Question

Which of Silver's assumptions is most important for their giving Trump a higher chance of winning than the other mainstream forecasters? (Just a sentence or two will suffice.)

Nate Silver's election forecasting model uses t distributions instead of normal distributions. What does this change about the types of estimates they will generate? (In this case, Silver says it doesn't matter much.)

f you have a sample mean of 15.6, a standard error of 1.7, and a t score of 2.78, what is the confidence limit? Show your work.

Answer 1

Using t distributions instead of Normal distribution means the estimated probabilities for extreme scenarios will be higher than given by normal distribution. Also, the probabilities for average cases will be higher when we use t distribution than when we use normal distribution. Hence, t distribution will give higher probabilities for average and extreme outcomes for the elections.

Confidence limits for the given sample are given by:

+/- T-score*StdError

= 15.6 +/- 2.78*1.7

= (10.876, 20.326)