Question: correct way to use fixed SD in truncreg

[Jasper Neckarauer]

Dear all,

is this the correct way to apply a fixed standard deviation to a truncated regression? The goal is to use a known SD and gather a precise mean estimate from the truncreg:


display `sd' // 5.62 cm
5.6180141

. truncreg ht, ll(`ll') constraints(1)
(0 obs truncated)

Fitting full model:

Iteration 0: Log likelihood = -8641.1176
Iteration 1: Log likelihood = -8599.1094
Iteration 2: Log likelihood = -8599.0614
Iteration 3: Log likelihood = -8599.0614

Truncated regression
Limit: Lower = 161.04669 Number of obs = 3,061
Upper = +inf Wald chi2(0) = .
Log likelihood = -8599.0614 Prob > chi2 = .

( 1) [/]sigma = 5.618014
------------------------------------------------------------------------------
ht | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
_cons | 168.2695 .1202349 1399.51 0.000 168.0339 168.5052
-------------+----------------------------------------------------------------
/sigma | 5.618014 (constrained)
------------------------------------------------------------------------------


Thanks!