Stata Tutorial: Fixing Heteroskedasticity in OLS
Вставка
- Опубліковано 24 лис 2018
- Basic methods of mitigating the effect of a heteroskedastic error in a simple OLS setting. We look at respecification, Weighted Least Squares, and the White Robust Standard Error options.
Link to "Gentle Introduction to Stata"
www.amazon.com/gp/product/159...
Link to the excellent Introduction to Econometrics Textbook by AH Studenmund:
www.amazon.com/gp/product/933...
Link to Jeffrey Wooldridge Introductory Econometrics Textbook:
www.amazon.com/gp/product/813...
My Twitter is:
/ michaelrjonas
My Google Scholar Page:
scholar.google.com/citations?...
ResearchGate:
www.researchgate.net/profile/...
made me laugh when you said 'i have a lot of time in my'. While people use their free time entertaining themselves, you spent it by helping thousands. Thanks a lot, Mike. you've been my only ray of hope these past few months
it is fantastic for me to learn more about how to use stata.
i will definitely to watch all of these videos and cant wait to practice it.
thank you so much !!sir!
keep update plz!
Hey Mike, thank you for your helpful explanation!
Helped me alot in my seminarwork,
thanks and best wishes!
Thank you so much! I plan to watch all your videos - they are great!
Thank you-glad they are helpful!
Nice explanation, sir! thanks so much
Thanks very much for your valuable effort
which approach is better to deal with heteroskedasticity? WLS or robust? Is there any indication in the estimates as to which approach is better?
This is a great question. If you have a solid candidate for the weight to be used, WLS will give more efficient (lower variance) results, and is preferred. You can also estimate an optimal weight using the "feasible" WLS method. However, when the form of het is unknown, which is most of the time, using robust standard errors is preferred (and easy!)
thank you so much!!!!!! life saver
You're welcome!
First of all Mike, you are the real MVP. Would it be possible to contact you for private tutoring for econometrics? IF not could you please provide some help on how to solve for Heteroskedasticity for panel data? Particularly analysis on bank profitability. Thank!
Hi Professor Mike, thank you so much for your informative video! I need to kindly ask if "robust" can be used for panel data too? And if it can be put before or after "cluster(x)" to fix both heteroskedasticity and auto-correlation in panel data -OR "cluster (x)" alone is enough to fix both problems?
Hi, first of all, thanks for the great video, it's really helped me a lot. However, I wonder whether you could provide any literature concerning the formulas that can be used to calculate the different 'types' of standard error (normal vs. robust) 19:57 -
So, did robust address the issue on heteroskedasticity?
Yes, the robust standard error calculation corrects the variance, standard error, and t statistic for the presence of a heteroskedastic error. However, unlike WLS, it does not create a homoskedastic error.
Thank you sir, but i didn't understand a point: when we do the breusch-pagan test with f-statistic, do we reject null hp when p-value is LESS than .05?
Yes, correct. If F-stat is greater than 95% critical value, then p
noconstant code is really cool
Many thanx for is this video, stay blessed..
I am facing an issue please guide me. I have only hetro issue in my panel data, hasuman selected fixed effect model. But due to hetroskedicity issue I have to apply robust, it is showing all the variables are insignificant.. please guide me
I did the same process but after gen new variables like Wages_W and so on....my Education variable is showing omitted and giving a reason is that collinarity. why?
from where did you get aw=1/exp?
can you really just log transform only your y variable? Doesn't this have implications in your model since you need to antilog for the results, thank you great work you are doing
My model suffers from heteroskedasticity but whenever I log the dependent variable (price) the problem is solved yet my model loses almost half of its explanatory power (I'm an absolute noobie at stats) anyone got any idea of what could be going on?
where did the eststo command come from?
You will have to type "ssc install estout" first in order to use this. Very useful in formatting and exporting tables and results. See here for details and examples:
ua-cam.com/video/-BAUD9UD6UQ/v-deo.html
The tutorial is good but lengthy, meanwhile, thanks so much, sir.
The video quality is really poor and quite troublesome..with jarring distortions and bizarre images interruptions..😕 Could have done better Sir