ANKETEST is unable to find the matrix

[Chiara Pollio]

Dear Statalists,
I am trying to run the Anselin-Kelejian test for assessing spatial autocorrelation in my dataset. I have already generated a spatial matrix through spmat command, and I was able to save it as a .dta file and use it to run the Moran's I test through the spatgsa command. However, I am not sure I fully get the functioning of the anketest command, which I am trying to use to run the autocorrelation testl. I understand it is a post-estimation command, so I run a simple ols to do the test, but the command does not seem to find the matrix. I am not very familiar with working with matrices on Stata, but I think that the matrix is in the programme, and it is anyway stored as a .dta file in my working directory.
Can anybody help me to get the command to find the matrix? Here I attach the series of commands and part of the output that I run from the formation of the matrix till the error messages that the system reports after using ANKETEST.

Thank you!
Chiara

*I build the matrix
spmat contiguity Wmm using gdxy, id(id) replace normalize(minmax)
* I save it as a .dta file
spmat export Wmm using "Wmm.txt", noid replace
insheet using "Wmm.txt", delim(" ") clear
drop in 1
save "Wmm.dta", replace

*I use it to run the Moran's I test..
.
use "gddatabase.dta", clear
spatwmat using "Wmm.dta", name(Ws) 
spatgsa ln_output, weights(Ws) moran

*... and it works
Moran's I
--------------------------------------------------------------
          Variables |    I      E(I)   sd(I)     z    p-value*
--------------------+-----------------------------------------
          ln_output |  0.557  -0.012   0.077   7.346   0.000
--------------------------------------------------------------
*1-tail test

*I run the regression:
reg ln_output dummy_zones developmentzones nst ln_inv_t3 ln_fdi_t3 ln_work prd coast urban, robust
(output omitted)

*and here is the problem with the ANKETEST

anketest, wname(Wmm) wfrom(Stata) model(ols)
(11 missing values generated)
  
Matrix Wmm not found
r(601);

 anketest, wname(Wmm) wfrom(Mata) model(ols)
(11 missing values generated)

[Emad Shehata]

use the commands as follows:

clear all
use http://fmwww.bc.edu/repec/bocode/s/spregcs.dta, clear
regress y x1 x2
spatwmat using http://fmwww.bc.edu/repec/bocode/s/SPWcs.dta, name(W)
anketest, wname(W) model(ols) wfrom(Stata)

. clear all
. use http://fmwww.bc.edu/repec/bocode/s/spregcs.dta, clear
. regress y x1 x2

      Source |       SS       df       MS              Number of obs =      49
-------------+------------------------------           F(  2,    46) =   28.39
       Model |  7423.33055     2  3711.66528           Prob > F      =  0.0000
    Residual |  6014.88661    46  130.758405           R-squared     =  0.5524
-------------+------------------------------           Adj R-squared =  0.5329
       Total |  13438.2172    48  279.962858           Root MSE      =  11.435

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          x1 |  -.2739317   .1031986    -2.65   0.011    -.4816598   -.0662037
          x2 |  -1.597311   .3341306    -4.78   0.000    -2.269881   -.9247407
       _cons |   68.61898   4.735484    14.49   0.000     59.08694    78.15101
------------------------------------------------------------------------------

. spatwmat using http://fmwww.bc.edu/repec/bocode/s/SPWcs.dta, name(W)

The following matrix has been created:

1. Imported binary weights matrix W
   Dimension: 49x49

. anketest, wname(W) model(ols) wfrom(Stata)
 
Diagnostic tests for Spatial Dependence
 
Tests for Spatial Autocorrelation in the OLS Model residuals

+----------------------------------------+
|            Tests |  Statistic  P-Value |
|------------------+---------------------|
|        Moran's_I |       1.96   0.0503 |
|         LM_Error |       2.35   0.1255 |
|  Robust_LM_Error |       1.41   0.2351 |
|           LM_Lag |       1.12   0.2891 |
|    Robust_LM_Lag |       0.19   0.6663 |
+----------------------------------------+

[Chiara Pollio]

Dear Emad,
Thank you, it worked.
C

[Emad Shehata]

another second way is:

clear all
spatwmat using http://fmwww.bc.edu/repec/bocode/s/SPWcs.dta, name(W)
use http://fmwww.bc.edu/repec/bocode/s/spregcs.dta, clear
regress y x1 x2
spatdiag, weights(W)

. clear all

. spatwmat using http://fmwww.bc.edu/repec/bocode/s/SPWcs.dta, name(W)

The following matrix has been created:

1. Imported binary weights matrix W
   Dimension: 49x49

. spatwmat using SPWcs.dta, name(W)


The following matrix has been created:

1. Imported binary weights matrix W
   Dimension: 49x49

. use http://fmwww.bc.edu/repec/bocode/s/spregcs.dta, clear

. regress y x1 x2

      Source |       SS       df       MS              Number of obs =      49
-------------+------------------------------           F(  2,    46) =   28.39
       Model |  7423.33055     2  3711.66528           Prob > F      =  0.0000
    Residual |  6014.88661    46  130.758405           R-squared     =  0.5524
-------------+------------------------------           Adj R-squared =  0.5329
       Total |  13438.2172    48  279.962858           Root MSE      =  11.435

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          x1 |  -.2739317   .1031986    -2.65   0.011    -.4816598   -.0662037
          x2 |  -1.597311   .3341306    -4.78   0.000    -2.269881   -.9247407
       _cons |   68.61898   4.735484    14.49   0.000     59.08694    78.15101
------------------------------------------------------------------------------

. spatdiag, weights(W)

Diagnostic tests for spatial dependence in OLS regression


Fitted model
------------------------------------------------------------
y = x1 + x2
------------------------------------------------------------

Weights matrix
------------------------------------------------------------
Name: W
Type: Imported (binary)
Row-standardized: No
------------------------------------------------------------

Diagnostics
------------------------------------------------------------
Test                           |  Statistic    df   p-value
-------------------------------+----------------------------
Spatial error:                 |
  Moran's I                    |     0.645      1    0.519
  Lagrange multiplier          |     2.348      1    0.125
  Robust Lagrange multiplier   |     1.410      1    0.235
                               |
Spatial lag:                   |
  Lagrange multiplier          |     1.124      1    0.289
  Robust Lagrange multiplier   |     0.186      1    0.666
------------------------------------------------------------

[Emad Shehata]

Also, another third way is:

clear all
use http://fmwww.bc.edu/repec/bocode/s/spregcs.dta, clear
spregcs y x1 x2 , model(ols) run(ols) lmsp wmfile(http://fmwww.bc.edu/repec/bocode/s/SPWcs.dta)

==============================================================================
*** Binary (0/1) Weight Matrix: 49x49 (Non Normalized)
------------------------------------------------------------------------------
==============================================================================
* Non Spatial Regression: Model(ols) - Run(ols)
==============================================================================
  y = x1 + x2
------------------------------------------------------------------------------
 Sample Size        =          49
 Wald Test          =     56.7713   |   P-Value > Chi2(2)       =      0.0000
 F-Test             =     28.3857   |   P-Value > F(2 , 46)     =      0.0000
 R2  (R-Squared)    =      0.5524   |   Raw Moments R2          =      0.9186
 R2a (Adjusted R2)  =      0.5329   |   Raw Moments R2 Adj      =      0.9151
 Root MSE (Sigma)   =     11.4350   |   Log Likelihood Function =   -187.3772
------------------------------------------------------------------------------
- R2h= 0.5524  R2h Adj=  0.5329  F-Test =   28.39 P-Value > F(2 , 46)  0.0000
- R2v= 0.5524  R2v Adj=  0.5329  F-Test =   28.39 P-Value > F(2 , 46)  0.0000
------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          x1 |  -.2739317   .1031986    -2.65   0.011    -.4816598   -.0662037
          x2 |  -1.597311   .3341306    -4.78   0.000    -2.269881   -.9247407
       _cons |   68.61898   4.735484    14.49   0.000     59.08694    78.15101
------------------------------------------------------------------------------

==============================================================================
*** Spatial Aautocorrelation Tests                         (Model= ols)
==============================================================================
  Ho: Error has No Spatial AutoCorrelation
  Ha: Error has    Spatial AutoCorrelation
------------------------------------------------------------------------
- GLOBAL Moran MI            =   0.1410     P-Value > Z( 1.907)   0.0565
- GLOBAL Geary GC            =   0.7343     P-Value > Z(-2.041)   0.0412
- GLOBAL Getis-Ords GO       =  -0.6793     P-Value > Z(-1.907)   0.0565
------------------------------------------------------------------------------
- Moran MI Error Test        =   0.6445     P-Value > Z(7.838)    0.5192
------------------------------------------------------------------------------
- LM Error (Burridge)        =   2.3476     P-Value > Chi2(1)     0.1255
- LM Error (Robust)          =   1.4096     P-Value > Chi2(1)     0.2351
------------------------------------------------------------------------------
  Ho: Spatial Lagged Dependent Variable has No Spatial AutoCorrelation
  Ha: Spatial Lagged Dependent Variable has    Spatial AutoCorrelation
------------------------------------------------------------------------
- LM Lag (Anselin)           =   1.1240     P-Value > Chi2(1)     0.2891
- LM Lag (Robust)            =   0.1860     P-Value > Chi2(1)     0.6663
------------------------------------------------------------------------------
  Ho: No General Spatial AutoCorrelation
  Ha:    General Spatial AutoCorrelation
------------------------------------------------------------------------
- LM SAC (LMErr+LMLag_R)     =   2.5336     P-Value > Chi2(2)     0.2817
- LM SAC (LMLag+LMErr_R)     =   2.5336     P-Value > Chi2(2)     0.2817
------------------------------------------------------------------------------

[Bastian Longer]

Hello,

When using anketest in a panel data setting I get an error message:

anketest, wname(W) model(ols) wfrom(Stata)

classdef _b_stat() in use
(nothing dropped)
(160 lines skipped)
(error occurred while loading anketest.ado)
r(310);

Can anyone help me? Or tell another way how to test which case of spatial dependence is present?

Thank you