dear all,
How to understand the difference between OLS regression and Quantile Regression more intuitively without referring to mathematical notations and theorems.
thanks in advance
Please see this link which addresses the difference between
them:
https://www.google.com/url?sa=t&source=web&rct=j&url=https://stats.stackexchange.com/questions/39002/when-is-quantile-regression-worse-than-ols&ved=2ahUKEwjovtn_7dfnAhWRDOwKHb30D5YQFjACegQIDhAG&usg=AOvVaw1YBnZSNva2Zmpcc5Ag7Mna&cshid=1581917577905
Dear Srikanth Potharla ,
Imagine that you want to evaluate the effect of ethnicity on income in India. If you are interested in the effect of ethnicity on the average income of the whole population in the country, you can use an OLS regression of income against ethnicity (and of course control covariates).
If you are interested in the effect of ethnicity on the income of the poor population, you will run a quantile regression of income against ethnicity (again with control covariates) for the lower tail (the first quantiles) of the income distribution in India.
If you are interested in the effect of ethnicity on the income of the rich population, you will run a quantile regression of income against ethnicity (again with control covariates) for the upper tail (the last quantiles) of the income distribution in India.
That's the overall idea.
If you are interested in the mean, use OLS, if in the median, use quantile.
- Badges
- Science topic