Help with hurdle model

[charlotte constable]

Hello -

i'm having some trouble running a churdle model. I have observational data (4,800 observations) and i want to look at the association between minutes spent exercising and perceived safety of neighbourhood (categorical variable - 3 responses). There are a large number of 0's in the minutes of exercise hence why i'm trying a hurdle model. Other variables i want to consider in the model are ethnicity, family income and education but i'm not entirely sure where to put these!

i have tried a basic code so far which just seems to run for ages and ages and give no output:
churdle linear MVPA i.safety, select (i.safety) ll(0)

i also tried just running the first part of the model with truncreg MVPA i.safety, ll(0) - this came with an error message of 'cannot compute an improvement -- discontinuous region encountered'

Any help or advice would be much appreciated!

thanks

[Jeff Wooldridge]

I wonder if the variable i.safety is perfectly predicting a zero outcome. What happens if you just use a probit for selection with i.safety?

[Bilal Caliskan]

Hello all,

I am working on developing hurdle model to analyze food consumption of households, especially for food consumed away from home by households.

To accomplish my analysis, I applied churdle command with following code. By the way, my data is survey data and I applied svy command too.


svy: churdle linear totalpaid_hh i.white i.anyvehicle i.rural i. northeast i.south i.midwest ln_income i.snapnowhh hhsize free_event_num i.child , select(i.prim_employed i.prim_age i.prim_sex prim_educ i.child) ll(0). Then, I got the attached linear output file. In this part, all linearized standard errors, t and p-values are missing.

However, when I applied exponential churdle model with following code

svy: churdle exponential totalpaid_hh i.white i.anyvehicle i.rural i. northeast i.south i.midwest ln_income i.snapnowhh hhsize free_event_num i.child , select(i.prim_employed i.prim_age i.prim_sex prim_educ i.child) ll(0). The output seems ok. I also, attached the exponential outcome output.

I would like to ask your help to understand why linear churdle command didnot work but, exponential churdle command work. Also, do you have any idea to make linear churdle command worked in my analysis? I am expecting to get your comments and helps. Thanks in advance.

Best,