I am having difficulty wrapping my mind around conceptualizing the impact of one event on another.
For the scope of this question, let's define the following: Event A: An assault occurs; Event B: A rival gang completes a drug sale in hostile turf.
Consider perfect data, such that these events can be ideally quantified and analyzed. The period of data used for the analysis is 6 months.
I wish to discover the impact of Event B on Event A; however, as opposed to a boolean occurred/didn't occur, I want to take into consideration the number of assaults that occurred during the scope of Event B (severity of impact). Am I contaminating my question? Notably, (below) my formula fails to address the number of days in which assaults occurred.
From a transactional perspective, to determine a theoretical weighting of Event B upon A, I would assess the situation as follows:
$Days_T:$ Total number of days in the 6-month period
$Days_B:$ Total number of days on which a rival gang completes a drug sale in hostile turf in the 6 month period
$Assault_T:$ Total number of assaults in the 6-month period
$Assault_B:$ Total number of assaults on days in which a rival gang completed a drug sale in hostile turf in the 6 month period.
$(Assault_B \div Assault_T) \times (Days_T \div Days_B)$
Should I be using total number of days on which assaults occurred instead of the total number of days in the 6-month period?
Ideally, I am aiming to predict assaults based on a number of available precipitators, but I'm very much doubting my own logic at the moment.
Thanks for any help you can offer!