Hi,

Im running some regressions to investigate the effects corruption has on enrolment rates. I am using panel data for different countries across different years. One of my control variables is lnGDP, but it was been suggested to me to use dummy variables for High income/Middle income/Low income countries. I have done this and I am getting some strange results. I would have thought that including lnGDP would mean I wouldn't need to include this dummy variable?

Any advice would be appreciated,

Thanks!