![]() ![]() If we have a total of N samples to compare two land uses, then, in the absence of further information, the best design is to have N/2 in each of the two groups. ![]() But the design is not biased for the objective of comparing land uses. the mean number of beetles per m 2 ) without accounting for the design, then the result may be biased, as different land uses may not be represented in the sample with frequencies that are proportional to their occurrence in the study area. ![]() If the data were used to make statements about the overall study area (e.g. It is sometimes suggested that this approach is ‘biased’, as the land use classes to sample are determined a priori. Used in this sense, the strata are land areas under different uses, and the idea is to deliberately sample from each of these. One approach to improving the sampling design (relative to SRS or a single grid) is to use ‘stratification’ to ensure that we do indeed have adequate sample sizes of each land use. This requires comparison of different land uses. Consider the example of the objective of discovering and understanding land use effects on BGBD. In the introduction to this paper, I suggested that focussing on objectives of a study will increase the efficiency of the design.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |