A GKM Legacy Holdings company · contracting in Maryland, USA · delivering from Accra, Ghana
InsightsWhere we workFrançais
Strategic Advisory
& AI
Book a consultation
Looking up along the concrete ceiling of a wide hall. One recessed line of light comes in from the left and divides at a single point into several lines, one of them yellow; below, a curved wall of glass looks on to the sea.

Research methodology decision aid

See which designs fit your question, and why.

Try the decision aid
What the film shows, in words
  1. Which designs fit your question?
  2. Five kinds of question. Eight designs.
  3. What is associated with what: 1 design to consider.
  4. Interviews within reach: 2 designs.
  5. Cause and effect, and you cannot assign: 1 design.
  6. Each design says what it assumes, and what you must learn.

ExampleWorked example: a question about what is associated with loan repayment. One answer changes at each step. Nothing is saved or sent.

Watch a closed branch open.

Step 1 of 5

A question about association: one design to consider.

1 design to consider: Cross-sectional survey or analysis of existing data. The branch to “Explanatory mixed methods” is closed: it needs interviews within reach.

The junctionFive kinds of question, two points, eight designs, read off the tool’s own rules.

Your question asks

  • How much or how many
  • What is associated with what (the example’s question)
  • Whether something caused an effect
  • How or why people experience something
  • What works, for whom, in what conditions

Interviews within reach

Not yet

Can you decide who gets an intervention?

No (the example’s answer)Yes

Used only for cause-and-effect questions

Data you can reach

  • Existing dataset: yes
  • Survey: yes
  • Interviews: no
  • Observation: no
  • Documents: no

Realistic number of participants

250

  2. A design to consider

    Cross-sectional survey or analysis of existing data

    Why
    Measures several factors across many cases.
    Assumes
    Associations are not causes; confounders must be measured.
    You learn
    Correlation, regression, handling missing data.
7 other designs are closed
  • Cross-sectional surveyClosed because the question asks what is associated with what. It opens when the question asks how much or how many.
  • Explanatory mixed methodsClosed because no interviews are within reach. It opens when interviews are within reach.
  • Randomised controlled trialClosed because the question asks what is associated with what. It opens when the question asks whether something caused an effect, and you can decide who gets the intervention.
  • Quasi-experimental design (comparison group or before-and-after)Closed because the question asks what is associated with what. It opens when the question asks whether something caused an effect.
  • Qualitative interviews (thematic analysis)Closed because the question asks what is associated with what. It opens when the question asks how or why people experience something.
  • Case studyClosed because the question asks what is associated with what. It opens when the question asks how or why people experience something.
  • Realist evaluation or mixed methodsClosed because the question asks what is associated with what. It opens when the question asks what works, for whom, in what conditions.

1 design to consider7 of 8 closed. None is ranked.

No design is automatically valid. Your supervisor and ethics committee judge whether it fits your question.

Now set your own points.

Answer what your question asks, what data you can reach and whether you can influence who gets what. The aid explains which designs fit and why, the assumptions each makes, and what analysis you will need to learn to do yourself.

The drawing is the form. The five lines are the question: choose one. Set the switch on the cause-and-effect line, then tick the data you can reach.

Live

The junctionFive kinds of question, two points, eight designs, read off the tool’s own rules.

Your question asks

Interviews within reach

Can you decide who gets an intervention?

Used only for cause-and-effect questions

Data you can reach

Interviews and Observation are the two that change what the aid shows.

  2. A design to consider

    Cross-sectional survey or analysis of existing data

    Why
    Measures several factors across many cases.
    Assumes
    Associations are not causes; confounders must be measured.
    You learn
    Correlation, regression, handling missing data.
7 other designs are closed
  • Cross-sectional surveyClosed because the question asks what is associated with what. It opens when the question asks how much or how many.
  • Explanatory mixed methodsClosed because no interviews are within reach. It opens when interviews are within reach.
  • Randomised controlled trialClosed because the question asks what is associated with what. It opens when the question asks whether something caused an effect, and you can decide who gets the intervention.
  • Quasi-experimental design (comparison group or before-and-after)Closed because the question asks what is associated with what. It opens when the question asks whether something caused an effect.
  • Qualitative interviews (thematic analysis)Closed because the question asks what is associated with what. It opens when the question asks how or why people experience something.
  • Case studyClosed because the question asks what is associated with what. It opens when the question asks how or why people experience something.
  • Realist evaluation or mixed methodsClosed because the question asks what is associated with what. It opens when the question asks what works, for whom, in what conditions.

1 design to consider7 of 8 closed. None is ranked.

No design is automatically valid. Your supervisor and ethics committee judge whether it fits your question.

1designs that fit
250participants planned

Designs to consider

  1. 01
    Cross-sectional survey or analysis of existing dataWhy: Measures several factors across many cases. Assumes: Associations are not causes; confounders must be measured.

Your analysis learning plan

  • Cross-sectional survey or analysis of existing dataLearn: Correlation, regression, handling missing data.

Things to check

  • Note: No design is automatically valid.Your supervisor and ethics committee judge whether it fits your question.

The researcher writes every word. This tool gives questions, structure and checks, not assessed text.

Your methods planDesign options explained, plus a learning plan for the analysis you will do yourselfTry this form

Practice 11: Research supervision, academic coaching & publication supportFree · instant results on this pageTool owner: Prof. Wisdom Gagakuma and Michael GagakumaVersion 1.0 ·

Your route, as it stands

Question asks
What is associated
Designs
1
Participants
250

Designs explained, not chosen for you. It moves with every answer you change above.

Your methods plan

Design options explained, plus a learning plan for the analysis you will do yourself, by email once delivery is connected.

For each design that fits your answers: why it fits, what it assumes and what you would need to learn. The analysis, and every word, stay yours.

Preview only. No email will be sent. Your results are available above.

Nothing you enter here is sent anywhere. Privacy notice

· a free brief, or book the Professor · from USD 150
Close under the place where two channels in a concrete ceiling part: one runs on as a line of yellow light, the other is cut the same way but stays dark.
An answer opens one way and shuts another. The shut one is still there to be seen.

How it works, and its limits.

The rules are published so you can check them. The aid explains designs; your supervisor and your ethics committee judge the fit.

Method

  1. 01

    Designs are matched to the type of question, whether you can assign an intervention, and the data you can reach.

  2. 02

    Each design lists its core assumptions and the analysis skills it needs.

  3. 03

    Sample sizes below 100 are flagged for quantitative questions.

What it cannot tell you

  • It explains options; it does not choose or approve a design.
  • Discipline norms vary; your supervisor has the final word.
  • Analysis is learned and done by the researcher.

Quality ruleTeach assumptions; no automatic claim of a valid design.

Prof. Wisdom Gagakuma Book the Professor · from USD 150
Michael Gagakuma Book Michael · from USD 90

When the choice needs a person

Methods support and teaching, not supervision. Nothing is written or analysed for you: the design, the analysis and every word stay the researcher’s, and your supervisor and ethics committee decide.

Research methods & statistics supportService · 11 Research supervision, academic coaching & publication supportResearch-methods bootcamps & supervisor trainingService · 11 Research supervision, academic coaching & publication support Choose a research method by question, evidence and constraintsGuide 11 Research supervision, academic coaching & publication supportPractice Research topicsFree toolResearch-proposal structure & review checklistFree toolReference-list consistency checkerFree toolJournal finderFree tool

What needs to move forward?

Tell us the decision, the challenge or the opportunity. We reply with a scoped approach, a named principal and a fee before any work begins.

Tell us what you need Book the Professor · from USD 150

No charge to submit an enquiry.