As we have discovered, when we do a
study we get a representative point-estimate statistic. The real question
behind accuracy is how confident we can be that this statistic represents
the true, actual population statistic (assuming the sample is a good
representation of the population).
Therefore, accuracy
is all about how well our sample statistic reflects the true population
statistic. Imagine that you could know the true population value (you
cannot of course, but imagine this). Simplistically, the true population
statistic might be characterized as existing or not existing in reality.
The question is whether your statistical test can adequately reflect
what exists in reality.
This ...