1. Summary of opportunity costs
The difference in observed proportions is not consistent with the null hypothesis.
2. Significance
Only seven of the 1000 permutations for a p-value of (point) 007 were smaller than the observed data value. Even if we had performed a more conservative two-sided test, the p-value would still have been equal to just 14 out of 1000, or (point) 014.
3. Causation
Because the p-value is substantially less than (point) 05, we conclude that it was not simply random variability that led to fewer students buying the DVD when being reminded to save.
Because the study was randomized, that is, the individuals were randomly assigned the choices, there was nothing systematically different about the participants in the treatment and control groups. The only difference in the two groups was the set of options they received.
Therefore, any difference in DVD buying rates is due to the options given, that is, being reminded to save. A causal inference can be made in this setting.
4. Random sample
Importantly, however, the 150 individuals in the sample were not randomly sampled from all people. Indeed, they were said to be students. Students are certainly different from the adult population in many ways.
In order to generalize to a larger population, we would need more information about the students and who they represented.
5. Let's practice!
OK, now that you have the hang of testing, let's move on to inferential estimation.