1
$\begingroup$

Borusyak (2021) writes the following:

We further discuss the implications of our results when treatment is simultaneous rather than staggered, when it can switch on and off, and when multiple treatment events can happen in the same unit.

There are three treatment styles in the above sentence. From my understanding, "when it can switch on and off" means that the laws can be implemented and removed after a period in the same country. But I don't understand the definition of the other two treatment styles: treatment is simultaneous and multiple treatment events can happen in the same unit.

From my understanding, "treatment is simultaneous" means that all countries passed a law at the same time. Is this correct? And what is multiple treatment events can happen in the same unit?

$\endgroup$

1 Answer 1

2
$\begingroup$

"From my understanding," treatment is simultaneous" is that all countries passed a law at the same time, is this correct?"

Yes.

And what is multiple treatment events can happen in the same unit ?

The same individual can be treated several times. i.e. suppose treatment is a binary variable for "the minimum wage increased". A region may have several increases in the minimum wage over the course of a data set.

$\endgroup$
7
  • $\begingroup$ Thanks @Micheal, but it is not totally clear to me about the "same individual can be treated several times". So, in each time the minimum wage increases, what is the movement of the binary variable ? $\endgroup$ Sep 16, 2021 at 21:04
  • 1
    $\begingroup$ @Louise I agree with you that "multiple treatment events" is a bit obscure without more information. For example, multiple treatments could be a unit experiencing, say, two sequential treatments, meaning they happen one after another. For example, the policy variable could represent different levels of intensity. On the other hand, a multi-treated unit could also mean a unit experiences a treatment intermittently; this is more characteristic of the 'on' and 'off' exposure pattern you referenced, such as when policies sunrise and sunset repeatedly over time. $\endgroup$ Sep 16, 2021 at 22:22
  • $\begingroup$ Agreed! I got what you mean, thanks @Thomas $\endgroup$ Sep 16, 2021 at 22:25
  • 1
    $\begingroup$ @Louise It could be binary. In other words, it could be a discretized measure (e.g., low, medium, high, etc.) representing the different exposures. Software will usually 'dummy out' the different levels for you. So, technically, they are binary variables representing the different levels of treatment. Does that make sense? $\endgroup$ Sep 16, 2021 at 22:42
  • 1
    $\begingroup$ @Louise By the way, a continuous measure can replace the binary treatment indicator. If you have a continuous treatment variable, then include it! I'm just letting you know that the model can also handle more than one 'discrete' treatment variable. $\endgroup$ Sep 16, 2021 at 22:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.