Hello everyone,
I'm reaching out to the community for some insights regarding an analysis I've been conducting using the did_multiplegt_dyn routine. My focus has been on studying the simultaneous effect of a treatment alongside its impact 1, 2, and 3 years after the treatment has started. The results I've obtained are quite intriguing and have led me to a couple of questions I'm hoping to discuss with you all.
In my analysis, the immediate effect (contemporaneous effect) of the treatment is negative and statistically significant. However, as we look further down the line, by the third year, this effect turns positive and also becomes significant. This shift in the treatment's impact over time is puzzling and raises the question of the relevance of conducting a placebo test in this scenario. Given the change in direction and significance of the treatment effect, do you think a placebo test would still provide valuable insights, or might it be redundant?
Moreover, I'm contemplating the empirical validity of the model given these results. The varying significance and direction of the treatment effect over the years make me wonder about the model's ability to accurately capture the dynamics at play. Does this scenario raise concerns about the model's empirical soundness, or could it be indicative of a more complex interaction that the model is successfully identifying?
I'm eager to hear your thoughts, experiences, or any advice you might have on interpreting these findings. Specifically, I'm interested in your views on the necessity and utility of a placebo test in this context and how to assess the empirical validity of the model based on the observed results.
Thank you in advance for your time and insights!
Best,
Carlo
I'm reaching out to the community for some insights regarding an analysis I've been conducting using the did_multiplegt_dyn routine. My focus has been on studying the simultaneous effect of a treatment alongside its impact 1, 2, and 3 years after the treatment has started. The results I've obtained are quite intriguing and have led me to a couple of questions I'm hoping to discuss with you all.
In my analysis, the immediate effect (contemporaneous effect) of the treatment is negative and statistically significant. However, as we look further down the line, by the third year, this effect turns positive and also becomes significant. This shift in the treatment's impact over time is puzzling and raises the question of the relevance of conducting a placebo test in this scenario. Given the change in direction and significance of the treatment effect, do you think a placebo test would still provide valuable insights, or might it be redundant?
Moreover, I'm contemplating the empirical validity of the model given these results. The varying significance and direction of the treatment effect over the years make me wonder about the model's ability to accurately capture the dynamics at play. Does this scenario raise concerns about the model's empirical soundness, or could it be indicative of a more complex interaction that the model is successfully identifying?
I'm eager to hear your thoughts, experiences, or any advice you might have on interpreting these findings. Specifically, I'm interested in your views on the necessity and utility of a placebo test in this context and how to assess the empirical validity of the model based on the observed results.
Thank you in advance for your time and insights!
Best,
Carlo
