I'm new to the forum. I am trying to conduct an eventstudy on the effect of a particular event on daily airline stock returns. I'm not sure which model would suit best. I started with a classic market model and included a dummy variable for event days. But I am afraid that theevent has also affected the indices which I use as proxy for market returns. Is that a problem? Would a simple ARMA model with a dummy variable for event days be more suitable?

The event is about a month long. How many days do you suggest I include before the event and should I include any days after it?

Should I use log returns or absolute ones. Finally, should I use actual close prices or ones adjusted for dividends and splits?

Any advice would be greatly appreciated. Thanks!