Add DiD section

[grantmcdermott]

Closes #10 (updates to some IV packages notwithstanding)

Note that I wrapped DiD together with IV and RDD under a "canonical research designs" heading. This seemed to me the most logical way to include these three sections. But we can easily change back if you don't like.

[zeileis]

Grant @grantmcdermott thanks for this! Feel free to merge this into the main branch.

The only question/comment I have: Is "canonical research designs" a commonly-used umbrella term for this? Maybe "research designs for causal inference" would be a bit more explicit? But I'm not sure what most readers would find intuitive here.

[grantmcdermott]

Great.

RE: Canonical Research Designs. I think I stole the phrasing from @paulgp, who in turn inherited it from Gary Chamberlain (I believe, although I might be making this up entirely). Happy to switch it out to something like your suggestion if that's more palatable.

Let me think it over tonight. I might add a quick synthetic control bullet point under the DiD heading too. It's a bit awkward having SC and DiD in the same section. But there is an overlap i.t.o. packages and it's not a bad compromise given that SC is mostly a topic for the Casual Inference TaskView.

[paulgp]

I think I just made it up (although wouldn't put it past me to have subconsciously gotten it from Gary Chamberlain) for my class lectures: https://github.com/paulgp/applied-methods-phd

It's a useful distinction from just pure "strict ignorability" type causal inference, such as RCTs, given that there seems to be a set number of research approaches that everyone needs in their toolhouse (DiD, RD, and IV).

[grantmcdermott]

It's a useful distinction from just pure "strict ignorability" type causal inference, such as RCTs, given that there seems to be a set number of research approaches that everyone needs in their toolhouse (DiD, RD, and IV).

Yeah, I like it a lot for that reason. I'll fully credit you (and you only!) from now on ;-)

[paulgp]

Hah! I don't need the credit, but glad it's catchy and useful. :-)

On Thu, Dec 29, 2022 at 5:25 PM Grant McDermott @.***> wrote:

It's a useful distinction from just pure "strict ignorability" type causal inference, such as RCTs, given that there seems to be a set number of research approaches that everyone needs in their toolhouse (DiD, RD, and IV).

Yeah, I like it a lot for that reason. I'll fully credit you (and you only!) from now on ;-)

— Reply to this email directly, view it on GitHub https://github.com/cran-task-views/Econometrics/pull/15#issuecomment-1367614786, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABDTDFSXTAQN2RGOOMWGLBTWPYFVXANCNFSM6AAAAAATLAXDVM . You are receiving this because you were mentioned.Message ID: @.***>

[zeileis]

Nice! Grant @grantmcdermott feel free to add a short reference/link in the introductory paragraph to that section in the task view.