Panel Data robustness

[taliagale]

Hi,

I am using least squares, panel data from 2006 to 2010 for 26 developing countries. Therefore, looking at 130 observations. I am using fixed effects to take into account the differences between countries. I am looking at the relationship between the political representation of women and sexual crime (dependent variable). I am using control variables including GDP per capita, male population, and strength of the legal system.

Are there any robustness checks I should undertake i.e. to ensure robust standard errors and how do I do these?

I am inputting: sexv c pol gdp malp leg

If anyone could offer advice would be greatly appreciate as panel data is new to me!

Thanks

[taliagale]

Forgot to mention, am using Eviews 8

[taliagale]

Anyone, how do I check for robust standard errors?

[startz]

Under panel options choose an appropriate coef covariance option

[taliagale]

Thank you. Also there may be endogeneity between the political representation of women on sexual violence(dependent) controlling for GDP per capita and male population. I tried using lags, however because I am using panel data between 2006-2010 lags may not be helpful.

Anyone know how to control for the endogeneity between women in parliament and sexual violence?


Thank you,

Talia

[taliagale]

How would I include an Instrumental Variable? This is quite new? anyone familiar with this??

[startz]

In the Estimate tab change the method to TSLS.