st: Hotdeck problem

[Joelle M Anderson <anders35@uwm.edu>]

For my thesis, I am using the hotdeck program to impute values for 
missing cases in my income variable. Currently, I am trying to hotdeck 
my income variable (176 missing) using 3 variables (age=27 missing; 
education=13 missing; gender=0 missing; although with 9 overlapping missing
values the combination of these three variables only has 31 missing 
cases total). Yet when Stata creates my new, hotdecked income variable, 
there are an additional 19 missing cases that I can't account for 
(missing=50). Does anyone know why this might be? Another strange thing 
is that, when I try to rename my hotdecked income measure before merging 
it with my full dataset, all 50 missing cases remain missing after 
merging; when I do not rename my hotdecked income
measure before merging, only 42 missing cases remain missing after 
merging. I have pasted my Stata output below. Any help would be greatly 
appreciated!

Joelle Anderson
Graduate Student, Sociology
University of Wisconsin-Milwaukee
anders35@uwm.edu

//First hotdeck imputation, renaming the income variable BEFORE merging 
with full dataset

. hotdeck incomeR using IncomeHD, store by(education sex ageR) keep(resp 
incomeR)
DELETING all matrices....

Table of the Missing data patterns
* signifies missing and - is not missing

Varlist order: incomeR

   pattern |      Freq.     Percent        Cum.
------------+-----------------------------------
         * |        176       11.72       11.72
         - |      1,326       88.28      100.00
------------+-----------------------------------
     Total |      1,502      100.00
WARNING: When the <command> option is not selected
then no analysis is performed on the imputed datasets


. use "C:\data\IncomeHD1.dta", clear

. tab incomeR

 RECODE of |
    income |
  (income. |
last year, |
that is in |
2004, what |
  was your |
total famil |      Freq.     Percent        Cum.
------------+-----------------------------------
         1 |        103        7.09        7.09
         2 |        164       11.29       18.39
         3 |        222       15.29       33.68
         4 |        148       10.19       43.87
         5 |        162       11.16       55.03
         6 |        265       18.25       73.28
         7 |        178       12.26       85.54
         8 |        124        8.54       94.08
         9 |         86        5.92      100.00
------------+-----------------------------------
     Total |      1,452      100.00

. rename incomeR incomez

. merge resp using "C:\Documents and Settings\anders35\My 
Documents\Thesis_3_29.dta", unique sort

. tab incomez

 RECODE of |
    income |
  (income. |
last year, |
that is in |
2004, what |
  was your |
total famil |      Freq.     Percent        Cum.
------------+-----------------------------------
         1 |        103        7.09        7.09
         2 |        164       11.29       18.39
         3 |        222       15.29       33.68
         4 |        148       10.19       43.87
         5 |        162       11.16       55.03
         6 |        265       18.25       73.28
         7 |        178       12.26       85.54
         8 |        124        8.54       94.08
         9 |         86        5.92      100.00
------------+-----------------------------------
     Total |      1,452      100.00

//Second hotdeck imputation, renaming the hotdecked income variable 
AFTER merging with full dataset

. hotdeck incomeR using IncomeHotD, store by(education sex ageR) 
keep(resp incomeR)
DELETING all matrices....

Table of the Missing data patterns
* signifies missing and - is not missing

Varlist order: incomeR

   pattern |      Freq.     Percent        Cum.
------------+-----------------------------------
         * |        176       11.72       11.72
         - |      1,326       88.28      100.00
------------+-----------------------------------
     Total |      1,502      100.00
WARNING: When the <command> option is not selected
then no analysis is performed on the imputed datasets

. clear

. use "C:\data\IncomeHotD1.dta", clear

. tab incomeR

 RECODE of |
    income |
  (income. |
last year, |
that is in |
2004, what |
  was your |
total famil |      Freq.     Percent        Cum.
------------+-----------------------------------
         1 |         98        6.75        6.75
         2 |        162       11.16       17.91
         3 |        220       15.15       33.06
         4 |        153       10.54       43.60
         5 |        159       10.95       54.55
         6 |        267       18.39       72.93
         7 |        178       12.26       85.19
         8 |        126        8.68       93.87
         9 |         89        6.13      100.00
------------+-----------------------------------
     Total |      1,452      100.00

. merge resp using "C:\Documents and Settings\anders35\My 
Documents\Thesis_3_29.dta", unique sort

. rename incomeR incomey

. tab incomey

 RECODE of |
    income |
  (income. |
last year, |
that is in |
2004, what |
  was your |
total famil |      Freq.     Percent        Cum.
------------+-----------------------------------
         1 |         98        6.71        6.71
         2 |        164       11.23       17.95
         3 |        222       15.21       33.15
         4 |        153       10.48       43.63
         5 |        159       10.89       54.52
         6 |        270       18.49       73.01
         7 |        178       12.19       85.21
         8 |        126        8.63       93.84
         9 |         90        6.16      100.00
------------+-----------------------------------
     Total |      1,460      100.00

*
*   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/