dear statalisters.
I created a person-period data set using prsnperd command to conduct a
discrete time hazard model and found a potential error in the data set
created. Here's the syntax I used:
. prsnperd childid minccage1 censored, t(72)
No variation in event occurence for period 22!
No variation in event occurence for period 32!
No variation in event occurence for period 66!
No variation in event occurence for period 68!
No variation in event occurence for period 72!
If I understand correctly, _period variable created by prsnperd should
have values ranging from 1 to 72 because minccage1, my length-to-event
variable in the model, has a maximum value of 72. And here's what I
found strange:
. tab _period if _period>72&_period<100, miss
_period | Freq. Percent Cum.
------------+-----------------------------------
73 | 12 3.86 3.86
74 | 12 3.86 7.72
75 | 12 3.86 11.58
76 | 12 3.86 15.43
77 | 12 3.86 19.29
78 | 12 3.86 23.15
79 | 12 3.86 27.01
80 | 12 3.86 30.87
81 | 12 3.86 34.73
82 | 12 3.86 38.59
*** I intentionally cut here to save the space...*******
94 | 11 3.54 82.32
95 | 11 3.54 85.85
96 | 11 3.54 89.39
97 | 11 3.54 92.93
98 | 11 3.54 96.46
99 | 11 3.54 100.00
------------+-----------------------------------
Total | 311 100.00
. tab _period if _period>1000, miss
_period | Freq. Percent Cum.
------------+-----------------------------------
1001 | 1 0.20 0.20
1002 | 1 0.20 0.40
1003 | 1 0.20 0.60
1004 | 1 0.20 0.81
1005 | 1 0.20 1.01
1006 | 1 0.20 1.21
1007 | 1 0.20 1.41
1008 | 1 0.20 1.61
1009 | 1 0.20 1.81
1010 | 1 0.20 2.02
1011 | 1 0.20 2.22
1012 | 1 0.20 2.42
I like to hear from you all regarding whether I missed something while
I use it, or it is a known glitch. Thanks.
Jinseok Kim, Ph.D.
Assistant Professor
College of Social Work
University of South Carolina
Columbia, SC 29208
Tel: 803-576-6082
Fax: 803-777-3498
e-mail: [email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
