Ecological sampling: estimates and causation

Total count = 16, so mean count = 16/8 = 2 plants per quadrat. Density = 2/0.50 = 4 plants m⁻². Estimated field total = 4 × 2000 = 8000 plants, assuming the sample adequately represents the defined field.

• Sum 16 plants and divide by 8 to obtain 2 plants per quadrat.
• Divide by 0.50 m² to obtain 4 plants m⁻².
• Multiply by 2000 m² to estimate 8000 plants.

Common confusion: Multiplying the per-quadrat mean directly by field area ignores the quadrat’s area.

Higher moisture readings are generally associated with larger counts: the 15% and 18% quadrats have no plants, while those at 32% and 35% contain four and five. The pattern is not perfectly proportional; for example, the counts at 22% and 25% are both one.

• Identify a generally positive association in these observations.
• Compare zero counts at 15% and 18% with four or five plants at 32% and 35%.
• Recognise that the pattern is not a perfectly proportional relationship.

Common confusion: A description of association should not automatically assert that moisture caused the differences.

The observations show an association, but moisture was not independently manipulated. Light, nutrients or disturbance could covary with moisture and affect abundance. A controlled comparison varying moisture while making other relevant factors comparable would better test its causal effect. The field table alone does not predict the size of that effect.

• Recognise that the data are observational rather than a manipulated-moisture experiment.
• Identify a plausible covarying factor such as light, soil nutrients or disturbance.
• Explain why a controlled comparison is needed to isolate moisture’s effect.

Common confusion: Naming a statistical association or collecting more observational quadrats does not remove every confounding factor.

Random placement gives field positions a defined chance of inclusion and reduces deliberate choice of plant-rich patches. Enough independent samples are needed to assess spatial variation. A slope pattern instead calls for position-based sampling along a justified transect, measuring both abundance and relevant environmental conditions.

• Explain reduced deliberate placement bias in the defined field.
• Connect sufficient independent quadrats with a representative mean and variation.
• Suggest a defined transect with recorded positions and environmental measurements for the slope pattern.

Common confusion: The sampling design should match the question; a transect and a whole-field random sample answer different questions.

Increase the number of independently placed random quadrats and report the distribution of counts or a justified uncertainty estimate. This better assesses spatial variation and the precision of the estimated mean. It does not make every field location identical or turn 8000 into an exact census.

• Propose more independent random quadrats across the sampling area.
• Explain improved representation of spatial variability.
• Report spread or an appropriate uncertainty estimate rather than claiming the total becomes exact.

Common confusion: Repeatedly recounting the same quadrat is not equivalent to adding new independent locations.