Statistical Methods - Text Mining Techniques for Healthcare Provider Quality Determination - page 78

Figure 2. Sample mean with sample=5
Once we have computed the means, we can graph them using kernel density estimation (a smoothed
histogram). We show the difference between the distribution of the population, and the distribution of
the sample mean for the differing sample sizes. Figures 2-5 show the distribution of the sample mean
compared to the distribution of the population for differing sample sizes. To compute the distribution of
the sample mean, we collect 100 different samples using the above code. We compute the mean for the
patient length of stay using the National Inpatient Sample.
In Figure 2, the sample mean peaks slightly to the right of the peak of the population distribution; this
peak is much more exaggerated in Figure 3. The reason for this shift in the peak is because the sample
Figure 3. Sample mean with sample=30