asclogit can't convergence

[Hung Hsiang]

Hi everyone,

I've recently encountered a problem with a logit model that can't converge. My research topic is about understanding whether a company's characteristics influence people's choices among different alternatives.

I have four alternatives, each under two conditions: with and without a companion. This setup results in eight dummy variables (4 alternatives * 2 conditions). I included seven of these dummy variables in my model. However, the model fails to converge even after I tried increasing the maximum number of iterations and changing the base dummy variable. The log likelihood keeps repeating when it reaches a certain value.

Below are my data and the command I used. Can someone give me some advice on this? Thanks!

(alternative1-4 stands for terminal, cargo offsite, neither sequently)
gen accompany_terminal=accompany*const1
gen accompany_cargo=accompany*const2
gen accompany_offsite=accompany*const3
gen accompany_neither=accompany*const4

gen Dnoaccompany_terminal=0
replace Dnoaccompany_terminal=1 if accompany_terminal==0 &alt==1
gen Dnoaccompany_cargo=0
replace Dnoaccompany_cargo=1 if accompany_cargo==0 &alt==2
gen Dnoaccompany_offsite=0
replace Dnoaccompany_offsite=1 if accompany_offsite==0 &alt==3
gen Dnoaccompany_neither=0
replace Dnoaccompany_neither=1 if accompany_neither==0 &alt==4

gen Dwaccompany_terminal=0
replace Dwaccompany_terminal=1 if accompany_terminal>0 &alt==1
gen Dwaccompany_cargo=0
replace Dwaccompany_cargo=1 if accompany_cargo>0 &alt==2
gen Dwaccompany_offsite=0
replace Dwaccompany_offsite=1 if accompany_offsite>0 &alt==3
gen Dwaccompany_neither=0
replace Dwaccompany_neither=1 if accompany_neither>0 &alt==4

asclogit AW dailydiscountedcharge totaltime Ddaily Dhourlim indoormix contact spaceprob contactspace Dnoaccompany_terminal Dwaccompany_terminal Dnoaccompany_cargo Dwaccompany_cargo Dnoaccompany_offsite Dwaccompany_offsite Dnoaccompany_neither ,case(scenario) alternatives(alt) vce (cluster num) nolog

(I remove some column of variable since the Message Board can't fit in )

number	ver	circum	scenario	alt	accompany	days	AW	const1	const2	const3	const4	accompany_terminal	accompany_cargo	accompany_offsite	accompany_neither	Dnoaccompany_terminal	Dnoaccompany_cargo	Dnoaccompany_offsite	Dnoaccompany_neither	Dwaccompany_terminal	Dwaccompany_cargo	Dwaccompany_offsite	Dwaccompany_neither
1	Ａ	1	1	1	5	5	0	1	0	0	0	5	0	0	0	0	0	0	0	1	0	0	0
1	Ａ	1	1	2	5	5	1	0	1	0	0	0	5	0	0	0	0	0	0	0	1	0	0
1	Ａ	1	1	3	5	5	0	0	0	1	0	0	0	5	0	0	0	0	0	0	0	1	0
1	Ａ	1	1	4	5	5	0	0	0	0	1	0	0	0	5	0	0	0	0	0	0	0	1
1	Ａ	2	2	1	5	3	0	1	0	0	0	5	0	0	0	0	0	0	0	1	0	0	0
1	Ａ	2	2	2	5	3	1	0	1	0	0	0	5	0	0	0	0	0	0	0	1	0	0
1	Ａ	2	2	3	5	3	0	0	0	1	0	0	0	5	0	0	0	0	0	0	0	1	0
1	Ａ	2	2	4	5	3	0	0	0	0	1	0	0	0	5	0	0	0	0	0	0	0	1
1	Ａ	3	3	1	0	7	1	1	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0
1	Ａ	3	3	2	0	7	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0	0	0
1	Ａ	3	3	3	0	7	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0	0
1	Ａ	3	3	4	0	7	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0
2	Ｅ	1	4	1	3	6	1	1	0	0	0	3	0	0	0	0	0	0	0	1	0	0	0
2	Ｅ	1	4	2	3	6	0	0	1	0	0	0	3	0	0	0	0	0	0	0	1	0	0
2	Ｅ	1	4	3	3	6	0	0	0	1	0	0	0	3	0	0	0	0	0	0	0	1	0
2	Ｅ	1	4	4	3	6	0	0	0	0	1	0	0	0	3	0	0	0	0	0	0	0	1
2	Ｅ	2	5	1	3	7	0	1	0	0	0	3	0	0	0	0	0	0	0	1	0	0	0
2	Ｅ	2	5	2	3	7	0	0	1	0	0	0	3	0	0	0	0	0	0	0	1	0	0
2	Ｅ	2	5	3	3	7	1	0	0	1	0	0	0	3	0	0	0	0	0	0	0	1	0
2	Ｅ	2	5	4	3	7	0	0	0	0	1	0	0	0	3	0	0	0	0	0	0	0	1
2	Ｅ	3	6	1	0	11	0	1	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0
2	Ｅ	3	6	2	0	11	1	0	1	0	0	0	0	0	0	0	1	0	0	0	0	0	0
2	Ｅ	3	6	3	0	11	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0	0
2	Ｅ	3	6	4	0	11	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0

[Carlo Lazzaro]

Hung:
the usual recipe is starting with a more parsimonious model, add one more predictor at a time and see when Stata starts to complain about convergence.

[Hung Hsiang]

Hung:
the usual recipe is starting with a more parsimonious model, add one more predictor at a time and see when Stata starts to complain about convergence.

Dear Carlo,

Yes, that's true.
I already tried the model without the variable which are related with "accompany"
And it does has the result.
Probably I should try variable "accompany" ,if it can converge.
Then breakdown it into more detailed alternative specific variable.
Thanks for reply.
(I use STATA 15.1)

Best Regards,

[Hung Hsiang]

Answer my own questions,
Since the dummy variable are highly associate with ASCs.
I just dropped some of it. And the model did converges.