#1 Paired study design can show differences between two specific data points. This can be good for before and after weights of individual people when studying weight loss#2 refer to attachmetn#3 Big box sampling puts all the data together as one population, therefore destroying the structure of the groups. When doing a confidence interval we must maintain the structure of the data so we can determine the difference. We can use big box sampling when it is safe to assume the medians of the populations are similar; in that csae we can use NHST. Big box sampling implies there will be no difference between the two groups because we are putting them together so doing a confidnece inerval wont be productive. We cant use big box sampling to calculate confidence intervals becasue the groups shouldnt be similar enough to group together#4 We use multiple testing correction when trying to find an alpha for a datasett. Before making lots of 2-group comparisons, must do 'omnibus testing'to ask is ther significant difference between all of the groups, then can proceed to group by group comparison. lowering alpha will decrease the false positve rate#5 Even though both type 1 and 2 errors can be harmful type two can be more problematic in most stiuations becasue a false negative will give the indication that everything is ok even though someones may need serious medical attention#6 The plunger plot on the left does not indicate the sample size or the test used, it doesnt show the skew of the data, and it is unclear what the bars and stars mean. The graph on the right is good becasue it clearly shows that the data is symetrical, it indicates the confidenceinterval, and we can determine the sample size.#7 no you could not used paired testing on the data becasue we are not looking at the difference or growth of specific individuals in a group, we are not looking at "repeated measures". You could use paired testing for pretest/posttest for example the weight loss of women in group that just started the qeto diet.
fromIPython.displayimportImage# To display figures in notebookImage(filename='python1.png',width=300)