Hotdeck Imputation - Mander's module

[Diana Rossi]

Dear Stata Users,

I'm trying to use hot deck imputation to fit my mlogit model, but I can't work out how I should choose the argument of the parameter option

parms(varlist)

.

I've understand I cannot use the name of the outcome variable as this is clearly stated in the instructions:

"The estimated coefficients are put automatically by STATA into the matrix e(b), note the column headings are the parameter names that hotdeck uses. So you can not use the simple syntax of just parms(male) because this refers to two parameters." from http://fmwww.bc.edu/repec/bocode/h/hotdeck.html

However, to put numbers does not improve the situation either:

hotdeck care sexr oold estold sexpm PKD PSY totill hhadults ch chinhhALL eqtotinc_CPI18 edqual318 using htd, command (mlogit care sexr oold estold survprt sexpm  ADLlow ADLmedium  HeartC STR CLD CAN Dem PKD PSY totill hhadults ch chinhhALL eqtotinc_CPI18 edqual318) parms(1:2) store infiles(htd1 htd2 htd3 htd4 htd5)

Output:

Seed is set as 697384976
DELETING all matrices....
Attempted to extract 1:2 from e(b)

e(b)[1,84]
2: 2: 2: 2: 2: 2:
sexr oold estold survprt sexpm ADLlow
y1 .60388364 -.47986207 -.02133574 .08003106 -.17934151 -.10411431

2: 2: 2: 2: 2: 2:
ADLmedium HeartC STR CLD CAN Dem
y1 -.19624277 1.4683839 .20741606 .87230941 .50988057 .71984507

2: 2: 2: 2: 2: 2:
PKD PSY totill hhadults ch chinhhALL
y1 -30.589216 .14339682 -.38178285 -1.0966315 -.27700942 -.06422221

2: 2: 2: 3: 3: 3:
eqtotinc_CPI18 edqual318 _cons sexr oold estold
y1 -.00059397 .44474437 -.50076011 .53585904 .01953826 .54606489

3: 3: 3: 3: 3: 3:
survprt sexpm ADLlow ADLmedium HeartC STR
y1 .38804055 -.52440217 -1.4041877 -.72715985 .9990522 -.00726879

3: 3: 3: 3: 3: 3:
CLD CAN Dem PKD PSY totill
y1 .29380069 .20838243 .42931336 1.9444675 .67458425 -.22455736

3: 3: 3: 3: 3: 3:
hhadults ch chinhhALL eqtotinc_CPI18 edqual318 _cons
y1 -.42962679 -.09104878 -.44191831 -.00006068 .25832145 -.14729462

4: 4: 4: 4: 4: 4:
sexr oold estold survprt sexpm ADLlow
y1 .04941033 .70977285 1.4551533 -.24615277 -.05765219 -2.663845

4: 4: 4: 4: 4: 4:
ADLmedium HeartC STR CLD CAN Dem
y1 -1.5498775 .93627375 .41815352 .50168593 -.01009619 1.793868

4: 4: 4: 4: 4: 4:
PKD PSY totill hhadults ch chinhhALL
y1 2.5680556 .08771486 -.13541796 -1.6689431 -.36185486 .96320093

4: 4: 4: 5: 5: 5:
eqtotinc_CPI18 edqual318 _cons sexr oold estold
y1 .00068153 .20444907 .80968691 .36825866 -.03937672 -.92423027

5: 5: 5: 5: 5: 5:
survprt sexpm ADLlow ADLmedium HeartC STR
y1 -1.5842798 -.03762929 -1.3724876 -.99879307 -1.4595781 -.92070584

5: 5: 5: 5: 5: 5:
CLD CAN Dem PKD PSY totill
y1 -1.2906747 -.41663967 -.62433776 -31.007363 -1.3616278 .19237522

5: 5: 5: 5: 5: 5:
hhadults ch chinhhALL eqtotinc_CPI18 edqual318 _cons
y1 -.29617103 -.06239098 -.11138008 .00136503 .3125066 -.7754682
Check the matrix of estimates and only include column names in the parameters NOT variable names
r(111);


Can someone explain me what I should do and why?

Thank you very much,
Diana Rossi

[Diana Rossi]

P.S. I tried

parms(2:2)

or

parms(3:2)

or

parms(4)

and similar others, but I always get the error message.