Sampling Strategies: Choosing the Right Approach

⚖️ Sampling Strategies: Choosing the Right Approach

A sampling strategy determines where and how often you collect data. Your sampling strategy must match your enquiry question — and you must be able to justify it in the exam. Using the wrong sampling strategy can lead to biased results that do not reliably test your hypothesis.

Strategy	How It Works	Best For	Limitation
Systematic sampling	Data collected at regular intervals — e.g., every 100 m downstream along the river	Detecting downstream patterns and trends; comparing sites fairly	May miss important features if they happen to fall between sampling points (e.g., a major tributary at 150 m)
Random sampling	Sites chosen using random numbers (e.g., from a random number table or app); every point has an equal chance of being chosen	Pebble sampling within a site (eliminates selection bias when choosing individual pebbles)	May produce a cluster of points in one part of the study area while leaving gaps elsewhere; can miss important features
Stratified sampling	The study area is divided into zones (e.g., upper, middle, lower course); a proportional number of samples is taken from each zone	Ensuring all parts of the river are represented even if the zones are different lengths	More complex to design; requires prior knowledge of the zones; time-consuming
Opportunistic/pragmatic sampling	Sites chosen based on access, safety, and available time — "wherever we could get to safely"	Realistic when access is limited or some sites are dangerous	Introduces bias — results may reflect accessible rather than representative locations; difficult to justify in a rigorous enquiry

For most GCSE river investigations: use systematic sampling for sites (e.g., every 100–200 m downstream, or at 5 specified points from source to mouth) combined with random sampling within each site for pebble collection. This is the combination most likely to earn full marks in an exam question about sampling justification.

How many sites? A minimum of 5 sites is needed to identify a trend with reasonable confidence. More sites (8–10) produce a more reliable picture but require more time. Your choice should be justified: "We used 7 sites to ensure the entire river from moorland source to urban lower course was sampled, allowing us to test the Bradshaw Model predictions across the full downstream gradient."