Research methods guides

The five decisions that matter most — test choice, sample size, p-values, and effect size — and how they fit together.

A four-question decision guide and table, from t-tests to regression — with a free selector tool.

What it is, the five things it isn’t, and what to report instead of “p < 0.05” alone.

Effect size in plain English — benchmarks, why it matters, and how d, r, and odds ratios convert.

Why studies miss real effects, the four interlocking quantities, and how to run a power analysis.

The two jobs of statistics — summarising your sample vs generalising to a population — and when you need which.

Why a correlation never proves cause, the confounder and reverse-causation traps, and what causation takes.

What a CI really means, what makes it wider or narrower, and how it relates to the p-value.

One-sample, independent, paired — when to use each, the assumptions, and what the result tells you.

What it tests, why not to run lots of t-tests, one-way vs two-way, and why you need a post-hoc test.

The design decisions in order: methodology, design type, sampling, variables, and validity.

What each methodology is for, a side-by-side comparison, and how to choose from your question.

Experimental, cross-sectional, longitudinal, case study — what each answers and when to use it.

Probability vs non-probability sampling, the main methods, and how the choice affects generalizing.

IV vs DV, confounders and controls, and how to operationalize a variable so it can be measured.

Measuring the right thing vs measuring it consistently — the types of each and how to check them.

Methods are the techniques; methodology is the justification for why they fit your question and paradigm.

Positivism, interpretivism, pragmatism — their assumptions and how your paradigm shapes the whole design.

What a case is, the types, how to bound it, and how to answer the generalisability critique.

A snapshot vs the same subjects over time — what each tells you about change and causation.

The five-step process — scope, search, synthesize, find the gap, structure — in plain English.

Databases, keywords and synonyms, Boolean operators, and inclusion/exclusion criteria.

How to synthesize across sources instead of summarizing each — with a template.

Which type of review you need — purpose, rigor, and effort compared.

The types of gap, how to spot them, and how to turn one into a defensible question.

Organize by theme and argument — never author-by-author — with an outline template.

Descriptive vs critical annotations, what each should contain, and a worked example.

A practical appraisal checklist — design, sample, measures, analysis, claims, bias — for quant and qual papers.

How they differ in purpose, method, reproducibility, and effort — and when to use each.

Why it depends, rough ranges by degree, saturation, and why quality beats a raw count.

Analysis, coding, interviews, and rigor — how the pieces of a qualitative study fit together.

Braun & Clarke’s six phases, code vs theme, and reflexive vs coding-reliability TA.

Building theory from data — open/axial/selective coding, constant comparison, and the variants.

Inductive vs deductive coding, coding cycles, codebooks, and the software options.

Semi-structured interviews — open questions, probes, running it, and the ethics.

Credibility, transferability, dependability, confirmability — and where saturation fits.

When the interaction is the data — size and number, the moderator’s job, and group dynamics.

The systematic, sometimes-countable cousin of thematic analysis — and how the two differ.

The study of lived experience — descriptive vs interpretive (IPA), bracketing, and the analysis.

How qualitative sample size is justified — what saturation means, the types, and how to report it.

What makes a study genuinely mixed — choosing a design and actually integrating the two strands.

The four core designs — convergent, explanatory, exploratory, embedded — and how to choose from timing and priority.

Quantitative first, then qualitative to explain it — the two phases and how to pick the follow-up sample.

Qualitative first to explore and build an instrument, then quantitative to test it — and the make-or-break build step.

Both strands at once, analysed separately, then merged — how to merge and how to handle divergence.

The three integration strategies, how to build a joint display, and how to write meta-inferences.

The reproducible process end to end — protocol, eligibility, search, PRISMA, appraisal, synthesis.

Write and register your method before you start — what a protocol contains and why PRISMA-P/PROSPERO matter.

Turn your question into reproducible eligibility rules with PICO — and avoid the common mistakes.

The PRISMA 2020 boxes, how every number must reconcile, and how to fill it in — with a free generator.

RoB 2, ROBINS-I, Newcastle–Ottawa, CASP — and how GRADE rates the whole body of evidence.

Pooled effects, forest plots, fixed vs random effects, heterogeneity, and when not to pool.

The standard sections and the logic that ties them together — from problem to plan to budget.

The gap-to-significance structure, a fill-in template, and how it differs from a research question.

The one broad aim vs the specific, measurable objectives — and how to write each.

Null vs alternative, directional vs non-directional, and how to make a hypothesis testable.

What each is, how to build a conceptual framework from your variables, and how to draw one.

Specific Aims, Significance, Approach, reading the funder’s call, and a budget reviewers trust.

Who benefits and how — the “so what?” case, made concretely and without overclaiming.

What you cover, the boundaries you chose and why, and how they differ from limitations.

The established theory you use as a lens — and how it differs from a conceptual framework.

Direct vs indirect costs, the line items, and the justification reviewers actually read.

What scholarly writing is for, how a paper is built, and the skills that get it read and accepted.

Introduction, Methods, Results, Discussion — what each must do, and the hourglass shape.

The four moves of a structured abstract — written last, read first, often all a reviewer reads.

Clarity, concision, cohesion, and tone — the self-edit passes that make prose readable.

In-text vs reference list, when to cite, and how to avoid accidental plagiarism.

The point-by-point response letter that turns “major revisions” into “accepted.”

The funnel / CARS model — establish the territory, find the gap, occupy it with your aim.

Report, don’t interpret — ordering findings, tables and figures, tense, and the results/discussion line.

Restate the contribution, not a summary — answer the question, give implications, add no new data.

Why changing a few words is still plagiarism, how to genuinely paraphrase, and why you cite even when you do.

Choosing a method, building the instrument, and protecting data quality before you start.

Open vs closed questions, unbiased wording, ordering, length, and piloting.

Item vs scale, how many points, the neutral midpoint, and how to analyse the responses.

Nominal, ordinal, interval, ratio — and why the level decides which statistics are legal.

Internal consistency, what counts as a good value, and why reliability isn’t validity.

What a DMP covers, the FAIR principles, and why it’s written before collection.

Where AI helps across the research lifecycle, and the verify-and-disclose discipline that keeps it honest.

Where it helps, where it fails, and the verify-everything workflow that stops it inventing sources.

Role, context, task, constraints — the prompt patterns that get useful, checkable output.

Why AI invents plausible, non-existent references — and how to catch every one.

Why AI can’t be an author, what policies require, and how to write a disclosure statement.

Legitimate drafting and editing help — what’s acceptable, and how to stay your own author.

What it’s good at, what it’s dangerous at, a safe workflow, and the one verification rule.

How detectors really work, why they false-positive, and how to use AI safely and provably.

Code, don’t compute; AI-assisted qualitative coding; and the verification & confidentiality rules.

A map of the tools by job, why source-grounded beats a chatbot, and how to choose one.

The whole arc — deciding, supervisor, milestones, surviving the hard parts, and the viva.

The real differences in purpose, time, cost, and careers — and how to choose.

Fit over fame, how to assess a supervisor, the questions to ask, and running the relationship.

What each year looks like — proposal, upgrade, data, writing up — and how to stay on pace.

Imposter syndrome, motivation, isolation, and burnout — and what actually helps.

What examiners test, the questions to expect, the outcomes, and a preparation plan.

Structure and schedule beat genius — the chapters, the hard ones, and how to finish.

Why the two terms swap meaning between the US and UK — and which one your institution means.

The standard five-chapter shape, front and back matter, and the common variations.

Justify, don’t just describe — what to include and the methods-vs-methodology distinction.

Interpret, don’t restate — the structure, handling limitations, and not overclaiming.

Backward planning from your deadline, how long each phase takes, and building in buffer.