Unable to reshape multiply imputed data

[roxy kallsen]

Hello,

I am unable to reshape my data into long format to run a fixed effects model. As can be seen in the STATA command I imputed variables with missing data and later converted the data into flong. However, When I tried to reshape the data to long by indicating the specific variables for i and j I received the following error message.

set seed 832016
mi set wide 
mi register imputed INF_M6a INTSCORE_M6 PWAS_M_1 PWAS_M_2 FAM_M3 FAM_M6 OYAS_FINAL_C 
mi impute chained (nbreg) INF_M6a (poisson) INTSCORE_M6 (regress) PWAS_M_1 PWAS_M_2 (logit) FAM_M3 FAM_M6(ologit) OYAS_FINAL_C = INF_M3a INTSCORE_M3 PWAS_M_0 VIC_M3a AGE_AD MILES LENGTH2 dummy_race, add(25) force dots augment
mi convert flong
mi reshape long INF BAI BDI INTSCORE PWAS MILES AGE_AD FAM VIC OYAS LENGTH EXIT RACE, i(ID) j(time)

Thank you for your assistance with this matter, [HTML]/HTML]

[roxy kallsen]

Just following up on my last post. I noticed that the table did not copy in the delimiters so here is my second try.

[HTML]

. mi reshape long INF BAI BDI INTSCORE	PWAS	MILES	AGE_AD	FAM	VIC	OYAS	LENGT
> H EXIT RACE, i(ID) j(time)
reshaping m=0 data ...
(note: j = 2)
(note: INF2 not found)
(note: BAI2 not found)
(note: BDI2 not found)
(note: INTSCORE2 not found)
(note: PWAS2 not found)
(note: MILES2 not found)
(note: AGE_AD2 not found)
(note: FAM2 not found)
(note: VIC2 not found)
(note: OYAS2 not found)
(note: EXIT2 not found)
(note: RACE2 not found)
variable MILES already defined

/HTML]

Thank you in advance.

[Phil Bromiley]

I'm not sure it makes sense imputing before you reshape. I suspect that the reshape results in the assumptions in the impute not holding.

[Rich Goldstein]

have you looked at:

help mi reshape

if you want actual code help, please show the data using -dataex- (see the FAQ for more on this)

[roxy kallsen]

Thank you both for responding. I was able to figure out the problem. My ID was a string and the stubnames were followed by numbers such as 3 and 6 or 1 and 2. I think that was throwing STATA off in creating the time variable. I fixed it so all the variables that needed to be reshaped would be 1 and 2 only.