Data Science Discussion

For your edification, we have made the

RShiny Schlitz Taste-test Simulation

, which simulates the “Schlitz ad” from this week’s Naked Statistics reading. The link should open in another tab.

https://ybergner.shinyapps.io/schlitz/

Use the following questions to guide your exploration (note: you are asked to respond to one of the questions below in a forum post, and others will be part of this week’s quiz):

1. Start with the default settings (which include sample size = 1). Press both the “Run Once” and “Run 100 Times” buttons. Did the one person you sampled using “Run Once” prefer Schlitz? What proportion of your 100 sampled people preferred Schlitz?
2. Now change the sample size to 10. Again, press both the “Run Once” and “Run 100 Times” buttons. In your single, “Run Once” experiment, how many people preferred Schlitz? Now look at your 100 runs of the experiment: In what proportion of experiments (remember that, in each experiment, 10 people are sampled), did at least 50% of people prefer Schlitz? In what proportion of experiments did at least 40% of people prefer Schlitz? (hint: to get this number, you will need to change the “Acceptable (Desirable) Outcome” slider.)
3. Question for the forum: Leave sample size = 10 (the other sliders can be at any values), and press the “Run Once” button. Look at the percentage of people (in this single run of the experiment) who preferred Schlitz. Now, press “Run Once” a few more times and observe this percentage each time. With these settings, is it possible for the percentage preferring Schlitz to be equal to 65%? Why or why not? (respond to this question in the forum link on the next page).
4. Now, choose some values for all sliders except sample size (you can choose whatever you like, but the probability of a participant preferring Schlitz should be somewhere between 30-70%). Now, start with sample size = 1, then increase sample size to each of the available options. For each sample size option, press “Run 100” and observe the resulting histogram. What changes about the histogram as you increase sample size? Can you explain your observations from this exercise?
5. Set sample size = 500, and slowly increase the probability of a participant preferring Schlitz (start at 10% and increase in increments of 20% until you get to 80%). Press “Run 100 Times” for each probability. As you increase the probability of a participant preferring Schlitz, how does the histogram change? Can you explain your observations from this exercise?

For your edification, we have made the RShiny Schlitz Taste-test Simulation, which
simulates the “Schlitz ad” from this week’s Naked Statistics reading. The link should
open in another tab.
https://ybergner.shinyapps.io/schlitz/
Use the following questions to guide your exploration (note: you are asked to
respond to one of the questions below in a forum post, and others will be part of this
week’s quiz):
1. Start with the default settings (which include sample size = 1). Press both the
“Run Once” and “Run 100 Times” buttons. Did the one person you sampled
using “Run Once” prefer Schlitz? What proportion of your 100 sampled people
preferred Schlitz?
2. Now change the sample size to 10. Again, press both the “Run Once” and “Run
100 Times” buttons. In your single, “Run Once” experiment, how many people
preferred Schlitz? Now look at your 100 runs of the experiment: In what
proportion of experiments (remember that, in each experiment, 10 people are
sampled), did at least 50% of people prefer Schlitz? In what proportion of
experiments did at least 40% of people prefer Schlitz? (hint: to get this number,
you will need to change the “Acceptable (Desirable) Outcome” slider.)
=
3. Question for the forum: Leave sample size = 10 (the other sliders can be at any
values), and press the “Run Once” button. Look at the percentage of people (in
this single run of the experiment) who preferred Schlitz. Now, press “Run Once”
a few more times and observe this percentage each time. With these settings,
is it possible for the percentage preferring Schlitz to be equal to 65%? Why or
why not? (respond to this question in the forum link on the next page).
4. Now, choose some values for all sliders except sample size (you can choose
whatever you like, but the probability of a participant preferring Schlitz should
be somewhere between 30-70%). Now, start with sample size = 1, then
increase sample size to each of the available options. For each sample size
option, press “Run 100” and observe the resulting histogram. What changes
about the histogram as you increase sample size? Can you explain your
observations from this exercise?
=
5. Set sample size = 500, and slowly increase the probability of a participant
preferring Schlitz (start at 10% and increase in increments of 20% until you get
to 80%). Press “Run 100 Times” for each probability. As you increase the
probability of a participant preferring Schlitz, how does the histogram change?
Can you explain your observations from this exercise?