hi
can I run a quantile regression on a cross sectional dataset of 100 firms only? My independent variables are 12...
hii
Quantile regression as a method for estimating functional relations between variables for all portions of a probability distribution. details on this stat test at the website https://www.google.jo/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=quantile%20regression
The method of least squares estimates,, which approximate the conditional mean of the response,, variable given certain values of the predictor variables, quantile regression aims at estimating either the conditional median or other quantiles of the response variable.
Now, the small sample can be used but the possiblity of Type II errors increase. you may dont get significant finding due to the small sample size. Thus, you need to consider this as a limitation, and give recommendation for next studies about this point.
hii
Quantile regression as a method for estimating functional relations between variables for all portions of a probability distribution. details on this stat test at the website https://www.google.jo/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=quantile%20regression
The method of least squares estimates,, which approximate the conditional mean of the response,, variable given certain values of the predictor variables, quantile regression aims at estimating either the conditional median or other quantiles of the response variable.
Now, the small sample can be used but the possiblity of Type II errors increase. you may dont get significant finding due to the small sample size. Thus, you need to consider this as a limitation, and give recommendation for next studies about this point.
58
One problem with quantile regression is that quantiles are empirical quantities. Your precision will depend both on the sample size and on the quantile you are modelling. Data are sparser at the extremes of the distribution, so modelling extreme quantiles like the 10th or 90th will have lower associated precision than modelling the median.
Muayyad's answer, copied from Wikipedia, is not clear on the distinction between least-squares regression and quantile regression, but the Wikipedia article itself is useful reading, and does explain the difference, further down the page.
https://en.wikipedia.org/wiki/Quantile_regression
of course the answers from other colleagues...went to the same idea which is called in research "Type II errors" and I have already defined it in text shown in my previous answer to this question.
Regards
Dear Saadia,
A small size will be a problem. As said before the estimated conditional quantiles are empirical, so they could be affected easily for small size samples. See in attach a figure that probably will help you to depict the problem.
Since quantile regression models the conditional quantile think we want to predict (for example) the quartiles of Y (given each value of X), a covariate. Look at extremes when the number of observations is smaller. Extreme quantiles also will be harder to estimate.
So a greater sample size is need in order to assure a good estimation for all levels of quantiles.
Good luck, :)
You are welcome Saadia,
Once you started the analysis and need further help, dont hesistate.
Good luck
Someone, please correct me if I am wrong. I am under the impression that quantile regression uses all of the observations at every quantile analyzed. For example at the 10%, all observations are still analyzed but with different weights on the residuals. Is this correct? If so, wouldn't this mean that there are just as many data points analyzed at the 10% as there are at the 50%?
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression