Yep, the whole series is about my book, which contains code for most of the topics, including CEM. See near the end of this section of Chapter 14,. www.theeffectbook.net/ch-Matching.html
@@NickHuntingtonKlein can you tell me what the 'w' in this code from the chapter you linked above represents? I am having trouble figuring out how to incorporate the CEM data into my regression analysis. In other words, once I have run the cem() code, what are the next steps? It looks like you re-weight your data based on CEM and then specify this weighting methodology in the regression? brcem % mutate(cem_weight = c$w) lm(responded~leg_black, data = brcem, weights = cem_weight).
@@cheriseregier4729 the w is the CEM weight. See further up the section to see how this is calculated. And yep, once you have the weight you can use it as a weight in your analysis to apply your CEM matching set/weights to any regression, means comparison, etc.
You just saved me at least 2 hours of after hours reading - so concise and helpful!!
Wow! This is the best series on matching techniques I've seen! Thank you!
Entropy balancing is such an interesting idea! Thank you for a great explanation, Nick 😊
Wow this video gave a really helpful explanation of entropy balancing (and coarsened exact matching as well)! Thank you!
Thank you for the video, and the book!! Very helpful :)
Very helpful video, greetings from Chile!
Very clear explanation!
Good explanation
Thank you. Do you know of an informative resource that lays out the R code for CEM?
Yep, the whole series is about my book, which contains code for most of the topics, including CEM. See near the end of this section of Chapter 14,. www.theeffectbook.net/ch-Matching.html
@@NickHuntingtonKlein Wonderful, thank you again.
@@NickHuntingtonKlein can you tell me what the 'w' in this code from the chapter you linked above represents? I am having trouble figuring out how to incorporate the CEM data into my regression analysis. In other words, once I have run the cem() code, what are the next steps? It looks like you re-weight your data based on CEM and then specify this weighting methodology in the regression?
brcem %
mutate(cem_weight = c$w)
lm(responded~leg_black, data = brcem, weights = cem_weight).
@@cheriseregier4729 the w is the CEM weight. See further up the section to see how this is calculated.
And yep, once you have the weight you can use it as a weight in your analysis to apply your CEM matching set/weights to any regression, means comparison, etc.
@@NickHuntingtonKlein Thank you🙏