Drawing Conclusions Beyond the Data
In their pre-study, the Mythbusters found a 3/7 split (statistic = 3 or 3/10 = 0.30). Sure that's not 50/50 but could it have happened by chance alone?
As in Lab 1, we want to make a decision about the value of the parameter.(b) If the mechanism does not have a "built-in bias," what value would that presume for the long-run probability of dropped toast landing top-side down?
(c) If the mechanism is biased towards NOT landing on the "top" side, what range of valuesdoes that imply for the long-run porbability of dropped toast landing top-side down?
Notice how these are statements about the unknown parameter for the process in general. The first statement is our "null model" or null hypothesis (by chance alone), and the second statement is often called the alternative hypothesis. Our goal is to decide which hypothesis we believe more.
As in Lab 1, we will assume the null model to be true, and see what types of values we find for the statistic when we know the null is true.
(d) Include a screen capture of your output, showing the approximate p-value, in your lab report.
(e) Do you consider this p-value small? Do you consider it convincing evidence that the long-run probability of landing top down is larger than 0.50?
