As discussed
in the previous chapter, when you undertake statistical analysis,
you generally wish to generate one or more representative numbers
(statistics, statistical coefficients and parameters) that tell you
something about the world and specifically about your research question.
The chapter case above gives an example, namely the findings of what
changes a drug like Crestor can bring about in patients.
Other examples of statistical
effects include:
-
You may wish to know the average
return on one type of financial instrument versus another.
-
You may generate a correlation
– as per Chapter 8 – between two variables.
-
You might assess change in a variable
from month to month.
There are two crucial
elements of almost any statistic. These are size and accuracy.
The following two sections explain and explore the difference.