Dear all,
I have looked into the similar question (recently asked here: http://www.statalist.org/forums/foru...-in-panel-data) , but my data structure is slightly different.
I have a dependent variable which takes values from 0 to 15, and there are almost 50 % observation with zero value. My data is the pooled cross-sections of annual household survey, (1990-2010), and consist of state identifier and time identifier. Also, my explanatory variable is measured at the state level. In particular, I am trying to examine the impact of state unemployment rate (UR) on doctors visits (Visits) at the household level (dependent variable). .I was wondering whether it is possible (biased/consistency) to run the zinb model with controls for state dummies, time dummies, and linear state specific time trend like the following :
i.state i.year i.state #c.year
Thank you.
I have looked into the similar question (recently asked here: http://www.statalist.org/forums/foru...-in-panel-data) , but my data structure is slightly different.
I have a dependent variable which takes values from 0 to 15, and there are almost 50 % observation with zero value. My data is the pooled cross-sections of annual household survey, (1990-2010), and consist of state identifier and time identifier. Also, my explanatory variable is measured at the state level. In particular, I am trying to examine the impact of state unemployment rate (UR) on doctors visits (Visits) at the household level (dependent variable). .I was wondering whether it is possible (biased/consistency) to run the zinb model with controls for state dummies, time dummies, and linear state specific time trend like the following :
i.state i.year i.state #c.year
Thank you.