Compute the assignmentMath Talk 9. Why use a p-value anyway?

Summary

The video introduces p-values by imagining that the presenter has to choose whether to cycle down a busy road or along the back streets. The presenter also has a data set where he had recorded the journey times on a few occasions. He analyzes the data set, and then also a larger one after taking more data, to obtain p-values, but in both cases ignores the conclusions, taking the main road when the p-value is high (because he had to rush home from an appointment), but the back streets when the p-value is low (because it is more pleasant to be on the back streets). The point of the video is to emphasize that what p-values do is to test hypotheses, but our decisions are not determined by the results of hypothesis tests.

1. In the video, the scientist focuses pretty much exclusively on the p-value. Is this really what scientists do? Should they? This can form the basis for discussing the fact that many scientists and other non-statisticians who work with statisticians often seem to think that statistics is exclusively about p-values and that statisticians are simply machines for producing p-values.

2. The statistician ended up ignoring the p-value and focusing on how long it would take to get home on average following each route. This reflects two different sides of statistics. What are they called?

3. Why bother with inference at all? Why not just focus on estimation?