Why and how we sample

We cannot count everything, so we sample and infer. Sampling should be representative and avoid bias:

• Random sampling — positions chosen by chance (e.g. random coordinates);
• Systematic sampling — at regular intervals, often along a transect to study change across a gradient.

Estimating populations: the Lincoln index

For mobile animals we use mark–recapture. Capture, mark and release a first sample; later take a second sample and see how many are marked. The Lincoln index estimates population size:

N = (n₁ × n₂) ÷ m
n₁ = number marked first, n₂ = number in second sample, m = number of marked individuals recaptured.

Your turn — Lincoln index

Quadrats and abundance

For plants and slow-moving organisms we use quadrats to measure density (number per unit area), frequency or percentage cover. Taking several quadrats and averaging improves reliability.

Your turn — mean count

Your turn — total population estimate

Choosing a statistical test

• Spearman’s rank correlation tests for a relationship between two variables that can be ranked.
• Chi-square tests whether observed frequencies differ significantly from expected ones (association or goodness of fit).
• Standard deviation measures the spread of data around the mean.

Choosing a test

Correlation, reliability and validity

A correlation shows two variables change together, but correlation does not prove causation — a third factor may be involved. Good data are reliable (consistent when repeated) and valid (they actually measure what was intended).

Correlation and causation

Which kind of tool?

Sort each item into the correct category.

Term to meaning

Match each term to its meaning.

The big ideas to know

Choose the right test for your data, and interpret it with care. Press Finish.