ENV330H5 Lecture Notes - Lecture 3: Random Number Table, 2D Computer Graphics, Stratified Sampling
Document Summary
Much stronger inference can be drawn about the site differences if we have controlled for other variables at the sites, we have confidence that differences among sites are due to the lead in the soil. Occurs when measures are treated as individual replicates but remain dependent on a common variable that makes them more likely to be similar than true replicates. Because chicks share nests, they are more likely to become co-infected with other chicks. Parasites per chick is not independent of nest. In this study, that would reduce the number of subjects in each test group from. Not much power to test for differences among parks, but we can look at within- nest and among-nest variability. Be smart in allocating your sampling effort. Probability-based designs: each sample has a known probability of selection this doesn"t mean that every sample has the same probability of being selected (e. g. , stratified sampling design)
