Whether, Why, and For Whom: Research Methods for Applied Computing

Context

Most applied computing practitioners who enter doctoral programs have deep methodological literacy in one narrow sense: they know how to design and evaluate software systems. What they typically lack is literacy in research paradigms — the philosophical frameworks that determine what kind of knowledge a study can produce and what warrants a claim.

The title encodes the three prior questions that every research design decision depends on:

• Whether — is an empirical approach warranted, or is this a design, evaluation, or theoretical contribution?
• Why — what paradigm is appropriate given the nature of the claim being made?
• For whom — whose problem does the research address, and what would count as a useful answer for them?

What the Paper Covers

Research Pipeline Model

A visual and structured model tracing the full path from research question formation to dissertation-grade contribution:

1. Question formation — from practitioner observation to researchable question
2. Paradigm selection — positivist, interpretivist, pragmatist, or critical, based on the nature of the claim
3. Design choice — quantitative, qualitative, or mixed-methods, based on paradigm and question
4. Data collection — instruments, sampling, ethical considerations
5. Analysis — procedures appropriate to the design and paradigm
6. Contribution — what the results warrant claiming, and for whom

The pipeline makes explicit the dependencies that are often implicit: design choice depends on paradigm, which depends on question, which depends on the practitioner’s ontological and epistemological commitments.

Mixed-Methods Design Framework

Six mixed-methods design types mapped to common applied computing research goals:

• Convergent parallel — quantitative and qualitative data collected simultaneously, compared and integrated at interpretation
• Explanatory sequential — quantitative results explain which cases or phenomena to investigate qualitatively
• Exploratory sequential — qualitative findings inform the design of a quantitative instrument or framework
• Embedded design — one strand is nested inside a dominant strand (e.g., interviews embedded in a controlled experiment)
• Transformative design — a theoretical lens (critical, feminist, disability) drives the overall design
• Multiphase design — multiple sequential studies, each building on prior findings, across a program of research

Each design type is evaluated against common applied computing research goals: system evaluation, practitioner study, artifact development, and comparative framework assessment.

Design Principle 1 — Question Before Method

The research question must be formed before the method is selected. Practitioners frequently select methods based on familiarity or availability rather than fit to the question — producing studies that answer a different question than the one that motivated the research. Question-first design prevents this failure mode.

Design Principle 2 — Paradigm Coherence

Ontological, epistemological, and methodological choices must be coherent with each other and with the research question. Mixing paradigm elements (e.g., positivist data collection with interpretivist analysis claims) produces inconsistencies that undermine the warranted claim the study can make.

Conceptual Comparison of Evaluation Stances

A structured qualitative analysis of four evaluation paradigms:

• Positivist: observable, measurable phenomena; generalizable findings; hypothesis testing
• Interpretivist: meaning-construction by participants; context-specific findings; thick description
• Pragmatist: what works for whom under what conditions; action-oriented; mixed-methods natural fit
• Critical: power structures, whose interests are served, emancipatory potential

The comparison replaces a prior matrix with descriptive analysis that acknowledges the paradigms are incommensurable on some dimensions and complementary on others.

Why It Matters (Portfolio Angle)

This monograph directly seeds my dissertation methods chapter. The research pipeline model structures how I am approaching the empirical components of my doctoral research in AI governance:

• Whether: empirical study of practitioner behavior and organizational governance is warranted — not a pure design contribution
• Why: pragmatist paradigm, because the research question is “what works, for which organizations, under which conditions” — not “what is universally true”
• For whom: AI governance practitioners and doctoral researchers who need structured methods guidance

The mixed-methods framework will directly inform how I integrate quantitative governance assessment data with qualitative case study evidence in the dissertation.

Palayil, A. B. (2026). Whether, Why, and For Whom: Research Methods for Applied Computing (Version 1.1) [Technical report]. Engineering-to-Research Monograph Series, Vol. 9. Zenodo. https://doi.org/10.5281/zenodo.20829174