Here is my problem: I have monthly income data and have used the "xtile" command to calculate the 5% quantiles
xtile income_decile=bbh5101, nq(20)
How can I find out which borders Stata used to allocate the observation to a certain quantile bin, e.g. first quantile bin from 0 to 800€, second quantile bin from 801 to 1600€ and so on?
I believe you just want the percentiles. Use the corresponding _pctile command. For example:
clear all
set more off
sysuse auto
xtile q = weight, nq(10)
_pctile weight, nq(10)
sort weight
list weight q
return list
Checking those two lists, should be useful. See also the Methods and formulas section in [D] pctile.
The result:
. list weight q
+-------------+
| weight q |
|-------------|
1. | 1,760 1 |
2. | 1,800 1 |
3. | 1,800 1 |
4. | 1,830 1 |
5. | 1,930 1 |
|-------------|
6. | 1,980 1 |
7. | 1,990 1 |
8. | 2,020 1 |
9. | 2,040 2 |
10. | 2,050 2 |
|-------------|
11. | 2,070 2 |
12. | 2,110 2 |
13. | 2,120 2 |
14. | 2,130 2 |
15. | 2,160 2 |
|-------------|
16. | 2,200 3 |
17. | 2,200 3 |
18. | 2,230 3 |
19. | 2,240 3 |
20. | 2,280 3 |
|-------------|
21. | 2,370 3 |
22. | 2,410 3 |
23. | 2,520 3 |
24. | 2,580 4 |
25. | 2,640 4 |
|-------------|
26. | 2,650 4 |
27. | 2,650 4 |
28. | 2,670 4 |
29. | 2,690 4 |
30. | 2,730 4 |
|-------------|
31. | 2,750 5 |
32. | 2,750 5 |
33. | 2,830 5 |
34. | 2,830 5 |
35. | 2,930 5 |
|-------------|
36. | 3,170 5 |
37. | 3,180 5 |
38. | 3,200 6 |
39. | 3,210 6 |
40. | 3,220 6 |
|-------------|
41. | 3,250 6 |
42. | 3,260 6 |
43. | 3,280 6 |
44. | 3,300 6 |
45. | 3,310 6 |
|-------------|
46. | 3,330 7 |
47. | 3,350 7 |
48. | 3,370 7 |
49. | 3,370 7 |
50. | 3,400 7 |
|-------------|
51. | 3,420 7 |
52. | 3,420 7 |
53. | 3,430 8 |
54. | 3,470 8 |
55. | 3,600 8 |
|-------------|
56. | 3,600 8 |
57. | 3,670 8 |
58. | 3,690 8 |
59. | 3,690 8 |
60. | 3,700 8 |
|-------------|
61. | 3,720 9 |
62. | 3,740 9 |
63. | 3,830 9 |
64. | 3,880 9 |
65. | 3,900 9 |
|-------------|
66. | 4,030 9 |
67. | 4,060 9 |
68. | 4,060 9 |
69. | 4,080 10 |
70. | 4,130 10 |
|-------------|
71. | 4,290 10 |
72. | 4,330 10 |
73. | 4,720 10 |
74. | 4,840 10 |
+-------------+
.
. return list
scalars:
r(r1) = 2020
r(r2) = 2160
r(r3) = 2520
r(r4) = 2730
r(r5) = 3190
r(r6) = 3310
r(r7) = 3420
r(r8) = 3700
r(r9) = 4060
You can put the percentiles in a variable. Just use:
pctile p = weight, nq(10)