From 6d8ed4f5c7c85c8367923ac1b4bff961111f2f90 Mon Sep 17 00:00:00 2001 From: Joan Concha Date: Thu, 10 Sep 2020 04:58:25 +0100 Subject: [PATCH] lab finished --- Lab_answers.md | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) create mode 100644 Lab_answers.md diff --git a/Lab_answers.md b/Lab_answers.md new file mode 100644 index 0000000..443befe --- /dev/null +++ b/Lab_answers.md @@ -0,0 +1,23 @@ +## Challenges + +### Challenge 1: What is the difference between expected value and mean? + +The mean is total sum of observations, divided by the total number of observations, as simple as that. +In other hand, the expected value is the average value we expect to get when we repeat an experiment a lot of times. + +### Challenge 2: What is the "problem" in science with p-values? + +There are two problems in my opinion: the first one is that the p-value is misinterpretated a lot of times, for example, let's see this +example: url = https://blog.minitab.com/blog/adventures-in-statistics-2/how-to-correctly-interpret-p-values + +For example, suppose that a vaccine study produced a P value of 0.04. This P value indicates that if the vaccine had no effect, you’d obtain the observed difference or more in 4% of studies due to random sampling error. The correct way to interpretate this is: + +Correct: Assuming that the vaccine had no effect, you’d obtain the observed difference or more in 4% of studies due to random sampling error. + +Incorrect: If you reject the null hypothesis, there’s a 4% chance that you’re making a mistake. + +The second problem is that, it is generally expected that a study should have a P < 0.05 as way to avoid considering 'statistically non-significant' outliers, that threshold can lead scientists to disregard data to keep within the P < 0.05, without considering whether the data is significant or not. + +### Challenge 3: Applying testing to a specific case: A/B testing. + +I chose the netflix article. On this article they try to demonstrate if changing the art of a picture they were able to increase the active time of the subscribers. I think that it would also be interesting to test different regions, and to point the main characteristics of the art with bigger results. \ No newline at end of file