PES6251 Chap.1 Research Logic and Design Families
Research Logic and Design Families
From question to interpretation
Research is systematic, question-led problem solving. The scientific sequence moves from a defined problem and literature-informed hypothesis through design, data collection, analysis and interpretation.
Research can be basic or applied, quantitative or qualitative, experimental or non-experimental, and primary or secondary at the same time because those labels classify different features. The key decision is to match the design to the claim rather than choose a technique first.
Classify the decision, not the topic
A sports-coaching topic does not dictate one method.
A practical club problem can generate applied research, numerical or narrative evidence, direct observation or analysis of existing records. Experimental status depends on whether the researcher manipulates an independent variable, not whether the project uses scores, takes place in a laboratory or includes comparison. State purpose, evidence type, manipulation and data origin separately.
This avoids the common error of treating labels as one menu and makes the eventual conclusion easier to delimit.
Audit the chain backwards
Write the strongest conclusion the project hopes to support and work backwards through the evidence required.
A causal conclusion needs an intervention contrast and a credible alternative; an association needs defined variables and confounder planning; a descriptive estimate needs a population, frame and denominator. Then travel forward again, checking that each procedure produces data used in the analysis.
An unconnected claim is overreach, while an unconnected measure consumes participant effort without answering the question.
Use limitations as design information
Every design exchanges one strength for another. Laboratory control may reduce sporting realism, while field observation preserves context but admits more competing explanations.
Qualitative evidence can explain experience deeply without estimating prevalence, and secondary data can cover large populations while limiting what was measured. Name the trade-off that matters for the actual question. A limitation is useful when it changes the claim, analysis or next study, not when it appears as a generic apology.
What this chapter covers
- 01
Scientific process from question to interpretation
- 02
Basic and applied purposes
- 03
Quantitative and qualitative evidence
- 04
Experimental and non-experimental control
- 05
Primary and secondary data
- 06
Operational definitions and claim boundaries
Worked example · free
Repair a method-first proposal
- 2Identify the design mismatch.
- 3Rewrite the question with measurable variables.
- 3State the largest defensible conclusion.
Key terms
- Applied Research
- Research designed to address an immediate practical problem while using a systematic method.
- Independent Variable
- The exposure or condition manipulated or compared as a possible explanation of outcome variation.
- Dependent Variable
- The measured response that a study is designed to explain or compare.
- Operational Definition
- A statement of how a concept will be observed, delivered, measured or scored in the study.
- Primary Research
- Research collecting original observations specifically for the question under investigation.
Research Logic and Design Families FAQ
Can one study belong to several research families?
Yes. The labels classify different aspects of a project. A study may be applied because it addresses a club problem, quantitative because it analyses scores, non-experimental because exposure is only observed, and primary because the researchers collect new data. State each label only when it helps explain what the design can support.
Does a field setting make a study observational?
No. Setting and manipulation are separate. A field experiment still assigns an intervention, while a laboratory observation may measure variables without changing them. Ask what the researcher controls, how participants enter comparison groups and when outcomes are measured before choosing the design label.
What is the fastest way to check a research chain?
Write the hoped-for conclusion first and trace backwards. A causal sentence needs an assigned intervention, credible comparator and outcome after treatment. An association needs defined measures and confounder planning. A description needs a clear population, sampling route and denominator. Any missing link exposes an overclaim or unnecessary measurement.
Assessment move
Classify three studies by purpose, evidence type, manipulation and data origin. Then write one permitted conclusion for each. Keep the labels on separate lines so you do not treat them as competing choices. For your proposal, underline every outcome claim and connect it to a measurement, timing decision and analysis step.
Add a backward audit: begin with the proposed conclusion, list the evidence it would require, and test whether the stated procedures actually create that evidence. If a procedure has no analytical destination, remove or justify it. If a conclusion has no supporting procedure, narrow it or redesign the study. End by naming the most important trade-off between control, realism, depth and generalisability.
Repeat the audit from a participant's perspective to detect burdens, unclear transitions and measures that the research question does not justify.
Working through Research Logic and Design Families in PES6251? Sia is AskSia’s AI Statistics tutor — ask any PES6251 Research Logic and Design Families question and get a clear, step-by-step explanation grounded in how PES6251 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
