—MIND LAB—
—MIND LAB—
Answer eight questions about your project. Watch a defensible methodology assemble itself — design, notation, instruments, sampling, analysis, and a week-by-week flow diagram you can paste straight into your report.
A methodology is not a menu. Every answer closes some doors and opens others.
Work top to bottom. The specification panel fills in as you go, and the flow diagram appears once you have named a design and a track. Nothing here blocks you — if you build a weak design, the tool will tell you exactly how a reviewer would attack it.
Twenty-two worked projects. Load one, see its finished methodology, then edit it into your own.
Each card is a complete gate configuration. Loading it fills the builder and jumps you to the output, so you can read a finished framework before you attempt your own. Filter by track or by domain.
Every design buys you one kind of claim and forbids another.
Grouped as in Prof. Mandal's taxonomy: descriptive, associational, experimental, time-structured, and interpretive families, plus the non-traditional paradigms. Notation follows Campbell and Stanley — R randomisation, O observation or measurement, X treatment.
Five ways to get data out of people. Each one changes the rest of your project.
The track you pick decides your kit, your consent form, your sample size, your analysis, and roughly two-thirds of your timeline. Read the failure mode for each before you commit — it is almost always the thing that sinks a semester project.
Do not write your own scale. Somebody has already validated one.
A borrowed scale comes with a reliability coefficient, a validation sample, and a citation — three things a reviewer will ask for and you cannot manufacture in a semester. Use the library below, then declare exactly what you used in a Fig. 3.2 style table.
Filter by construct. Every entry lists item count, response format, published internal consistency, and whether you need written permission. Verify the current version and licence before you administer anything.
Four thousand and sixty-three published instruments, indexed from the Psychological Scales & Instruments Database. Search by name or filter by construct; each result opens the database entry, where you will find the items and the source paper. Reliability and licensing are not in this index — check them on the entry page before you commit to a measure.
How many, and how you got them, decides what you are allowed to generalise to.
Sampling is where student projects quietly lose their claims. Convenience sampling is acceptable in a semester project — pretending it was random is not.
Your analysis is not a choice. It is a consequence of four things you already decided.
Number of groups, whether the same people appear in every condition, the measurement level of your outcome, and your sample size. Walk the tree; the leaf you land on is your test. This is Prof. Mandal's plan-of-study chart, made walkable.
Name the confounds yourself, before your examiner does.
Quantitative work defends validity; qualitative work defends trustworthiness. They are different vocabularies for the same obligation — showing that your finding is about the world and not about your procedure.
Consent is a conversation you can prove happened.
Voluntary participation, informed consent, debriefing, and sharing results back — the four obligations in the methods lecture — plus data protection, which now does most of the work in practice.