Sampling and research design: defining the population and selecting a sample, recognizing sampling and design choices that affect validity and reliability, and designing the inquiry (variables, controls, instruments) so the data can actually support the conclusion.

What this topic is asking

Designing the inquiry means deciding who or what you will study and how the study is structured so that the data you gather can actually support your conclusion. Two ideas sit at the center: sampling (choosing your participants or cases from a population) and research design (the variables, controls, and instruments that shape the data). Choices here directly affect how much you can trust and generalize your findings, so they must be deliberate and defensible. This page shows how to make them.

Population and sample

Your population is the whole group your question is about; your sample is the subset you actually study. You almost never study an entire population, so you choose a sample and then reason about how far its results apply. The selection method matters:

• Random sampling gives every member an equal chance and supports generalization, but is often impractical at this level.
• Convenience sampling uses who is available; easy, but the sample may not represent the population.
• Purposive sampling deliberately selects cases that fit the inquiry, common and appropriate in qualitative work.

Validity and reliability

Two qualities tell you whether a design can be trusted:

• Validity asks whether you are actually measuring or capturing what you intend, and whether your conclusions follow from your data.
• Reliability asks whether your method would yield consistent results if repeated.

Design choices affect both: a leading interview question threatens validity; an inconsistently applied coding scheme threatens reliability. You will not eliminate every threat, but you should recognize and address the ones your design creates.

Designing so the data answers the question

A good design connects every element back to the question. Define your variables or focus, decide whether you need a comparison or control, and choose instruments (surveys, interview guides, coding schemes) that produce data capable of answering what you asked. Pilot or check your instruments where you can, because a flawed instrument quietly corrupts everything downstream.

Why this matters for the paper and defense

The strength of your conclusions is bounded by your design, and markers judge whether the data you gathered can actually support what you claim. A paper that over-claims from a small convenience sample loses credit; one that draws careful conclusions and acknowledges its limits gains it. In the oral defense, expect questions about why you chose your sample and how you would improve the design, so you must understand the trade-offs you made.

Try this

Q1. Define validity and reliability in one sentence each. [Recall]

• Cue. Validity is whether you are measuring or capturing what you intend and whether conclusions follow from the data; reliability is whether the method would give consistent results if repeated.

Q2. A student studies 25 friends through a survey and concludes their result holds for all teenagers nationally. Identify the design flaw and how to fix the claim. [Short explanation]

• Cue. The convenience sample of 25 friends is too small and unrepresentative to generalize to all teenagers, threatening external validity; the fix is to limit the conclusion to the studied group and acknowledge the sampling limit, rather than claiming national generalisability.