How to find equation numbers for use with etable option equations()

[Kate OHara]

I would like to display two melogit models side-by-side using etable. They have different outcome variables but identical controls. On first attempt the control variables are not aligned so I attempt to fix this using the equations() option. However, non of my attempts are working. I would like to know how I can see what the equation numbers are. Can I list these somehow?

And is the following expected behaviour? The first attempt, with no equations() option does not show the level two variance (turn var(_cons)). When I attempt to specify any equations() option, the table is rendered with only the "Number of observations" row.

Thank you.

This code recreates my problem:

webuse auto, clear

melogit foreign rep78 || turn:,
est store m1

drop if mpg > 38

gen high = headroom > 4
melogit high rep78 || turn:,
est store m2

etable, estimates(m1 m2)
etable, estimates(m1 m2) equations(1)
etable, estimates(m1 m2) equations(1:1)
etable, estimates(m1 m2) equations(1:1, 2:2)

The four respective outputs look like this:

[Andrew Musau]

See #11: https://www.statalist.org/forums/for...date-for-v17-0

[Kate OHara]

Thank you, Andrew. I didn't spot that one when searching.

So to update, using the information from https://www.statalist.org/forums/for...date-for-v17-0 , I can align the tables by getting the equation names using:

etable, replay showeq

And then recode the equation names as I specify the etable. The following code solves my particular problem.

webuse auto, clear

melogit foreign rep78 || turn:,
est store m1

drop if mpg > 38

gen high = headroom > 4
melogit high rep78   || turn:,
est store m2

etable, estimates(m1 m2)

--------------------------------------
                       foreign   high 
--------------------------------------
Repair record 1978       2.287        
                       (0.793)        
Intercept              -10.010        
                       (3.217)        
var(_cons[turn])         4.332   0.000
                       (3.842) (0.000)
Repair record 1978              -0.949
                               (0.528)
Intercept                        0.333
                               (1.506)
Number of observations      69      68
--------------------------------------

etable, replay showeq novarlabel

--------------------------------------
                       foreign   high 
--------------------------------------
foreign                               
  rep78                  2.287        
                       (0.793)        
  _cons                -10.010        
                       (3.217)        
/                                     
  var(_cons[turn])       4.332   0.000
                       (3.842) (0.000)
high                                  
  rep78                         -0.949
                               (0.528)
  _cons                          0.333
                               (1.506)
Number of observations      69      68
--------------------------------------


etable, estimates(m1 m2) eqrecode(foreign=xb high=xb)

--------------------------------------
                       foreign   high 
--------------------------------------
Repair record 1978       2.287  -0.949
                       (0.793) (0.528)
Intercept              -10.010   0.333
                       (3.217) (1.506)
var(_cons[turn])         4.332   0.000
                       (3.842) (0.000)
Number of observations      69      68
--------------------------------------

Using numbers as the manual says, e.g. any of the following, still produces the strange table with only the N row. I do wonder if that's a bug. However, I do have a solution to my particular problem (thanks again, Andrew).

etable, estimates(m1 m2) equations(1)
etable, estimates(m1 m2) equations(1:1)
etable, estimates(m1 m2) equations(1:1, 2:2)

All produce:

-----------------------------------
                       foreign high
-----------------------------------
Number of observations      69   68
-----------------------------------

[Andrew Musau]

Using numbers as the manual says, e.g. any of the following, still produces the strange table with only the N row. I do wonder if that's a bug. However, I do have a solution to my particular problem (thanks again, Andrew).

etable, estimates(m1 m2) equations(1)
etable, estimates(m1 m2) equations(1:1)
etable, estimates(m1 m2) equations(1:1, 2:2)

All produce:

-----------------------------------
foreign high
-----------------------------------
Number of observations 69 68
-----------------------------------

The option -equations()- in etable simply orders the equations in the resulting table:

equations(eqlist) specifies the equations to be included in the table and the order in which they are reported.

Since there is no equation named 1 or 2, the table ends up empty. My guess is that you are mixing up this option with the -equations()- option from estout (from SSC), which does what you assume:

equations(eqmatchlist) match the models' equations

But the two options do not do the same thing.