Please use D2L to turn in both the PDF or Word output and your R Markdown file in.
Use the following script to create your data for this lab
playoff.free.throws <- tbl_df(data.frame(
player = c('James','Curry','Irving','Durant','Wall','Leonard','Beal','Harden','Love','Horford','Aldridge','Green','Paul','Olynyk','Parker','Favors','Jordan','Oladipo'),
position = c('F','G','G','F','G','G','G','G','F','F','F','F','G','F','G','F','F','G'),
FTA = c(162,114,84,103,93,102,61,115,75,29,55,67,33,30,14,23,56,6),
FTM = c(113,103,76,92,78,95,50,101,63,22,42,46,29,22,14,11,22,6)))
playoff.free.throws <- playoff.free.throws %>% arrange(position, player)
playoff.free.throws
## # A tibble: 18 x 4
## player position FTA FTM
## <fct> <fct> <dbl> <dbl>
## 1 Aldridge F 55.0 42.0
## 2 Durant F 103 92.0
## 3 Favors F 23.0 11.0
## 4 Green F 67.0 46.0
## 5 Horford F 29.0 22.0
## 6 James F 162 113
## 7 Jordan F 56.0 22.0
## 8 Love F 75.0 63.0
## 9 Olynyk F 30.0 22.0
## 10 Beal G 61.0 50.0
## 11 Curry G 114 103
## 12 Harden G 115 101
## 13 Irving G 84.0 76.0
## 14 Leonard G 102 95.0
## 15 Oladipo G 6.00 6.00
## 16 Parker G 14.0 14.0
## 17 Paul G 33.0 29.0
## 18 Wall G 93.0 78.0
Fit independent models using a beta prior for \(\theta\) for the players Curry, Beal, and Oladipo. Describe the sampling model and priors for these models.
Compare the posterior HDI from these models with those found using the hierarchical models. Note the HPDinterval function can be applied directly to samples from a beta distribution as HPDinterval(mcmc(data = rbeta(n = 1000, shape1 = a.star, shape2 = b.star)))
,
Reflect on the differences/similarities in the credible intervals and the posterior mean in each situation. If you were going to bet on the players shooting percentages for the next season, which would you prefer?