For the study with the headline "coffee Gives Jolt to Lifespan, over 400,000 adults aged 50 to 71 were studied for 13 years as part of the NIH-AARP Diet and Health Study. They completed a questionnaire asking about 124 dietary items in addition to demographic and lifestyle variables. Results were reported separately for men and women and for differing amounts of coffee consumption. The hazard ratio for death (somewhat similar to a relative risk) for men who drank 2 or 3 cups of coffee a day was estimated to be 0.88, with a 95% confidence interval from 0.84 to 0.93, indicating that they were less likely to die during the study than men who drank no coffee. Explain how each of the following numbered items from the Ten Guiding Principles on would apply to this study. Guiding principle 2. Guid

Guiding principle 2 states that cause-and-effect relationships generally cannot be inferred from observational studies, but only from randomized experiments. This research was an observational study, so we cannot say for sure that there exists a cause-and-effect relationship between the two variables.
Guiding principle 6 states that very large samples are likely to result in statistical significance, even if there isn't any practical importance to the research, and that having a confidence interval to accompany the results of significance tests is a good idea.
The sample size in this research was extremely large, so it had no problem reaching statistical significance, even though the actual magnitude of the relationship is insignificant. The confidence interval for the hazard ratio of death tells us about that magnitude (it is significant).
Guiding principle 8 states that we should examine the consequences of Type l and Type 2 errors before stating the level of significance.
Remember that a Type l error occurs when the null hypothesis is true, but we mistakenly reject it. In this example, a Type l error would be made if coffee consumption doesn't increase lifespan, but we conclude that it does. The consequence is that we would probably drink large amounts ofcoffee believing it would reduce the risk of death, whereas in fact it has no effect.
Recall that a Type 2 error occurs when the alternative hypothesis is true, but we mistakenly conclude that we cannot reject the null hypothesis. In this case, a Type 2 error would be made if drinking coffee increased a person's lifespan, but we are assured that it didn't. The consequence is that we wouldn't consider drinking more coffee even though it would reduce the risk of death.
We can agree that a Type 2 error sounds more serioUS, so we should consider using higher level of significance, such as \alpha = 0.10, for instance (to make it easier to reject the null hypothesis).
Guiding principle 9 states that studies which perform multiple hypotheses testing could have one or more statistically significant results simply by chance.
Notice that the participants were asked about 124 dietary items (in addition to lifestyle and demographic variables). So, there is a great chance that at least one Type l error was made, and it might just have been the result that linked coffee consumption with longer lifespan.
Result:
a) This was an observational study, so we cannot infer cause-and-effect relationship; b) The sample size was large so it easily reached statistical significance, but the accompanied confidence interval speaks of practical significance as well; c) Type 2 error is more serious than Type 1 error, so use higher level of significance when doing hypotheses testing; d) Many significance tests were done, so this one could have been a false positive