1: You
will get this particular SAS output by choosing the PROC MEANS program
(e.g. see “Code06a Continuous descriptives”) and using
the keywords “N MEAN STD STDERR PROBT LCLM UCLM.”
[return]
2: For instance, the bootstrap generally
cannot find confidence intervals for the maximum or minimum. Also,
if a sample is very small (say < 25 observations) bootstrapping
may not work for many techniques.
[return]
3: This
is not always true: sometimes you have all the data on the complete
population (what we would call a census), for instance you may only
be interested in studying your own employees and you have all required
data on all your employees. In this case, there is not really a “broader
population” in the short term. But, when your employee pool
changes due to turnover and new hires, it becomes a new population,
so your statistics on the current employee pool could be seen as a
sample of the broader employee population over time.
[return]