- Видео 25
- Просмотров 91 096
Paul Goldsmith-Pinkham
Добавлен 5 июл 2013
EJMR Username Prediction (Part 1 of 3)
The username scheme shown here was in effect from July 8, 2013 to May 17, 2023. It was slightly different from December 21, 2010 to July 8, 2013. Notice that there was no salt for the hash. This can be verified in the WayBack Machine. See, in particular, this post by the EJMR administrator: tinyurl.com/ejmr-username
In our paper, we show how the statistical properties of this scheme allow many posts to be assigned to IP addresses using only publicly available data. Our paper does not include personally identifiable information.
Join us at NBER SI on July 20, 2023 in person and streaming live on You Tube: www.nber.org/conferences/si-2023-digital-economics-and-artificial-intelligence
In our paper, we show how the statistical properties of this scheme allow many posts to be assigned to IP addresses using only publicly available data. Our paper does not include personally identifiable information.
Join us at NBER SI on July 20, 2023 in person and streaming live on You Tube: www.nber.org/conferences/si-2023-digital-economics-and-artificial-intelligence
Просмотров: 1 708
Видео
EJMR IP Determination (Part 2 of 3)
Просмотров 835Год назад
The username scheme shown here was in effect from July 8, 2013 to May 17, 2023. It was slightly different from December 21, 2010 to July 8, 2013. Notice that there was no salt for the hash. This can be verified in the WayBack Machine. See, in particular, this post by the EJMR administrator: tinyurl.com/ejmr-username In our paper, we show how the statistical properties of this scheme allow many ...
EJMR IP Statistics (Part 3 of 3)
Просмотров 805Год назад
The username scheme shown here was in effect from July 8, 2013 to May 17, 2023. It was slightly different from December 21, 2010 to July 8, 2013. Notice that there was no salt for the hash. This can be verified in the WayBack Machine. See, in particular, this post by the EJMR administrator: tinyurl.com/ejmr-username In our paper, we show how the statistical properties of this scheme allow many ...
Identification Song (work-in-progress)
Просмотров 2,4 тыс.2 года назад
Song is a work in progress (and mainly a joke as I was reviewing my slides with Phil Haile‘s comments re: identification)
Lecture 21: Canonical Research Designs IX: RD II: The Checklist
Просмотров 1,5 тыс.3 года назад
Lecture 21 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 22: Canonical Research Designs X: RD III: Extensions
Просмотров 1,2 тыс.3 года назад
Lecture 17 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 19: Canonical Research Designs VII: Examiner Designs aka Judge IV
Просмотров 2,1 тыс.3 года назад
Lecture 19 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 20: Canonical Research Designs VIII: RD I: Identification and Groundwork
Просмотров 1,6 тыс.3 года назад
Lecture 20 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 17: Canonical Research Designs V: Instrumental Variables III
Просмотров 2,7 тыс.3 года назад
Lecture 17 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 15: Canonical Research Designs III: Instrumental Variables I
Просмотров 3,2 тыс.3 года назад
Lecture 15 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 16: Canonical Research Designs IV: Instrumental Variables II
Просмотров 2,4 тыс.3 года назад
Lecture 16 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 14: Canonical Research Designs II: Event Studies, Synthetic Control + Synthetic DinD
Просмотров 3,7 тыс.3 года назад
Lecture 14 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd/blob/main/lectures/14_synthetic_dind.pdf Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 13: Canonical Methods I: Difference-in-Differences
Просмотров 5 тыс.3 года назад
Lecture 13 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd/blob/main/lectures/13_dind.pdf Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 12: Likelihood Methods IV: Hierarchical Models + Bayesian Shrinkage
Просмотров 2,2 тыс.3 года назад
Lecture 8 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd/blob/main/lectures/12_hierarchical_bayes.pdf Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed!
Lecture 11: Likelihood Methods III: Duration Models
Просмотров 2 тыс.3 года назад
Lecture 11 from my Applied Metrics PhD Course. Materials here: github.com/paulgp/applied-methods-phd/blob/main/lectures/11_duration_models.pdf Please note that I've cut out questions from students, so some of the cuts in the video might feel slightly disjointed! In this video, I had to cut the end portion because I was taking care of my son simultaneously.
Lecture 10: Likelihood Methods II: Multiple Discrete Choices
Просмотров 1,2 тыс.3 года назад
Lecture 10: Likelihood Methods II: Multiple Discrete Choices
Lecture 9: Likelihood Methods 1: Discrete Choice, GLM and Computational Methods
Просмотров 1,6 тыс.3 года назад
Lecture 9: Likelihood Methods 1: Discrete Choice, GLM and Computational Methods
Lecture 7: Linear Regresion III: Quantile Regression
Просмотров 9 тыс.3 года назад
Lecture 7: Linear Regresion III: Quantile Regression
Lecture 8: Linear Regression IV: Penalized Regression
Просмотров 2 тыс.3 года назад
Lecture 8: Linear Regression IV: Penalized Regression
Lecture 5: Linear Regression I: Inference
Просмотров 3,6 тыс.3 года назад
Lecture 5: Linear Regression I: Inference
Lecture 4: Interference, Spillovers, and Dynamics
Просмотров 3,6 тыс.3 года назад
Lecture 4: Interference, Spillovers, and Dynamics
Lecture 1: Potential Outcomes and Directed Acyclic Graphs
Просмотров 19 тыс.3 года назад
Lecture 1: Potential Outcomes and Directed Acyclic Graphs
Lecture 2: Research Design, Randomization and Design-Based Inference
Просмотров 6 тыс.3 года назад
Lecture 2: Research Design, Randomization and Design-Based Inference
Lecture 6. Linear Regression II: Semiparametrics and Visualization
Просмотров 4,5 тыс.3 года назад
Lecture 6. Linear Regression II: Semiparametrics and Visualization
amazing lectures
Hey, i'm in the 12th grade and i'm very interested in what you were talking about but majority of it didn't make sense .. could you summarize?
loved it, had some doubts and this series turned out super useful
Glad to hear it!
Thank you so much for all these incredible videos!
Thanks from Pakistan
Love it!
I foun this while scrolling on Twitter. This is exactly what I needed for my class!! Thank you. You gained a new subscriber!!
Is lecture 18 available? Cannot find it in your uploads
52:55 Bailey and Goodman-Bacon
Thank you professor!
That’s really helpful for us students. And where is the lecture 18? :)
Definitely a work of art!
Thanks god for allowing us humans this kind of contents watchable at home
Thank you!!!
could you please make a video about your paper Bartik iv in AER please
I was reading your newly published AER paper on Bartik instrument, and I tried to find one video presentation, and today I just found your channel. Thank you so much.
Is duration model same as survival model?
Thank you a ton for making these public! Could you please post lecture 18? I think it’s missing
Going to rerecord it -- I didn't like the lecture I gave, and it's important for me to give a good lecture on Bartik IV b/c it's part of my research portfolio.
@@paulg-p Looking forward to it! Thanks a lot for sharing!
@@paulg-p Plz Plz Plz !! I need Bartik IV desperately !! 😭😭😭
@@paulg-p Hi Professor, is the Bartik lecture will come up anytime soon? Thank you
@@paulg-p Dear professor, can we have an idea on when do you plan to post the Bartik lecture? Thanks a lot!!
On your slide of SUPD, by 'with logit errors, new products do not crowd the original market,' do you mean that logit erorr always incorporate some external margin of the market (substitution from the outside good) as opposed to preserving the internal share of the market? Similar to what the "pure characteristics model" tried to address..?
Yes, that's what I meant! Conditional on the observables, the substitution patterns are proportional across all goods.
Highlights of the video: 1.Definitely bring up Hodges' estimator at parties 2. the Cher-vengers. But jokes aside, can the control function approach be framed as a problem of nuisance parameter?
Hi Paul. What an amazing course! Will you cover machine learning? Cheers.
Thanks a lot! I want to get those videos up but lost one of the best ones on ML. I will likely try to get them up in the spring.
@@paulg-p Oh that's a shame :( . Do you mean spring 2022? Sorry, but here, in south hemisphere, spring is in september 23th. Please, tell me that you are in the south hemisphere too :)
Amazing course! I can’t wait to listen to your lectures on machine learning. Maybe you can film a short vid? If not the entire lecture? I find your teaching so inspiring
Hi Professor, is lecture 18 absent?
Hi, yes, unfortunately I somehow lost 18 and the second ML lecture. I'm going to try to rerecord in the spring when I teach the course again. (Or potentially this November for lecture 18).
@@paulg-p Great! Learn a lot from your lecture!!
31:35 Slide 14 has been updated in the github page which explicitly constrains P(D) s.t E[D_i/(D_1+...+D_n)]=1
Thanks for this incredible course Paul! It has been very useful, especially the Diff-Diff TWFE part. Are you planning to add the remaining videos at some point? (thx also for posting the slides!)
Excellent lecture. Thanks.
Your classes are great Paul! If you can, please put some more here on RUclips =)
We are getting close to Bartik \o/
Great lecture, prof. ! I have a question on more general forms of censoring and wanted to know if you can point me to any resource on that. Particularly, I have two examples which bother me on a daily basis. 1. I want to estimate the effect of treatment T on outcome Y on the long term profitability of a customer (say 12 or 24 months). I don't observe long term profitability for the recent cohorts, but I think they often contain valuable information (specially in crazy times like COVID). How can I include them in my estimation? 2. When estimating the effect of credit limit T on credit card spending, Y, spending is censored by the credit limit. I know about the tobit model can be useful with censoring, but how can I account for that kind of censoring where there isn't one censoring threshold, but one threshold for each individual?
Thank you for the lecture Prof! Super interesting, specially the part on visualisation! Prof., what is the relationship between binscatter and ML? It looked to me like bin-scatter is just a flexible way to estimate the CEF even if it is non linear. In that spirit, can't you use any ML function approximation to display the relationship between the treatment and the outcome (or bayesian ML if you confidence intervals)?
Hi Matheus, yes, for sure -- there's a clear connection between estimating E(y|X) (the conditional expectation) and binscatter. The trick is always how to visualize the "right" relationship. Say there are two covariates: x_1 and x_2. If you allow E(y|x1,x2) = g(x1,x2) to be fully flexible, it's challegning to visualize the relationship between x1 and y --- it really depends on your choice of x2 (this is true of logit, for example). What the binscatter models tend to assume is some additive linearity -- E(y|x1,x2) = g(x1) + x2*beta, such that you can residualize the effect of x2, and then just plot the relationship between x1 and y. The "On Binscatter" paper talks about this in more detail, but that's the crux of the main difference between more general flexible models and binscatter style approaches.
Thanks a lot for sharing this material. It is super helpful and so generous!!
Thank you for your sharing this lecture video. Can you upload more sessions from the lecture 14 to 26?
Can se say that X is proxy for D used in the DAGs?
Funny how your hair gets messier and messier throughout these videos.
Thank you so much for this. Really helps. I was wondering if you would upload the rest of the lectures?
thanks for sharing those materials,you are so kind
Thanks for sharing, Paul. I will try to follow the whole course. Best regards from Brazil.
It is absolutely fantastic that you shared this. It is super helpful for Ph.D students everywhere, even in top programs. The videos cleared some doubts that I have been carrying for a while as my econometric classes were not so intuitive. The audio and video quality is perfect.
I'm so glad to hear that! Glad it was helpful. One unexpected silver lining of having to do everything online this year.
hooray
Best Applied Econometrics Course Ever!
Thanks a lot, your lectures are great! I was wondering if you eventually plan on publishing the whole course?
Hi, yes, I need to get them up! They will be up eventually.
@@paulg-p great news! Thanks again :)
Thank you so much for sharing this. I watched all of them and learned a lot. Will there be more videos? Many thanks!
Hi, yes, I need to get them up! They will be up eventually.
Super cool lectures
Will there be more content coming soon =) ?
Yes!
Lp norm is summing up the components to the p power and then taking the p-th root
Yes, I did a terrible job explaining it on the spot!
@@paulg-p still really cool stuff, very interesting to hear about this material from economics perspective
So, what assumptions are necessary for identifying the effect of just one dose? That was such a cliffhanger hahaha Also, thank you so much for uploading these classes. I already saw three and I pretend to see all. Your explanations are great =)
Hah! Good question. I haven't solved it out formally, but you would need to pick some assumptions. The most natural one might be monotonicity -- e.g. the effect of the first dose won't go down in the future. Then the lower bound for the one dose regime is just the efficacy right before the second dose is given. One upper bound is just the 2 dose efficacy. That'll give you a pretty decent range.
Thank you so much!
I don’t know how to be thankful for this! I’ve been a fan of your papers and now this. This teaches me a lot. Appreciate so much 👍👍👍🙏🙏🙏💕💕💕👏👏👏
Thank you so much for sharing these. Have been visiting the materials you shared on Twitter and these are really great support to the slides. More power and blessings!
Glad you like them!
what is the account on twitter 😂
Thanks for sharing these, great lectures!
Glad you like them!
Dear Prof. Is it possible to get access to the recordings of the last five lectures too? I am very keen to go through the slides and other teaching material as well. Looking forward to your response. Regards Gaurav
I'm slowly putting them up, 1-4 are now available.
@@paulg-p Thank you so much.