I am studying the impact of having a data breach on the tone/financial reporting quality of the 10ks and to avoid confounding variables, I investigated that propensity score matching was the solution. However, I don't know or dispose of the variables that affect the choice of treatment i.e being breached but I have the variables that "control" the outcome i.e Tone. Does anyone have proposition??
One of the important assumptions of the propensity score analysis is no unmeasured confounding. Also, studying the impact and exploring the association are two different things. Please be specific. Here are some recommendations on which variables to include in the propensity score model (obviously, it requires subject area-expertise):
○ Include only pre-baseline measures
○ Confounders: important to include
○ Risk factors/predictors of the outcome: include to reduce the standard error
○ Instrumental variable/predictors of the exposure only: avoid
○ Mediator: avoid
The following can be a good resource: Article A Tutorial and Case Study in Propensity Score Analysis: An A...
Thank you very much!!
Yes, the problem is initially my regression was going to be:
Dependent variable- tone
Interest independent variable- breach
Control variables for tone
However, I saw that I couldn't measure correctly the propensity score, as I don't have the covariates that affect whether a firm is breached or not.
Do you have any suggestion, maybe differences in differences model with multiple time periods? With a within panel data structure
You may test parallel trend if the parallel trend is exist in prior treatment then you can use DID, PSM DID, Two Step System GMM...
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India