Add confidence intervals to median

[Buyadaa Oyunchimeg]

Dear Stata users,

I would like to add 95% confidence intervals to q50. But it does not appear to be an option of logistic quantile regression.


Codes look as below:

mkspline uacr_ = uacr, nknots(3) cubic

lqreg re_slope1 uacr_1 uacr_2, quantiles(50) nodots seed(123)

lqregpred adjs, for(uacr_1 uacr_2) plotvs(uacr) ci 

Could anyone help me.

Thanks a lot.

Oyun

[Rich Goldstein]

since you don't supply any data, I can't see what is happening in your case; however, using the example given in the help file, I do see 95% CI's; I suspect we are not talking about the same thing; please read the FAQ on how to ask questions and how to post results and point out what is missing

[Marcos Almeida]

If I understood right, you need to use the - centile - command, like in the toy example below:

. sysuse auto
(1978 Automobile Data)

. sum mpg, detail

                        Mileage (mpg)
-------------------------------------------------------------
      Percentiles      Smallest
 1%           12             12
 5%           14             12
10%           14             14       Obs                  74
25%           18             14       Sum of Wgt.          74

50%           20                      Mean            21.2973
                        Largest       Std. Dev.      5.785503
75%           25             34
90%           29             35       Variance       33.47205
95%           34             35       Skewness       .9487176
99%           41             41       Kurtosis       3.975005

. centile mpg

                                                       -- Binom. Interp. --
    Variable |       Obs  Percentile    Centile        [95% Conf. Interval]
-------------+-------------------------------------------------------------
         mpg |        74         50          20              19          22

Hopefully that helps.

[Nick Cox]

I fear that the question is about a community-contributed command


SJ-12-3 st0231_1 . . . . Software update for lqreg, lqregpred, and lqregplot
. . . . . . . . . . . . . . . . . . . . . . . N. Orsini and M. Bottai
(lqreg, lqreg_postestimation, lqregpred, lqregplot if installed)
Q3/12 SJ 12(3):570
cluster() option has been fixed

SJ-11-3 st0231 . . . . . . . . . . . . Logistic quantile regression in Stata
. . . . . . . . . . . . . . . . . . . . . . . N. Orsini and M. Bottai
(lqreg, lqreg_postestimation, lqregpred, lqregplot if installed)
Q3/11 SJ 11(3):327--344
estimation, prediction, and graphical representation of
logistic quantile regression

-- which Buyadaa really is expected to explain more fully, rather than assume we all recognise the code.

[Buyadaa Oyunchimeg]

Dear all, thank you so much for your reply and sorry about unclear post.

I fear that the question is about a community-contributed command


SJ-12-3 st0231_1 . . . . Software update for lqreg, lqregpred, and lqregplot
. . . . . . . . . . . . . . . . . . . . . . . N. Orsini and M. Bottai
(lqreg, lqreg_postestimation, lqregpred, lqregplot if installed)
Q3/12 SJ 12(3):570
cluster() option has been fixed

SJ-11-3 st0231 . . . . . . . . . . . . Logistic quantile regression in Stata
. . . . . . . . . . . . . . . . . . . . . . . N. Orsini and M. Bottai
(lqreg, lqreg_postestimation, lqregpred, lqregplot if installed)
Q3/11 SJ 11(3):327--344
estimation, prediction, and graphical representation of
logistic quantile regression

Mr.Cox you are right. I have used lqreg and lqregpred commands for my analyses. I am wondering, is it possible to plot median and mean using lqpred command (instead of 95% CI)?

Thank you again.

[Nick Cox]

Sorry, I have never used any of these excellent commands. But my guess has to be that quantile regression is not in any sense focused on predicting the mean.

[Buyadaa Oyunchimeg]

Dear Mr.Cox,

Thank you so much for your prompt reply.
If so is there anyway to plot both median mean after cubic splines and quantile regression (lqreg re_slope1 uacr_1 uacr_2, quantiles(50) nodots seed(123)) (not using lqpred)?

Thank you

[Nick Cox]

Again, I haven't used any of these programs and I really don't know. Unless no-one else answers your best chance is to contact the program authors directly.