Economics
Saved & recent

Sampling and measurement

A sample describes a target population only through its selection and measurement process. Nonresponse, missing observations and inconsistent definitions can bias results even with many observations.

Read first
Distributions, means and medians
Related
Sampling uncertainty and confidence intervals, Causal identification
Further reading
CORE Econ: Measuring management practices — Working in Excel
population 100 100 200 selected 40 40 80 responded 36 20 56 valid 34 18 52 group A group B
Hypothetical counts, not outcome values. Group A is violet and B blue throughout. Equal initial selection does not prevent differential nonresponse and invalid measurement from changing composition; missing values are not coded as zeros.

View in the full Economics collection