Analysis - Adjusting Sample Weights
|
|
Prior to analyzing GRTS Design probability data, it is necessary to ensure that
each sampled site is assigned the correct weight. Commonly, weights are
adjusted following the completition of the Design Status File to reflect the
implementation of the design. Most often the design is not implemented exactly
as initially specified, commonly fewer or more sites, thus requiring the weight
adjustment.
The initial design weights are assigned during the GRTS design process and
individual site weights reflect the amount (stream lenght, estuarine area,
etc.) of the target population represented by the site. For example, if the
frame contains 1000 km of stream length and the design sample size is 100, then
each site has a design weight of 10. If during the evaluation and sampling
phase only the first 90 sites are used, then the initial design weights are
adjusted:
adjusted weight(i) = sum(design wts)/sum(used wts)xdesign weight(i)
These weights are then used when calculating estimates for the Target Population
that is included in the Sample Frame.
-
Estimates based on Target Sampled sites apply to the Sampled Population with no
additional assumptions.
-
Estimates based on Target Sampled sites can apply to the portion of Target
Population within the Sample Frame ONLY IF assume that the Access Denied,
Target Not-Sampled, etc., sites occurred randomly and independently of site
characteristics
-
Estimates for Target Population NOT ONLY require assumptions above BUT ALSO
that portions of Target Population that are not included in the Sample Frame
have same characteristics as the Sampled Population
Back to Analysis Specifics