Help on choosing the right regression

[Eato Clement]

Hello!

I am a real beginner with Stata and would like to know if I am doing the right things, I am doing research on panel data of venture capital investments done in start-ups. It would be great if someone could let me know if I am choosing the right regressions for this:

IV: Investment amount
DV: Start-up performance, measured by;

• Going public (Yes/No)
• Employment growth rate
• Measured after the 1^st investment date + 5 years and +10 years

Moderator 1: Industry Relatedness

• Measured from 1 to 5 (1 being no relatedness and 5 fully related)

Moderator 2: Investment Stage

• Measured by Early or Late Stage, Early stage being 1 and Late being 0

Regression types:
IV -> DV Going Public : xtprobit
IV -> Growth rate : xtreg
Moderator 1: Interaction variable added Investment amount x Relatedness xtprobit & xtreg
Moderator 2: Interaction variable Investment amount x Investment stage xtprobit & xtreg
Control variables: industry_growth_rate, industry_size, total_funding_received, cvc_parent_size

Question: How can you let Stata do the analysis from the first investment done in the company, or is this done automatically by Stata? Because I also have panel data from before the first investment.


Thank you very much for your guidance!

[Carlo Lazzaro]

Eato:
welcome to this forum.
1) -xtreg. and -xtprobit- are ok for your different research goals. You may want to consider -xtlogit- as an alternative to -xtprobit-, as -xtlogit- allows the consitional fixed effect estimator (whereas -xtprobit- does not);
2) see -fvvarlist- notation for creating categorical variables and interactions. Pay attention to -i.- and -c.- prefixes as well as -#- and -##- symbols;
3) by default Stata omits all the observations with missing values in any variable (listwise deletion). Therefore, if the -investment- predictor has mssing values, those observations will be excluded from the regression.

[Eato Clement]

Dear Carlo,

Thank you very much for your answers! One more questions from my side; for the investments made in the start-ups throughout the years there are also years without an investment. Would it be wise to insert a 0 here or just leave the cell empty? I can't find any information on how this would influence the results of the analysis.

Kind regards

[Carlo Lazzaro]

Eato:
if you're 100% sure that no investment ina given year does not mean a missing value for that year, plugging in 0 value sounds OK.
Conversely, if you leave the value "empty" (that in Stata means "period" if the format of the variable is numeric), the observation will be excluded from the statistical procedure via listwise deletion.