Methodology Builder is the research methodology generator in the AI for Research Thesis Suite. You give it a research question and your constraints; it searches the literature and drafts a methodology plan with its choices cited to studies. It is for students and researchers who need a sound design for a thesis or proposal before collecting data.
What you give it and what you get
The form has three fields:
- Research question (required), for example "What explains low uptake of solar irrigation among smallholders?"
- Approach: Suggest one for me, Quantitative, Qualitative, or Mixed methods.
- Context (optional): population, setting, time, budget and the data you can get.
The plan comes back in eight parts, as headings or as flowing paragraphs if you switch "With headings" off:
| Part | What it covers |
|---|---|
| Research design and why | The design for your question and the reason it fits better than the alternatives |
| Population and sampling | Who or what you study, how you select them, and a sample size rationale |
| Data collection and instruments | Surveys, interviews, records or measures, and how you collect them |
| Variables or concepts | What each variable or concept is and how it is measured or observed |
| Analysis plan | Named statistical tests or an analytic approach, such as thematic analysis |
| Validity and reliability | Or trustworthiness for qualitative work: how you make the findings credible |
| Ethics | Consent, risk, data protection points your ethics application must cover |
| Limitations | What the design cannot show, stated before an examiner does |
Methodological sources and similar studies are cited where they support a choice, as numbered markers linked to each paper.
If you start from a Deep Research result (its "Build on this" links, or a task reopened from History), the plan builds on that result: it reuses the result's sources instead of running a new search and keeps to its findings. This is the quickest route when you have already researched what is known about your question and now need to decide how to study it.
How it works, step by step
- Open the Thesis Suite and choose Methodology builder.
- Enter your research question. If you do not have one yet, run the Research question module first (FINER, PICO, PICOT or SPIDER).
- Pick an approach, and add your context: the more specific, the more usable the plan.
- Choose sources and filters. Under "Your own sources" you can attach a methods textbook chapter, your supervisor's notes or papers from your Library.
- Set your role and citation style (APA 7, MLA 9, Chicago or Harvard). Pro users can switch on High quality.
- Press Run (3 credits). The plan streams in, with the cited papers in a panel beside it.
- Press Open in editor to turn it into your chapter with the citations and reference list in place, or carry it on under "Build on this".
A worked example
Example (invented for illustration): a master's student in agricultural economics enters "What explains low uptake of solar irrigation among smallholders?", sets Approach to "Suggest one for me", and adds the context "two districts in northern Ghana, six months of fieldwork, a budget for about 300 household surveys, access to a farmer cooperative".
A plan for this question might propose an explanatory sequential mixed-methods design: a household survey first, analyzed with a model of adoption against cost, credit access and land tenure, then interviews with a subset of non-adopters to explain the patterns. It would justify the survey sample, describe how adoption and each factor are measured, name the analysis, cover consent for interviews and the anonymity of survey data, and flag limits such as self-reported income. The student would then check each cited study and replace the general choices with her real districts and cooperative.
The rest of the Thesis Suite
Methodology Builder sits with the other planning modules. Each costs 3 credits unless shown.
| Module | What it does | Credits |
|---|---|---|
| Research question | Refines a topic into one question checked against FINER, PICO, PICOT or SPIDER | 3 |
| Hypotheses | Three to five testable hypotheses with variables, null hypothesis and how to test each | 3 |
| Theoretical framework | Candidate theories, the one to use, key constructs and how variables relate | 3 |
| Methodology builder | The plan described on this page | 3 |
| Thesis outline | Chapters, sections and a word budget per chapter | 3 |
| Research timeline | Phases, milestones and risks over 6 to 48 months | 3 |
| Similar papers | Papers like yours from a title or abstract | 0 |
What it does not do
- It does not calculate a sample size. Use the Sample size tab in Data Lab for t-tests and two proportions, or software such as G*Power for other designs.
- It does not know your ethics board's rules or your department's chapter template.
- It runs one search on your question, so it cites up to about a dozen papers, mostly similar studies. Add classic methods texts yourself.
- It does not collect or analyze data. It plans; you do the study.
- It can suggest a design that is not feasible for you. Your context field and your supervisor are the checks.
Who uses it
Master's students planning a dissertation (see AI for master's students), PhD candidates preparing a proposal or confirmation review (see AI for PhD students), and undergraduates designing a final-year project (see AI for college students). The plan also feeds the Methods section of a proposal in Proposal Studio.
Learn more
- How to write a methodology chapter
- Theoretical framework vs conceptual framework, with examples
- How to write a research question
- How to write a hypothesis
Last updated 2026-11-12