Propensity Score Matching using STATA: Simple explanation and step by step estimation
Вставка
- Опубліковано 2 чер 2023
- This video shows how to use the STATA software to estimate The Propensity Score mMatching. It covers the concept in a very simple explanation.
Excellent! this is the most simplest explanation PSM i have ever seen. Thank you sir
This was great, Hamayooon. Thank you!
Very helpful video. thanks a lot
very helpful!! thank you so much
Great job
Good job
Informative sir
Thanks dear
Very educational video! How do I evaluate the quality of the match and is there a way I can extract the matched observations/pairs?
Excellent could you make a video without using those codes. Just by using the GUI
Excellent! I read papers on PSM, and they have a curve to show the "common area" between treatment and control. Do you mind sharing how to get that? How to evaluate the balance.
I will share soon. And Thanks for your appreciation
Thanks for this excellent video! I am trying to find a way to present my PSM scores in an article, but I can't find a good example how to do this. Any idea how to present findings in a table? Thanks again!
Am facing the same problem. Have you gotten any idea so far?
Please how do we present the ATT, ATE results in word? how do we construct or outreg the tables?
Hello, I am confused about the unbalance in the blocks, I am having this warnings on my code
thanks but for some reason i get "command pscore is unrecognized". how do i get round this? i already have psmatch2 installed
could you share the data for exercise
Thanks, would like to ask you, the pscore and psmatch2 are they two separate matching method? or should we do both for impact analysis?
@@Dhanya690 both are the same
@@hamayoonshah1990 thank u sir, so do we need to use both? or any one of them?
@@hamayoonshah1990 i used Pscore command, can we do balancing test if we opt pscore rather psmatch 2 command?
In nnm results, is 380 treated grps matched with 437 control grps? May I know how many total matched pairs is? 380 or? I'm really confused, thank you in advance for your answers.
I means 380 from treatment are matches with all the control. One treatment may be matched with multiple control at a time
@@hamayoonshah1990 Thank you so much for your reply. I understand now!
I got n. treat. 46536 and n. contr. 1711 , but still confused about how many the matched data? And how to drop the matched data and just keep the matched data cause I want to do logistics reg to analysis. Also is there a way to evaluate the match graphically and is there a way to present my PSM scores in an article.
do not drop the unmatched and just follow the steps, it will get to the ATET. As far as reporting the Pscores is concerned, so let me make it clear that we don't report them because our concern is only to reach to the ATET
Thank you for your reply. How about descriptive statistics for the matched sample?@@hamayoonshah1990
@@hamayoonshah1990 why stata says command pscore is unrecognized??can you help
@@bushrazaman4944 install the package. go to help pscore and install pscore package sr00-2
Actually, very nice lecture I take more awareness about PSM but your video quality is less so other time please improve it
thank u so much , could you please share this data ?
could you please provide this data (csv) link ?
Comment
Hello Sir, i followd all your steps one by one, in the end it give an error r(2000) no observatios.Please help solve this problem.
How many observations do you have? May b very few observations with no matches
@@hamayoonshah1990 i got the solution, my problem was that for the traning varibale i used alphabets 'Yes'/'No', i changed that to '0'/'1' it worked. Thanks for the analysis though.
Sir can you please share the data set
Yes let me put it in the description
I need your help can we connect over WhatsApp?