mi estimate and loop

[Ilias Geo]

Dear all,

I want calculate the mean of the hb0900 variable (property valuation) between 1942 and 2016 and years do not always run sequentially. A sample of my survey dataset looks like this:

hb0700	hb0800
2008	130000
2007	314465
1994	30000.5
1961	20000
1993	30000

My code is below and the he problem is that every time the mi estimate command runs the replace meanhb0800 replaces all 1,004 observations but I want it to replace it each observation at a time so I can have the meanhb0800 containing the mean for each respective year. {The mi estimate commands takes care of the missing observations so don't worry about that part.)

gen meanhb0800 = .

forval i = 1956/2016 {

mi estimate, esampvaryok vceok: svy: mean hb0800 if hb0700==`i'
scalar b_ih = el(e(b_mi),1,1)
replace meanhb0800 = el(e(b_mi),1,1)
}

Any help would be appreciated.
Ilias

[Raffaele Grotti]

What about

...
replace meanhb0800 = el(e(b_mi),1,1)  if hb0700==`i'
...

?

[Clyde Schechter]

I'm not sure I follow what you want here, but I think what you need to do is add -if hb0700 == `i'- to your -replace meanhb0800 = ...- command.

Added: Crossed with #2.

[Ilias Geo]

Yes, your suggestion worked.
Thank you both!
Ilias