Hi,
I'm trying to run a test manipulation for an RDD, and I have a naive question. My running variable is the time, with a specific date as the threshold (in this case, threshold = 0). For the moment, I'm using the following command:
In my database, I have one observation = one transaction (made on a specific date) (and information relative to volume sold during this transaction). I have therefore several observations linked to the same date. I'm testing whether there is a difference in the number of transactions carried out per day before and after the threshold date.
I'd like to be able to test whether there is a difference in total volume per day (and not number of transactions per day), before and after the threshold date, but I'm not sure whether it’s possible. I thought about running the following commands:
But it feels wrong as I will end up having only one observation for each date… I’m not familiar at all with the rddensity command and I’m looking for more insights.
I'm trying to run a test manipulation for an RDD, and I have a naive question. My running variable is the time, with a specific date as the threshold (in this case, threshold = 0). For the moment, I'm using the following command:
Code:
rddensity date, plot
I'd like to be able to test whether there is a difference in total volume per day (and not number of transactions per day), before and after the threshold date, but I'm not sure whether it’s possible. I thought about running the following commands:
Code:
collapse (sum) volume, by (date) rddensity date, plot
