> If the increase at night then it's unlikely to be because of SUVs, unless there's correlating evidence that shows it's states where SUVs are driven at night m

> If the increase at night then it's unlikely to be because of SUVs, unless there's correlating evidence that shows it's states where SUVs are driven at night more. Or if you replace “people walk” by “people walk at night” in my example above… That is to say, there are many parameters at stake here, and you cannot infer or refute anything from the aggregated data.