We are seldom interested in statistics for their
own sake. Instead, we want to use statistics such as means and standard
deviations to answer substantive questions such as:
- Is a certain combination of drugs effective in treating AIDS?
- How should we select students
so that they will have the best chance at succeeding in our college?
- How much will an
advertising campaign increase sales?
- Are environmentalists more likely to vote for Democrats
or for Republicans?
- Is a webbook like Seeing Statistics an effective
way to teach statistics?
In the preceding chapters we have developed the tools we need to answer
such questions. In particular, knowledge of probability and sampling distributions
will allow us to infer, for example, whether the typical patient outcome for the AIDS
group receiving the treatment is better than the typical patient outcome for the AIDS
group not receiving the treatment. Also, we are able to state our confidence in
estimates such as the amount that an advertising campaign will increase sales.
Inference and confidence turn statistics from mere numbers into answers to
important questions.
