mi xeq: predict leverage, hat --> Generates only missing values

[Tim Albrecht]

Hello there,


i am new to this forum and I have problem with using mi xeq 5: predict leverage, hat. After proceeding this command on m=5, there are only missing values generated and I cannot explain this. Maybe you can help me?

Thanks a lot!

I am using the following code:

mi estimate, dots: logit H_WOHNEIGENTUM i.PoM i.HP_SEX i.P_MIG c.HP_ALTER i.P_BIL_sum i.HP_BKAT i.aq_eink_sum i.H_EINK_sum if hhtyp_sum == 1
mi xeq 5: logit H_WOHNEIGENTUM i.PoM i.HP_SEX i.P_MIG c.HP_ALTER i.P_BIL_sum i.HP_BKAT i.aq_eink_sum i.H_EINK_sum if hhtyp_sum == 1
mi xeq 5: predict leverage, hat

[Rich Goldstein]

I believe that you will need to concatenate your 2 "mi xeq 5" commands by adding "; predict leverage, hat" to the end of the first mi xeq 5 command and then deleting "mi xeq 5: predict leverage, hat" - but there may be other things going on and we don't know what style your data is in; see

help mi xeq

[Tim Albrecht]

Hello Rich,

thanks for your fast reply! Now it works, but I have a following questions:


With mi xeq in m=5 there is no variable "leverage" generated, see:

mi xeq 5: logit H_WOHNEIGENTUM i.PoM i.HP_SEX i.P_MIG c.HP_ALTER i.P_BIL_sum i.HP_BKAT i.aq_eink_sum i.H_EINK_sum if hhtyp_sum == 1; predict leverage, hat

With mi xeq on each imputation, there is a variable leverae, see:

mi xeq: logit H_WOHNEIGENTUM i.PoM i.HP_SEX i.P_MIG c.HP_ALTER i.P_BIL_sum i.HP_BKAT i.aq_eink_sum i.H_EINK_sum if hhtyp_sum == 1; predict leverage, hat

What is the reason for this?

Thanks a lot,
Tim

[Tim Albrecht]

Hello Rich,

thanks for your fast reply! Now it works, but I have a following questions:


With mi xeq in m=5 there is no variable "leverage" generated, see:

mi xeq 5: logit H_WOHNEIGENTUM i.PoM i.HP_SEX i.P_MIG c.HP_ALTER i.P_BIL_sum i.HP_BKAT i.aq_eink_sum i.H_EINK_sum if hhtyp_sum == 1; predict leverage, hat

With mi xeq on each imputation, there is a variable leverae, see:

mi xeq: logit H_WOHNEIGENTUM i.PoM i.HP_SEX i.P_MIG c.HP_ALTER i.P_BIL_sum i.HP_BKAT i.aq_eink_sum i.H_EINK_sum if hhtyp_sum == 1; predict leverage, hat

What is the reason for this?

Thanks a lot,
Tim

[Rich Goldstein]

I don't know

but, what do you mean by "Now it works" if the variable is not generated?

[Tim Albrecht]

Well, the calculation of the leverage on m=5 works, so there are not only missings anymore. But only with mi exq on each imputation there is a variable leverage generated.

[Rich Goldstein]

sorry, but I am confused - if there values other than missings then there must be a variable???? maybe I'm missing something that someone else can help with

[Tim Albrecht]

That confuses me, too and i think that I am missing something as well. But I will go on trying this week, let's see if I find a solution for that.