Evaluation of reliability

'Descriptive statistics alone cannot prove a fieldwork hypothesis is correct.' Evaluate this statement, referring to a fieldwork enquiry you have studied or planned. [8]

Mark scheme (8 marks) — levelled:

• Level 1 (1–2): Simple agreement or disagreement, no example, vague generalisations ('statistics can be wrong').
• Level 2 (3–5): A balanced argument with a fieldwork example. Agree points: small sample size limits statistical power; correlation is not causation (Spearman's strong Rₛ doesn't prove a causal link); sampling bias may distort results; primary data captures only one moment in time. Disagree points: statistics provide objective, quantifiable evidence; significance testing (e.g. against critical values) does allow defensible conclusions; combined with secondary data, statistics strengthen inference.
• Level 3 (6–8): All of Level 2 PLUS a clear, justified judgement. E.g. 'Descriptive statistics are necessary but not sufficient. In my river study on the Glenarm, mean velocity differences supported the Bradshaw model, but I could only support — not prove — the hypothesis because n = 10, weather varied across sites, and confounding variables like channel substrate were not controlled. Reliability is strengthened by triangulation with secondary Environment Agency NI data, larger samples, and replication. Statistics are the evidence; geographical reasoning, theory and evaluation transform that evidence into a defensible conclusion.' Specific reference to confidence level, sampling strategy, or named geographical model elevates the response.