Differences between revisions 1 and 9 (spanning 8 versions)
Revision 1 as of 2020-10-22 18:21:22
Size: 1725
Comment:
Revision 9 as of 2025-11-03 01:52:44
Size: 0
Comment:
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
= Survey Sampling =

<<TableOfContents>>

----



== Sample Type ==

=== Propability sampling ===

All members of a population have a non-zero chance to be contacted in a survey instrument. Traditional statistics rely on this assumption.

Examples:
 * Administrative surveys
 * Surveys with random recruitment (as by random digit dialing)


=== Non-probability sampling ===

Some members of a population are certain to be contacted or not be contacted.

Examples:
 * Panel surveys
 * River surveys (i.e. surveys with open recruitment, as by banner ads)

----



== Survey Allocation ==

Allocation is the distribution of sample size across domains.


=== Designing Domains ===

The key considerations are:

 * are some splits more important to others?
   * if studying military recruitment, then sex/gender is a strong split
 * what splits will be used for reporting?
 + what is the expected response rate?
   * if too few responses are expected from a domain, then splits should be reconsidered
 * what is the desired margin of error?


=== Stratified Allocation ===

Stratification is the process of dividing the population into discrete stratum, and then sampling from the strata.

Within a stratified sample, allocation can be designed as:

 * equal from each stratum
 * proportional to the size of each stratum
 * an optimization of a key measure's margin of error against cost
   * if cost is assumed fixed per unit, it is a Neyman allocation
   * if cost is assumed variable, it is an optimal allocation
   * note that, if all measures are equally varied, proportional allocation is essentially the same as a Neyman allocation



----
CategoryRicottone