Data not MCAR

[Sherine Maui]

Hi statalist,

I am using little’s test to check if my data is MCAR, which it is not. The variable with missing data is income and approx. 29% of data is missing. My total sample size is 1,800 individuals.

. mcartest income wor1 wor2 wor3 wor4 wor5 wor6 wor7 wor8 wor9
 
Little's MCAR test
 
Number of obs       = 1726
Chi-square distance = 22.0935    
Degrees of freedom  = 9
Prob > chi-square   = 0.0086

I am not sure how to proceed. I need to run a fixed effects regression and include this variable as one of my covariates. Can I still try to impute the missing values?

* Example generated by -dataex-. To install: ssc install dataex
clear
input str14 ID float outcome byte(income wor1 wor2) long wor3 byte(wor4 wor5 wor6 wor7 wor8 wor9 wave)
"10504000000002"   -.3186359 2 2 4 2 1 0 0 0 0 1 1
"10504000000002"  -3.1873674 2 1 1 2 0 0 0 0 0 0 2
"10504000000043"  -2.4036455 . 1 2 1 0 1 0 0 0 1 1
"10504000000043"    2.613899 . 1 1 2 0 1 0 0 0 0 2
"10504000000113"    1.234423 1 2 3 3 0 1 0 0 0 0 1
"10504000000113"   .25891808 1 3 2 2 0 0 0 0 0 0 2
"10504000000231"   -2.949189 4 2 2 1 1 0 0 0 0 0 1
"10504000000231"    1.567996 4 1 4 3 0 1 1 0 0 1 2
"10504000000444"    .9008501 1 2 4 2 1 0 0 0 0 0 1
"10504000000444"    .8814207 1 3 2 2 1 0 0 0 0 0 2
"10504000000482"    2.613899 1 1 4 3 1 0 0 0 0 0 1
"10504000000482"   1.4012096 1 1 2 3 0 0 0 0 1 0 2
"10504000000535"  -1.0241696 3 1 1 1 0 0 0 0 0 0 1
"10504000000535"   -2.236859 3 1 4 1 0 1 0 0 0 0 2
"10504000000644"   2.4471126 . 1 4 1 0 0 0 0 0 0 1
"10504000000644"   1.9467533 . 1 4 2 1 1 0 0 1 0 2
"10504000000696"    -2.92976 . 1 1 2 1 0 0 0 0 0 1
"10504000000696"   -3.449549 . 1 1 1 1 0 0 0 0 0 2
"10504000000748"    1.567996 . 4 4 2 0 1 0 0 0 0 1
"10504000000748"   -1.370393 . 4 4 2 1 1 0 0 0 0 2
"10504000000821"   1.4012096 1 3 3 2 1 0 0 0 0 0 1
"10504000000821"   1.0676366 1 2 4 1 0 0 0 0 0 0 2
"10504000000978"   -2.419119 1 4 2 1 0 0 0 0 0 0 1
"10504000000978"    .3553065 1 2 2 3 1 1 0 0 0 0 2
"10504000000991"   1.4012096 2 2 2 3 0 1 0 0 0 0 1
"10504000000991"  -1.0390252 2 4 3 1 1 0 0 0 0 0 2
"10504000001036"    2.613899 2 1 4 3 1 1 0 1 0 0 1
"10504000001036"   1.3244863 2 3 1 3 1 0 0 0 0 0 2
"10504000001052"   .52209294 . 2 4 2 1 0 0 0 0 0 1
"10504000001052"   -3.449549 . 2 2 2 1 0 0 0 0 0 2
"10504000001073"   1.4012096 1 2 3 3 1 0 0 0 0 0 1
"10504000001073"  -1.5697132 1 1 3 2 1 0 0 0 0 0 2
"10504000001102"   .18347497 1 4 4 3 1 1 0 0 0 0 1
"10504000001102"   -2.236859 1 1 4 3 1 0 0 0 0 0 2
"10504000001142"   -2.236859 . 1 4 2 1 1 0 0 0 0 1
"10504000001142"    2.613899 . 1 2 3 1 0 0 0 0 0 2
"10504000001236"   -2.402652 1 1 4 2 0 0 0 0 1 0 1
"10504000001236"   -1.563859 1 1 4 3 0 0 0 0 1 0 2
"10504000001239"   1.9467533 . 1 1 1 0 0 0 0 0 0 1
"10504000001239"  -.03875893 . 2 1 1 1 0 0 0 1 0 2
"10504000001244"      .18852 1 4 4 2 0 1 0 0 0 0 1
"10504000001244"   -.8573831 1 2 4 3 1 1 0 0 1 0 2
"10504000001261"    2.613899 1 4 2 3 0 1 0 0 0 0 1
"10504000001261"    2.613899 1 3 4 3 1 1 0 0 1 0 2
"10504000001304"   1.4012096 1 4 4 3 0 1 0 0 0 0 1
"10504000001304"   1.4012096 1 1 4 3 0 0 0 0 0 0 2
"10504000001470"   -3.449549 1 1 1 1 1 0 0 0 0 0 1
"10504000001470"   -3.191941 1 1 4 1 0 0 0 0 0 0 2
"10504000001471"  -2.4997985 1 3 2 1 1 0 0 0 0 0 1
"10504000001471"   -.5364607 1 1 2 1 0 0 0 0 0 0 2
"10504000001499" -.005772657 2 1 4 2 0 1 0 0 0 0 1
"10504000001499"   -2.236859 2 4 4 2 1 1 0 0 0 0 2
"10504000001510"   -3.282762 . 1 1 1 1 0 0 0 0 0 1
"10504000001510"   -3.449549 . 1 1 1 0 0 0 0 0 0 2
"10504000001523"  -1.0241696 1 3 3 1 0 0 0 0 0 0 1
"10504000001523"   1.4012096 1 2 1 1 1 0 0 0 0 0 2
"10504000001624"   -1.832888 . 1 4 2 0 0 0 0 0 0 1
"10504000001624"   -3.449549 . 1 2 3 0 1 0 0 0 0 2
"10504000001632"   1.4012096 . 3 3 1 1 0 0 0 0 0 1
"10504000001632"   -2.236859 . 1 2 1 1 0 0 0 0 0 2
"10504000001704"   1.4012096 1 1 3 3 0 1 0 0 0 0 1
"10504000001704"    2.613899 1 1 4 2 0 0 0 0 0 0 2
"10504000001779"  -1.5697132 1 4 4 2 1 0 0 0 0 0 1
"10504000001779"  -1.7364997 1 2 2 1 1 0 0 0 0 0 2
"10504000001832"    1.567996 2 1 2 3 0 0 0 0 0 1 1
"10504000001832"   2.4471126 2 2 2 2 0 1 0 0 0 0 2
"10504000001834"   -.9287747 . 4 4 3 1 0 0 0 0 0 1
"10504000001834"   1.9467533 . 1 1 2 0 1 0 0 0 0 2
"10504000001909"    .3553065 . 4 2 1 0 0 0 0 0 0 1
"10504000001909"   -2.236859 . 4 2 2 1 0 0 0 0 0 2
"10504000001915"  -2.4036455 2 3 2 2 0 0 0 0 1 0 1
"10504000001915"   -.3366007 2 1 2 2 0 1 0 0 1 0 2
"10504000001932"   -2.615616 2 1 1 1 0 0 0 0 0 0 1
"10504000001932"   -3.449549 2 1 2 1 0 0 0 0 0 0 2
"10504000001943"    1.831171 1 1 2 2 1 0 0 0 0 0 1
"10504000001943"   2.3509598 1 1 4 3 0 0 0 0 0 1 2
"10504000001966"  -.52381015 2 1 2 2 1 0 0 0 1 0 1
"10504000001966"   -2.236859 2 1 2 1 1 0 0 0 0 0 2
"10504000001968"  -1.0241696 2 3 3 2 1 0 0 0 0 0 1
"10504000001968"   -.5364607 2 1 2 1 1 0 0 0 0 0 2
"10504000002069"    .4461275 1 4 4 2 1 0 0 0 0 0 1
"10504000002069"   .18243033 1 4 1 2 0 1 0 0 0 0 2
"10504000002139"   1.9318976 . 1 2 2 0 0 0 0 0 0 1
"10504000002139"   -3.449549 . 4 1 3 0 0 0 0 0 0 2
"10504000002297"   -2.615616 1 1 3 2 1 0 0 0 0 0 1
"10504000002297"   -3.449549 1 1 1 2 0 0 0 0 0 0 2
"10504000002304"  -1.7364997 1 2 2 2 1 0 0 0 0 1 1
"10504000002304"   1.4012096 1 2 1 3 1 1 0 0 0 0 2
"10504000002326"   -3.449549 1 4 4 2 0 0 0 0 0 0 1
"10504000002326"    2.613899 1 4 4 2 0 0 0 0 0 0 2
"10504000002390"   -1.190956 . 2 3 2 0 0 0 0 0 0 1
"10504000002390"  -1.0387385 . 1 1 1 0 1 0 0 0 0 2
"10504000002476"    .7814519 3 2 2 2 1 0 0 0 0 0 1
"10504000002476"   -3.449549 3 4 4 2 1 0 0 0 0 0 2
"10504000002694"    2.613899 1 2 3 2 1 0 0 0 0 1 1
"10504000002694"  -1.3275787 1 2 4 1 0 0 0 0 0 0 2
"10504000002727"   -1.190956 1 1 4 2 1 0 0 0 0 0 1
"10504000002727"   -3.282762 1 4 1 1 0 0 0 0 0 0 2
"10504000002766"   -1.454131 2 1 1 2 1 0 0 0 0 0 1
"10504000002766"   1.4012096 2 2 2 3 0 0 0 0 0 0 2
end
label values income Lma_COR22
label def Lma_COR22 1 "Less than 2500 MAD", modify
label def Lma_COR22 2 "2500 - less than 5000 MAD", modify
label def Lma_COR22 3 "5000 -less than 10000 MAD", modify
label def Lma_COR22 4 "10000 or more", modify
label values wor1 COR34
label def COR34 1 "Not at all", modify
label def COR34 2 "A little", modify
label def COR34 3 "Rather", modify
label def COR34 4 "Very", modify
label values wor2 COR35
label def COR35 1 "Not at all", modify
label def COR35 2 "A little", modify
label def COR35 3 "Rather", modify
label def COR35 4 "Very", modify
label values wor3 foodc
label def foodc 1 "High", modify
label def foodc 2 "Moderate", modify
label def foodc 3 "Low", modify
label values wor4 COR27
label def COR27 0 "Not Mentioned", modify
label def COR27 1 "Mentioned", modify
label values wor5 V104_A
label def V104_A 0 "Not Mentioned", modify
label def V104_A 1 "Mentioned", modify
label values wor6 V105_A
label def V105_A 0 "Not Mentioned", modify
label def V105_A 1 "Mentioned", modify
label values wor7 V106_A
label def V106_A 0 "Not Mentioned", modify
label def V106_A 1 "Mentioned", modify
label values wor8 V107_A
label def V107_A 0 "Not Mentioned", modify
label def V107_A 1 "Mentioned", modify
label values wor9 V108_A
label def V108_A 0 "Not Mentioned", modify
label def V108_A 1 "Mentioned", modify
label values wave wave
label def wave 1 "Wave 1", modify
label def wave 2 "Wave 2", modify

[Rich Goldstein]

multiple imputation assumes the data are MAR (which is not testable); so, given your knowledge of the context and the data is MAR reasonable? see

help mi

[Carlo Lazzaro]

Sherine:
if you assume that your data are MAR, you can go -mi- before -xtreg,fe-.

[Sherine Maui]

Thank you Rich and Carlo, I have decided to go with mi before xtreg.