4. (2) In the race to develop a commercial COVID19 vaccine, your company has just nished the Phase III trial, i.e, an RCT with a large sample size to test whether the vaccine provides protection against infection compared to a placebo. As the data analyst, you must analyze the trial data and test the null hypothesis that the vaccine provides no additional protection against infection compared to a placebo and report it to the company's shareholders and to the Food and Drug Administration (FDA) for approval. You run a regression of whether a subject contracted the illness or not on whether they received the vaccine or the placebo. The CEO (your boss) has been hinting that you use the old-fashioned standard error, probably because it makes it more likely to reject the null hypothesis than the robust standard error. But she is not a statistician/econometrician. You also know that if you are caught doing anything unethical you could get into trouble with the FDA, jeopardizing your career. You decide to go by what is statistically justified. Which standard error would you use for statistical inference? How would you justify your decision if asked to testify before a panel of experts at the FDA in charge of approving the vaccine