I'm a strict quant and even I don't give a s**t about your question about an interaction. Is it her treatment variable? If not then you look like the imbeciles trying to interprete a control variable in 2017. If you don't get why that is a huge critique of your empirical interpretation of a model than you are already out of date.
Moments at Job Talks that Made You Wince...

The correct response: That doesn't matter. It is still data, from which I got significant results. The probability of madeup data getting significant findings is < .05. I reject the null hypothesis of madeup data.
Q: how can you be sure that the survey firm in conflict torn country X didn't just make up all of your data?
A: [silence] 
Q: how can you be sure that the survey firm in conflict torn country X didn't just make up all of your data?
A: [silence]Well, silence is the wrong answer in this case. The survey research industry has welldeveloped (but not foolproof) methods for detecting data fabrication in these contexts. If you know your stuff, you also know this. BTW, I did survey work in Iraq for 5 years.

Q: how can you be sure that the survey firm in conflict torn country X didn't just make up all of your data?
A: [silence]Well, silence is the wrong answer in this case. The survey research industry has welldeveloped (but not foolproof) methods for detecting data fabrication in these contexts. If you know your stuff, you also know this. BTW, I did survey work in Iraq for 5 years.
... you did 'survey work' in Iraq for 5 years? BS.

This is really funny and exactly how I've heard formal theorists answer in the past.
Q: how do you know your formal model is right?A: I proved it in the appendix.
How else would one answer this?Quantard detected.... no one cares if the math works. Of course it will we can make a formal model in seconds doesn't mean it explains why dictators kill. What the theory? Why is *this* your model?

Q: how can you be sure that the survey firm in conflict torn country X didn't just make up all of your data?
A: [silence]Well, silence is the wrong answer in this case. The survey research industry has welldeveloped (but not foolproof) methods for detecting data fabrication in these contexts. If you know your stuff, you also know this. BTW, I did survey work in Iraq for 5 years.
If this is true, you're at a $hit program or not in poli sci. No one that matters and does opinion the Mid East/conflict zones has that on their CV.

This is really funny and exactly how I've heard formal theorists answer in the past.
Q: how do you know your formal model is right?A: I proved it in the appendix.
How else would one answer this?Quantard detected.... no one cares if the math works. Of course it will we can make a formal model in seconds doesn't mean it explains why dictators kill. What the theory? Why is *this* your model?
Sounds like you've never put together a formal model.