ABSTRACT
This chapter reviews the problem of collecting appropriate and defensible data for successful data analysis. In choosing the sampling method, the sample's capacity for resolving targeted variations and estimating desired parameters within tolerable error limits and other sufficiency issues must be considered. It also may address biasing problems like operator reading and other systematic errors in the sample. Data collected over a uniform grid have a short-wavelength or Nyquist resolution limit of two station intervals or three consecutive data points. Data collected over a random distribution of samples are always submitted for electronic analysis in an array format gridded either implicitly by the computer or explicitly by the investigator. Custer sampling, and adaptive cluster sampling when combined with random sampling, are efficient to measure and map small wavelength features and unbiased means of the population.
