1. The researcher is concerned about making a Type I error (which caused by incorrect rejection of a true null hypothesis), which concludes that there are differences between the placebo and medication groups when these are really due merely to chance.
2. In order to decrease the likelihood of type I error, the researcher could reduce her probability (alpha) leve to .01 or even .001.
Other method that she could do even though it's not popular is reducing her sample size.
3. The same method to solve type 1 errors would not work in other studies. As the possibility of type 1 error increased, the possibility of type 2 error will be decreased (for example type 2 error could be reduced by by increasing the power of your test)