A valid survey measures the variable it is intended to measure and produces data suitable for answering the research question. In IB Mathematics: Applications and Interpretation HL, this requires an appropriate sample, precise questions, unbiased wording, and well-designed response options.
This is HL-only content from AHL 4.12: Data collection, reliability and validity tests.
Designing the Survey
Begin by defining the target population and the variable being measured. Then choose an appropriate sampling method: for example, a stratified sample may be suitable when important groups within the population must be represented proportionally.
Each question should measure one clearly defined idea. Avoid vague terms such as “often” unless a specific time frame is supplied, and avoid asking two things in one question.
| Design feature | Valid approach | Problem to avoid |
|---|---|---|
| Wording | Neutral and precise | Leading or emotionally loaded language |
| Focus | One issue per question | Double-barrelled questions |
| Response choices | Exhaustive and mutually exclusive | Overlapping ranges or missing options |
| Sample | Representative of the population | Convenience sampling that creates bias |
| Privacy | Anonymous where appropriate | Questions encouraging dishonest responses |
For example, “Don’t you agree that school lunches are too expensive?” is leading. A more valid version is: “How would you rate the price of school lunches?” with balanced choices from “very inexpensive” to “very expensive”.
A pilot survey should be conducted with a small group before full data collection. This can reveal ambiguous wording, missing response categories, or inconsistent interpretations.
A common misconception is that reliability guarantees validity. Reliability means a method gives consistent results; validity means it measures what it claims to measure. A consistently biased questionnaire may be reliable but invalid.
Exam Technique
For a Paper 1, Paper 2, or HL Paper 3 response, identify the specific flaw and explain its effect on the data. Do not merely write “the question is biased”; quote or describe the problematic wording, explain how it may influence responses, and propose a precise correction.
