Evaluating a Land-Use Conflict with Incomplete Field Data
Students survey a 2 km stretch of coastal dune behind a growing town. They record vegetation cover at 10 quadrats: 15%, 20%, 18%, 60%, 22%, 25%, 19%, 70%, 24%, 21%. Two quadrats (60% and 70%) sit where a walking track and a car park edge the dune. The town council proposes a 300-space car park on the foredune, citing 'low vegetation cover and high visitor demand.' A conservation group cites the same survey to argue the dune is fragile and should be fenced. Both sides quote the mean of 29.4%.
1. Analyse the survey data to show why quoting the mean of 29.4% is misleading for both sides, and recommend a better summary.
Read the worked answer
The mean of 29.4% is pulled upward by two outliers (60% and 70%) that coincide with disturbed edges. Excluding them, the remaining eight quadrats average about 20.5%, a more representative figure for the intact dune. The median is 22.5%, sitting between the 22% and 25% values, and is far less sensitive to the two high readings. The range of 15–70% (55 percentage points) shows extreme variability that a single mean conceals.
So the council's 'low vegetation cover' claim is broadly supported for the intact dune (~20%), but its use of 29.4% actually overstates cover, weakening its own case. The conservation group's 'fragile' claim is supported by the low, sparse cover on most quadrats, but the high values at the track and car park edge show that disturbance raises local cover, likely through weeds and colonising species, not healthy dune vegetation.
A better summary reports the median, IQR and range, and separates disturbed-edge quadrats from intact ones. This gives both sides an honest basis: most of the dune has roughly 20% cover with high variability, and edges behave differently.
2. Identify the sampling limitation in the survey and explain how it undermines both sides' arguments.
Read the worked answer
The key limitation is sample size and placement: only 10 quadrats cover 2 km, and the two disturbed quadrats are not randomly placed but sit at the track and car park edge. This introduces bias: the sample over-represents accessible, disturbed ground and under-represents the interior dune. With n = 10, a single quadrat shifts the mean by several percentage points, so the 29.4% figure is unstable.
This undermines the council because its 'low cover' claim rests on a mean inflated by edge quadrats; a larger, stratified random sample might show either lower or higher cover. It undermines the conservation group because its 'fragile dune' claim generalises from a small, non-random sample and cannot establish that the whole 2 km is uniformly fragile.
A defensible recommendation is to stratify sampling (foredune, mid-dune, hind-dune, disturbed edges), increase quadrat number, and record species composition, not just cover, before any 300-space car park decision is made. Neither side's current evidence supports a firm conclusion.