative (p.11) and the qualitative (p.26), logic of mixed methods (p.47), study design (p.63), traditional and digital data collection (p.88), strategies for statistical, qualitative and computational analysis (p.97), triangulation and integration (p.119), criteria of rigor (p.135). For each step, it specifies the digital adaptations (online surveys, digital ethnography, social network mining, AI) and ethical safeguards. The guiding thread: "innovation with integrity" — adopting new tools without relaxing the demands of validity, reliability and transparency.

2. KEY POINTS OF THE DOCUMENT

1. Five reasons to combine methods (decision-making framework directly transposable). Drawing on Greene, Caracelli and Graham (1989), the work distinguishes triangulation, complementarity, development, initiation and expansion (p.47-48). This framework helps to explicitly justify why we mix — and not just to juxtapose a survey and interviews.
2. Formalised mixed designs: sequential, concurrent, nested. Chapter 3 describes how to chain or nest quantitative and qualitative phases, with the example of a prior qualitative phase (post-conflict Rwanda) to make a questionnaire culturally relevant (p.47-62).
3. Overview of computer-assisted analysis (CAQDAS). Chapter 6 directs towards NVivo, MAXQDA, ATLAS.ti and Dedoose for storing, coding and querying large qualitative volumes, while reminding that interpretation remains a human act (p.107-108).
4. An ethical checklist for digital collection. Informed consent, confidentiality, anonymisation, adherence to platform rules, reference to the AoIR 2019 guidelines and ethics committees (IRB): chapter 5 states that observing public interactions is generally accepted, but capturing private groups is unethical (p.92, p.94-95).
5. Rigor in the face of 'big social data'. boyd and Crawford (2012) are invoked to remind that more data does not equal better data: large corpora often come from convenience samples that amplify existing biases. Hence data audits, pre-registration of hypotheses (Bakker et al., 2019) and code sharing against 'p-hacking' (p.138-140).

3. ACTION PATHS FOR LOCAL ACTORS

1. Clarify the intention before mixing. Before an evaluation, choose your reason for combining methods from the five rationales of Greene et al. (p.47-48): corroborating a result, explaining it, preparing a tool, exploring a contradiction or broadening the scope.
2. Sequencing a participatory evaluation. Rely on the sequential designs of chapter 3: first gather the voices of the audiences (focus groups, interviews) to build a culturally adapted questionnaire, rather than the reverse (p.47-62).
3. Structure the qualitative coding. For a large collection of testimonies (local diagnosis), turn to a CAQDAS (NVivo, MAXQDA, ATLAS.ti or Dedoose) to code, query and visualise the themes (p.107-108) — while keeping the interpretation under human responsibility.
4. Secure the use of data from social networks. Apply the safeguards from chapter 5: do not exploit private groups, anonymise and aggregate even if the content is public, respect the conditions of the platforms, refer to the AoIR guidelines (p.92, p.94-95).
5. Cross the sources by triangulation. Use the "joint displays" / matrices from chapter 7 to confront numerical results and qualitative themes: convergences strengthen the conclusion, divergences signal a gap between what is said and what is done to investigate (p.119-134).
6. Ensure the reliability of any automated digital data. Before concluding from online reviews or traces, check the composition of the corpus, compare to known demographic data, redo a collection to test coherence (reliability audit), and pre-register your analysis plan (p.138-140). Unmet need: the work does not offer a ready-made protocol suitable for health promotion — adaptation to the context remains to be done by the reader.

4. ADDITIONAL REFERENCES

1. WHO — Artificial intelligence and evidence-informed policy: emerging challenges and opportunities (discussion document, June 2026). Extends chapters 6 and 8 on AI in the production and synthesis of evidence for public decision-making. Verified URL: https://www.who.int/news/item/02-06-2026-new-who-discussion-paper-sets-out-opportunities-and-risks-of-ai-in-evidence-informed-health-policy
2. Cochrane, Campbell Collaboration, JBI & CEE — Position statement on AI use in evidence synthesis / RAISE recommendations (2025). Methodological supplement on the responsible use of AI and the author's responsibility in evidence synthesis. Verified URL: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12594113/
3. ASPQ — Practical guide: Towards participatory public health (Association for Public Health of Quebec, published 2026). Bridges to the Francophone field: concrete guidelines for integrating citizen participation, where the work remains theoretical. Verified URL: https://aspq.org/guide-pratique-vers-une-sante-publique-participative/

5. FREQUENTLY ASKED QUESTIONS (FAQ)

1. Who is this book for? For researchers, doctoral students, and research-oriented professionals in social sciences and management (Introduction, p.1-9). It is not a field intervention manual.
2. Why combine quantitative and qualitative? To compensate for the weaknesses of each approach: the numbers provide scale and generalisation, the narrative provides meaning and context (p.47-48).
3. What are the main types of mixed designs? Sequential, concurrent, and nested, depending on whether the phases follow one another, run in parallel, or fit together (chapter 3, p.47-62).
4. What software is used to analyse qualitative data? NVivo, MAXQDA, ATLAS.ti, Dedoose — for organising, coding, and querying; interpretation remains human (p.107-108).
5. Can social media data be used freely? Not without precautions: observing public content is generally accepted, exploiting private groups is not; anonymisation and respect for platforms are essential (p.92, p.94-95).
6. What is triangulation, concretely? Cross-referencing methods, sources, theories or researchers to verify a result; convergences and divergences are treated differently (chapter 7, p.119-134).
7. How to maintain rigor with big data and AI? Check the biases of the corpus, pre-register your hypotheses, share data and code, maintain human control over 'black box' algorithms (p.138-140, Conclusion p.145).

6. REWRITING IN EASY TO READ LANGUAGE

What is this book about?

This book discusses methods for conducting research. It helps to study society. Today, much data comes from the Internet. The book shows how to use this data effectively.

Two ways to search

There are two main ways to search.

• Numbers: we count and we measure.
• Stories: we listen and we understand people.
  The book says: the two ways go well together.

Mixing the two well

We can conduct a survey with numbers. We can also conduct interviews. We put the two together. We see reality better.

Using the Internet with caution

We can use data from social networks. But we must respect rules.

• We ask for people's consent.
• We protect their name and their privacy.
• We do not take data from private groups.

Stay rigorous

A lot of data does not mean good data. It needs to be verified. We must be honest about the limitations. We note what we have done, step by step.

7. CROSS-ANALYSIS — VALUES OF HEALTH PRACTICES

• Literacy : low in direct — the work is a dense academic text in English, without a simplified version; it does, however, thematise the clarity of methodological terms (p.5-6).
• Empowerment : addressed indirectly through the transformative paradigm and participatory approaches (chapter 3), but the audiences are mostly "participants", rarely co-designers.
• Participation : present through digital ethnography and mixed designs engaging communities; co-construction is not centrally formalised.
• Community health : illustrated by an example (epidemiological data + focus groups on health beliefs, p.48), but the collective dimension remains one case among others.
• Ethics : strong point — consent, confidentiality, anonymisation, AoIR/IRB lines, algorithmic biases and "p-hacking" are addressed (chapters 5, 6, 8).
• Human rights : equity through inclusive sampling of hard-to-reach populations (p.13) and privacy protection; non-maleficence principle affirmed.
• Intersectorality : dimension mainly interdisciplinary (sociology, psychology, education, public policy, communication, management) rather than partnership in the field sense.
• Partnership : mentioned via triangulation by researchers (investigator triangulation); no formalised collaboration model with field actors.
• Fight against discrimination : addressed at the data level — algorithmic biases (fairness), non-representative corpora that marginalise certain groups, big data that amplifies existing biases (boyd & Crawford, 2012; p.138).

8. EVALUATION OF RESOURCE RELIABILITY

Scientific relevance: high. Academic publisher of reference (Routledge), identified peer review, Open Access CC-BY, mobilisation of classic literature (Greene 1989; Johnson & Onwuegbuzie 2004; boyd & Crawford 2012) and recent (Bakker 2019; sources 2023-2024). Transparency point to note : the authors openly declare having used ChatGPT as a "mediating voice" in the editing and harmonisation of the text (p.6). This disclosure is a mark of honesty, but invites the reader to maintain a critical perspective on a synthesis partially assisted by AI. Some citations in the text appear as "et al." with a lightened critical apparatus.

Operational relevance: moderate for research/evaluation, low for direct intervention. It is a doctoral reference resource, not a field toolkit. Directly mobilisable elements: decision-making framework of mixed methods, CAQDAS orientation, digital ethics checklist, triangulation by "joint displays", pre-registration. Adaptation to the field of health promotion remains the responsibility of the reader.

10. STRATEGIC HASHTAGS

#healthpractices #MixedMethods #ParticipatoryResearch #HealthEvaluation #DigitalData #ResearchEthics #ScientificLiteracy #AIinResearch