rqdeco rqdest(): 3200 conformability error

[Sabrina Weller]

Hi,
I am trying to run a simple rqdeco command:

rqdeco lnBrutto99 herstellen99, by(Beh_99) quantile(0.5)

I get the error term:

rqdest(): 3200 conformability error
<istmt>: - function returned error

I think the problem has to do with the default of the matrix...
The matrix that I need is 9 x 2

Do you have any idea how to solve the problem?
Thank you very much!

[Stephen Jenkins]

Welcome to Statalist. Help readers to help you and want to help you -- please follow the advice in the forum FAQ. So, ... (1) tell us where rqdeco, a user-written program, comes from. My search rqdeco yields no results. (2) Tell us something about your data and what you're trying to do. (We have no idea whether your remarks about matrix dimension refer to your data, your variable list, or program requirements.) (3) For legibility and other reasons, post your output using CODE delimiters (see the FAQ section 12 -- it's easy to do). And, (4), please re-register using your real name (firstname lastname). Again, this is easy to do: hit the Contact Us button at bottom right of screen and make your request. Thank you.

[Konrad Zdeb]

I think it may be this: http://www.econ.brown.edu/fac/Blaise...de_rqdeco.html

[Sabrina Weller]

Hi,

thank you for your replies.

(1) I am using the rqdeco user written demand by Blaise Melly in stata 12. Rqdeco decomposes differences in distribution using quantile regression. This decomposition is very similar to the Machado and Mata (2005) decomposition. http://www.econ.brown.edu/fac/Blaise...de_rqdeco.html

(2) I am working with the representative german BIBB/BAuA working population survey of the year 1999.
I am trying to do a decomposition analysis where I compare the median difference in lwage between disabled and non disabled employees in germany. I want to control for tasks items performed by the employees. In this model I am controlling the variable "producing" =(hestellen99)
my dependent variable is the hourly log wage = lnBrutto99.

The descriptive results of these variables are:

Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
herstellen99 | 11747 .2303567 .4210791 0 1


Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
lnBrutto99 | 11767 2.392273 .4381738 .0077594 3.984321


My comand is:

rqdeco lnBrutto99  herstellen99, by(Beh_99) quantile(0.5)

and the error term I get is:

                rqdest():  3200  conformability error
                 <istmt>:     -  function returned error
r(3200);

Stata tells me that the error which occured is

 a conformability error: A matrix, vector, or scalar has the wrong number of rows and/or
columns for what is required.  Adding a 2 x 3 matrix to a 1 x 4
would result in this error.

I hope I could explain my problem more precisely
Thank you very much

[Blaise Melly]

I have written this command. Please use cdeco instead of rqdeco. I am no longer updating rqdeco and cdeco has more capabilities. You can find how to install cdeco on this webpage: http://www.econ.brown.edu/fac/Blaise...e_counter.html
Send me an email with an extract of your data if the problem persists with cdeco. (blaise.melly at vwi.unibe.ch)
Thanks,
Blaise

[Biswa bhusan]

Dear Professor, I have used both rqdeco and cdeco for decomposition on one sample. The codes are:
generate female=(uniform()<0.5)
generate experience=4*invchi2(5,uniform())*(1-0.2*female)
generate lwage=2+experience*0.03-0.1*female +invnormal(uniform())*(0.6-0.2*female)


and results found:
rqdeco lwage experience, by(female) quantile(0.5)

Decomposition of differences in distribution using quantile regression
Total number of observations 1000
Number of observations in group 0 511
Number of observations in group 1 489

Number of quantile regressions estimated 100

The variance has not been computed.
Use the option vce if you want to compute it.

-----------------------------------------------------------------------------
Component| Effects Std. Err. t P>|t| [95% Conf. Interval]
-----------------+-----------------------------------------------------------
Quantile .5 |
Raw difference| -.31005 . . . . .
Characteristics| -.16811 . . . . .
Coefficients| -.14194 . . . . .
-----------------------------------------------------------------------------

Similarly, I used
cdeco lwage experience, g(female) quantile(0.5) noboot

Conditional model linear quantile regression
Number of regressions estimated 100

The variance has not been computed.
Do not turn the option boot off if you want to compute it.

No. of obs. in the reference group 511
No. of obs. in the counterfactual group 489


Differences between the observable distributions (based on the conditional model)
---------------------------------------------------------------------------------
| Quantile Pointwise Pointwise Functional
Quantile | effect Std. Err. [95% Conf. Interval] [95% Conf. Interval]
------------+--------------------------------------------------------------------
.5 | .31005 . . . . .
---------------------------------------------------------------------------------


Effects of characteristics
---------------------------------------------------------------------------------
| Quantile Pointwise Pointwise Functional
Quantile | effect Std. Err. [95% Conf. Interval] [95% Conf. Interval]
------------+--------------------------------------------------------------------
.5 | .16811 . . . . .
---------------------------------------------------------------------------------


Effects of coefficients
---------------------------------------------------------------------------------
| Quantile Pointwise Pointwise Functional
Quantile | effect Std. Err. [95% Conf. Interval] [95% Conf. Interval]
------------+--------------------------------------------------------------------
.5 | .14194 . . . . .
---------------------------------------------------------------------------------

.
But My concerns are on the sign of respective coefficient. When i am changing code, the results changed with sign. Although, you have said to use "cdeco", but my request you please have a look once more on this issues, and suggest which to be used for decomposition.

[sergio scicchitano]

Dear prof. Melly,
I used your command many times in many papers, by citing your good work: how can I control for sample selection?
best
Sergio

[Jay Dev]

I have written this command. Please use cdeco instead of rqdeco. I am no longer updating rqdeco and cdeco has more capabilities. You can find how to install cdeco on this webpage: http://www.econ.brown.edu/fac/Blaise...e_counter.html
Send me an email with an extract of your data if the problem persists with cdeco. (blaise.melly at vwi.unibe.ch)
Thanks,
Blaise

Dear Professor Melly

I recently downloaded CDECO,CDECO_JMP and COUNTERFACTUAL ado files from your web page (by following the link https://sites.google.com/site/blaise...ons?authuser=0).

While using CDECO and CDECO_JMP without the noboot option, I am getting the mata error 3301 with following message

I request you to please help me in this regards

Best

Jay Dev
test_boot(): 3301 subscript invalid
ev_boot(): - function returned error
<istmt>: - function returned error

[Stephen Jenkins]

Have you installed moremata? ssc install moremata, replace

[Blaise Melly]

Dear Professor Melly

I recently downloaded CDECO,CDECO_JMP and COUNTERFACTUAL ado files from your web page (by following the link https://sites.google.com/site/blaise...ons?authuser=0).

While using CDECO and CDECO_JMP without the noboot option, I am getting the mata error 3301 with following message

I request you to please help me in this regards

Best

Jay Dev
test_boot(): 3301 subscript invalid
ev_boot(): - function returned error
<istmt>: - function returned error

Dear Jay Dev,
I am unable to replicate your error message. Do you get it if you run the examples in the do-file example_cdeco.do (in the zip file)? If not, is it possible to get a replicable example with this error message?
Thanks,
Blaise

[Jay Dev]

Have you installed moremata? ssc install moremata, replace

Hello Professor Jenkins,

Yes I did it before. the problem still persist for "CDECO_JMP"

Thanks for kind help.

[Jay Dev]

Have you installed moremata? ssc install moremata, replace

Hello Professor Jenkins,

Yes I did it before. the problem still persist for "CDECO_JMP"

Thanks for your kind help.

[Jay Dev]

Dear Jay Dev,
I am unable to replicate your error message. Do you get it if you run the examples in the do-file example_cdeco.do (in the zip file)? If not, is it possible to get a replicable example with this error message?
Thanks,
Blaise

Hello Professor Melly

Example do file is now running well, but the problem persist with "cdeco_jmp".I sent you an email with the detailing of my problem. I request you to please see that and share your response. It will be a grate help.

Regards

Jay Dev