Variables. The independent variable is the type of area (mown lawn or unmown meadow); the dependent variable is Simpson's index of diversity, D.
Random sampling. In each area, lay two tape measures at right angles as axes and use a random number generator to give coordinates for each quadrat, so that every point has an equal chance of being sampled and bias is avoided.
Quadrats. Use a frame quadrat of the same size in both areas (e.g. 0.5 m × 0.5 m) and sample at least 10 quadrats per area (or until the running mean number of species levels off).
Recording. Identify each plant species with an identification key and count the number of individuals (n) of each species in each quadrat; sum each species over all quadrats to give n, and all individuals to give N, for each area.
Processing. Calculate n/N for each species, square, sum and subtract from 1 to give D for each area; compare the two values (higher D = higher diversity, as D close to 1 indicates high diversity).
Standardised variables. Sample both areas on the same day in the same season and weather, with the same total area sampled; choose areas with similar soil type, slope and aspect.
Abiotic checks. Measure soil water content, soil pH and light intensity in each area to check that they are similar, so mowing is the main difference.
Replication. Repeat the whole survey on at least two further occasions (or in more lawn/meadow pairs) and calculate a mean D for each area type.
Risk. Low risk: wear gloves when handling plants (some may be irritant or stinging) and wash hands afterwards; avoid picking or trampling plants unnecessarily.