Information Technology Reference
In-Depth Information
Figure 4. Sample mean with sample=100
mean is susceptible to the influence of outliers, and the population is very skewed. Because it is so
skewed, the distribution of the sample mean is not entirely normal. As the sample increases to 100 and
then to 1000, this shift from the population peak to the sample peak becomes much more exaggerated.
We use the same sample sizes for 1000 replicates (Figures 6-9).
It is again noticeable that the sample mean is shifted away from the peak value of the population
distribution because of the skewed distribution. However, the distribution of the mean is not normally
distributed. In other words, the sample converges on a value that is much greater than the actual popula-
tion peak.
Figure 5. Sample mean with sample=1000
Search WWH ::




Custom Search