API tools faq
paste
Guest User

Untitled

a guest
Oct 21st, 2018
81
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 5.53 KB | None | 0 0
raw download clone embed print report
  1. library (dplyr)
  2. library (tidyr)
  3.  
  4. testdata <- read.csv("https://stats.idre.ucla.edu/stat/data/binary.csv")
  5. testdata <- testdata %>% mutate (student.n0 = seq (1:400), location = sample(c(0,1), replace=TRUE, size=400), year = sample(c(0,1), replace=TRUE, size=400))
  6.  
  7. testdata$gre <- as.factor (case_when(
  8. testdata$gre < 520 ~ 1,
  9. testdata$gre > 519 & testdata$gre < 587 ~ 2,
  10. testdata$gre > 586 & testdata$gre < 660 ~ 3,
  11. testdata$gre > 659 ~ 4
  12. ))
  13.  
  14. testdata$gpa <- as.factor (case_when(
  15. testdata$gpa < 2.5 ~ 1,
  16. testdata$gpa > 2.4 & testdata$gpa < 3 ~ 2,
  17. testdata$gpa > 2.9 & testdata$gpa < 3.5 ~ 3,
  18. testdata$gpa > 3.4 ~ 4
  19. ))
  20.  
  21. testdata$admit <- as.factor(testdata$admit)
  22. testdata$location <- as.factor(testdata$location)
  23. testdata$year <- as.factor(testdata$year)
  24.  
  25. head (testdata)
  26. admit gre gpa rank student.n0 location year
  27. 1 0 1 4 3 1 1 0
  28. 2 1 4 4 3 2 0 1
  29. 3 1 4 4 1 3 1 0
  30. 4 1 3 3 4 4 0 0
  31. 5 0 2 2 4 5 0 0
  32. 6 1 4 3 2 6 0 0
  33.  
  34. #For location 0:
  35. testdata.0 <- testdata %>% filter (location == 0)
  36. #populate proportion table
  37. prop.admit.0 <- (round(prop.table(table(testdata.0$admit, testdata.0$year), margin = 2), 4)*100) %>% as.data.frame() %>% spread (key = Var2, value = Freq) %>% rename(Type = Var1) %>% add_column("Variable" = 'admit', .before = 1)
  38. prop.gre.0 <- (round(prop.table(table(testdata.0$gre, testdata.0$year), margin = 2), 4)*100) %>% as.data.frame() %>% spread (key = Var2, value = Freq) %>% rename(Type = Var1) %>% add_column("Variable" = 'gre', .before = 1)
  39. prop.gpa.0 <- (round(prop.table(table(testdata.0$gpa, testdata.0$year), margin = 2), 4)*100) %>% as.data.frame() %>% spread (key = Var2, value = Freq) %>% rename(Type = Var1) %>% add_column("Variable" = 'gpa', .before = 1)
  40. prop.rank.0 <- (round(prop.table(table(testdata.0$rank, testdata.0$year), margin = 2), 4)*100) %>% as.data.frame() %>% spread (key = Var2, value = Freq) %>% rename(Type = Var1) %>% add_column("Variable" = 'rank', .before = 1)
  41.  
  42. sb.prop.0 <- rbind (prop.admit.0, prop.gre.0, prop.gpa.0, prop.rank.0) %>% mutate(Variable_Type = paste(Variable, Type, sep = '_'))
  43.  
  44. sb.prop.0 <- sb.prop.0 %>% select(Variable, Type, Variable_Type, loc0_Pre_year = '0', loc0_Post_year = '1')
  45.  
  46. sb.prop.0 <- sb.prop.0 %>% mutate (loc0_DID = loc0_Post_year - loc0_Pre_year)
  47.  
  48. #For location 1:
  49. testdata.1 <- testdata %>% filter (location == 1)
  50. #populate proportion table
  51. prop.admit.1 <- (round(prop.table(table(testdata.1$admit, testdata.1$year), margin = 2), 4)*100) %>% as.data.frame() %>% spread (key = Var2, value = Freq) %>% rename(Type = Var1) %>% add_column("Variable" = 'admit', .before = 1)
  52. prop.gre.1 <- (round(prop.table(table(testdata.1$gre, testdata.1$year), margin = 2), 4)*100) %>% as.data.frame() %>% spread (key = Var2, value = Freq) %>% rename(Type = Var1) %>% add_column("Variable" = 'gre', .before = 1)
  53. prop.gpa.1 <- (round(prop.table(table(testdata.1$gpa, testdata.1$year), margin = 2), 4)*100) %>% as.data.frame() %>% spread (key = Var2, value = Freq) %>% rename(Type = Var1) %>% add_column("Variable" = 'gpa', .before = 1)
  54. prop.rank.1 <- (round(prop.table(table(testdata.1$rank, testdata.1$year), margin = 2), 4)*100) %>% as.data.frame() %>% spread (key = Var2, value = Freq) %>% rename(Type = Var1) %>% add_column("Variable" = 'rank', .before = 1)
  55.  
  56. sb.prop.1 <- rbind (prop.admit.1, prop.gre.1, prop.gpa.1, prop.rank.1) %>% mutate(Variable_Type = paste(Variable, Type, sep = '_'))
  57.  
  58. sb.prop.1 <- sb.prop.1 %>% select(Variable, Type, Variable_Type, loc1_Pre_year = '0', loc1_Post_year = '1')
  59.  
  60. sb.prop.1 <- sb.prop.1 %>% mutate (loc1_DID = loc1_Post_year - loc1_Pre_year)
  61.  
  62. #combine the DID tables
  63. combined.prop <- left_join(sb.prop.0, sb.prop.1, by = c('Variable', 'Type', 'Variable_Type'))
  64.  
  65. combined.prop <- combined.prop %>% mutate (DID = loc1_DID - loc0_DID)
  66.  
  67. Variable Type Variable_Type loc0_Pre_year loc0_Post_year loc0_DID loc1_Pre_year loc1_Post_year loc1_DID DID
  68. 1 admit 0 admit_0 71 64.52 -6.48 69.90 67.31 -2.59 3.89
  69. 2 admit 1 admit_1 29 35.48 6.48 30.10 32.69 2.59 -3.89
  70. 3 gre 1 gre_1 23 24.73 1.73 29.13 22.12 -7.01 -8.74
  71. 4 gre 2 gre_2 30 24.73 -5.27 27.18 22.12 -5.06 0.21
  72. 5 gre 3 gre_3 21 18.28 -2.72 17.48 17.31 -0.17 2.55
  73. 6 gre 4 gre_4 26 32.26 6.26 26.21 38.46 12.25 5.99
  74. 7 gpa 1 gpa_1 0 1.08 1.08 0.97 1.92 0.95 -0.13
  75. 8 gpa 2 gpa_2 17 20.43 3.43 11.65 14.42 2.77 -0.66
  76. 9 gpa 3 gpa_3 38 45.16 7.16 46.60 43.27 -3.33 -10.49
  77. 10 gpa 4 gpa_4 45 33.33 -11.67 40.78 40.38 -0.40 11.27
  78. 11 rank 1 rank_1 16 15.05 -0.95 11.65 18.27 6.62 7.57
  79. 12 rank 2 rank_2 34 41.94 7.94 31.07 44.23 13.16 5.22
  80. 13 rank 3 rank_3 29 22.58 -6.42 39.81 28.85 -10.96 -4.54
  81. 14 rank 4 rank_4 21 20.43 -0.57 17.48 8.65 -8.83 -8.26
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!