> Isn’t tobacco use trending down anyways? How could you possibly tease out the drop in tobacco use from one change to packaging? Anything in the real world h

> Isn’t tobacco use trending down anyways? How could you possibly tease out the drop in tobacco use from one change to packaging? Anything in the real world has many many factors affecting its consumption. What would it matter if it was trending down specifically? Or are you asking what models and statistical tools they use to make analyses like this when there are a complex set of factors at play?