Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: fixed effect or random effect model


From   Caliph Omar Moumin <[email protected]>
To   "[email protected]" <[email protected]>
Subject   Re: st: fixed effect or random effect model
Date   Sun, 6 May 2012 04:58:34 -0700 (PDT)

Many thanks to you John, that was so helpful. 
 
Kind Regards,
Caliph Omar Moumin

Email:  [email protected] 



----- Original Message -----
From: John Antonakis <[email protected]>
To: [email protected]
Cc: 
Sent: Sunday, May 6, 2012 1:49 PM
Subject: Re: st: fixed effect or random effect model

Right. You can go with the random-effects model; also, that the 
Breusch-Pagan test is significant means that there is significant 
variance in uj (in the random-effects specification); i.e., uj is not 
zero. Thus, you can use a random-effects model.

Best,
J.

__________________________________________

Prof. John Antonakis
Faculty of Business and Economics
Department of Organizational Behavior
University of Lausanne
Internef #618
CH-1015 Lausanne-Dorigny
Switzerland
Tel ++41 (0)21 692-3438
Fax ++41 (0)21 692-3305
http://www.hec.unil.ch/people/jantonakis

Associate Editor
The Leadership Quarterly
__________________________________________


On 06.05.2012 13:34, Caliph Omar Moumin wrote:
> Thank you John
>
> You told me important info.
> i applied it and the result as shown below
> is Sargan-Hansen statistic   0.051  Chi-sq(1)    P-value = 0.8219.
> So i think this is same result as hausman test. Meaning that we failed to reject null (both fixed and random effect model are ok).
> Therefore in My case i want to choose random effect model. 
> if you think otherwise, could you please let me know? 
> does Breusch and Pagan Lagrangian multiplier test for random effects makes any change of my choice of random
>   based on Sargan-Hansen statistic? The result of Breusch and Pagan Lagrangian multiplier test is  
> chibar2(01) =    59.40;  Prob>  chibar2 =   0.0000.
>
> Thank you again John
>
>
> xtreg cost duration sex age group, re cluster(id)
>
> Random-effects GLS regression                   Number of obs      =     16718
> Group variable: id                              Number of groups   =     14503
>
> R-sq:  within  = 0.0392                         Obs per group: min =         1
>         between = 0.0535                                        avg =       1.2
>         overall = 0.0578                                        max =        16
>
>                                                  Wald chi2(4)       =    371.51
> corr(u_i, X)   = 0 (assumed)                    Prob>  chi2        =    0.0000
>
>                                   (Std. Err. adjusted for 14503 clusters in id)
> ------------------------------------------------------------------------------
>               |               Robust
>          cost |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
> -------------+----------------------------------------------------------------
>      duration |   944.5671    152.539     6.19   0.000     645.5962    1243.538
>           sex |  -4476.141   781.0165    -5.73   0.000    -6006.905   -2945.377
>           age |     306.88   20.33477    15.09   0.000     267.0246    346.7354
>         group |   4442.876   1727.691     2.57   0.010     1056.665    7829.087
>         _cons |   922.7695   3769.766     0.24   0.807    -6465.835    8311.374
> -------------+----------------------------------------------------------------
>       sigma_u |  40329.125
>       sigma_e |  26038.659
>           rho |  .70578153   (fraction of variance due to u_i)
> ------------------------------------------------------------------------------
>
>
>
>
> . xtoverid
>
> Test of overidentifying restrictions: fixed vs random effects
> Cross-section time-series model: xtreg re  robust cluster(id)
> Sargan-Hansen statistic   0.051  Chi-sq(1)    P-value = 0.8219
>
>
>
> xttest0;
>
> Breusch and Pagan Lagrangian multiplier test for random effects
>
>          cost[id,t] = Xb + u[id] + e[id,t]
>
>          Estimated results:
>                           |       Var     sd = sqrt(Var)
>                  ---------+-----------------------------
>                      cost |   2.27e+09       47647.13
>                         e |   6.78e+08       26038.66
>                         u |   1.66e+09       40752.23
>
>          Test:   Var(u) = 0
>                               chibar2(01) =    59.40
>                            Prob>  chibar2 =   0.0000
>
>
>  
> Kind Regards,
> Caliph Omar Moumin
>
> Email:  [email protected]
>
>
>
> ----- Original Message -----
> From: John Antonakis<[email protected]>
> To: [email protected]
> Cc:
> Sent: Sunday, May 6, 2012 10:31 AM
> Subject: Re: st: fixed effect or random effect model
>
> It would be more correct to say that if the p-value for the Hausman test, where you compare random vs fixed-effects, is<  .05 then the random-effects estimator is no good (i.e., the test is in the form "hausman fe re").  The fixed-effects estimator is consistent; however, the random-effects estimator is more efficient. If the estimates using random effects are not significantly different from the fixed-effects estimator (i.e., the p-value is>  .05) then you can retain the random-effects estimator.
>
> In your case, it would be best to use the user-written -xtoverid- test (available from SSC) after having run
>
> xtreg cost duration sex age group, re cluster(id_indicator)
>
> (id_indicator is your panel identifier)
>
> The xtoverid test accommodates a cluster robust xtreg vce. Specifically, it is a Hausman-type test that constrains the covariance between uj (the fixed-effect) and the regressors to zero.  See "help xtoverid": here is the relevant extract from the help file:
>
> "A test of fixed vs. random effects can also be seen as a test of overidentifying
> restrictions.  The fixed effects estimator uses the orthogonality conditions that the
> regressors are uncorrelated with the idiosyncratic error e_it, i.e., E(X_it*e_it)=0.
> The random effects estimator uses the additional orthogonality conditions that the
> regressors are uncorrelated with the group-specific error u_i (the "random effect"),
> i.e., E(X_it*u_i)=0.  These additional orthogonality conditions are overidentifying
> restrictions.  The test is implemented by xtoverid using the artificial regression
> approach described by Arellano (1993) and Wooldridge (2002, pp. 290-91), in which a
> random effects equation is reestimated augmented with additional variables consisting of
> the original regressors transformed into deviations-from-mean form.  The test statistic
> is a Wald test of the significance of these additional regressors.  A large-sample
> chi-squared test statistic is reported with no degrees-of-freedom corrections.  Under
> conditional homoskedasticity, this test statistic is asymptotically equivalent to the
> usual Hausman fixed-vs-random effects test; with a balanced panel, the artificial
> regression and Hausman test statistics are numerically equal.  See Arellano (1993) for
> an exact statement and the example below for a demonstration.  Unlike the Hausman
> version, the test reported by xtoverid extends straightforwardly to heteroskedastic- and
> cluster-robust versions, and is guaranteed always to generate a nonnegative test
> statistic."
>
> HTH,
> J.
>
> __________________________________________
>
> Prof. John Antonakis
> Faculty of Business and Economics
> Department of Organizational Behavior
> University of Lausanne
> Internef #618
> CH-1015 Lausanne-Dorigny
> Switzerland
> Tel ++41 (0)21 692-3438
> Fax ++41 (0)21 692-3305
> http://www.hec.unil.ch/people/jantonakis
>
> Associate Editor
> The Leadership Quarterly
> __________________________________________
>
>
> On 06.05.2012 02:29, [email protected] wrote:
>> The Hausman test is actually use to select between fixed and random effect. To know which one to chose you proceed as follow: if the p value is greater than 0.5 then the fixed effect(fe ) is not good chose random effect(re ) and otherwise if reverse is the case. Secondly, to test for autocorrelation after the. 'xtreg' test, you use 'xttest0'
>> Sent from my BlackBerry wireless device from MTN
>>
>> -----Original Message-----
>> From: Caliph Omar Moumin<[email protected]>
>> Sender: [email protected]
>> Date: Sat, 5 May 2012 07:46:33
>> To: [email protected]<[email protected]>
>> Reply-To: [email protected]
>> Subject: st: fixed effect or random effect model
>>
>>
>> Dear all
>>     For the past two weeks i spent to decide whether i apply fixed effect or random effect model in my strongly unbalanced panel data. But I couldn't decide  it.
>> These are the tests i applied so could you please give a minute and advice me what to apply? I understood the my hausman test impllies that i can apply either fixed or random effect modells. Is that so? If that is correct then i choose to apply the random effect model becuase of some time in-variant involved.
>>     What about Breusch-Pagan Lagrange multiplier (LM) test? I have no clue as to how interperate this test? Could any help me?
>>     xtdescribe
>>          id:  6, 9, ..., 809378                                 n =      14503
>> nadmission1:  1, 2, ..., 16                                  T =         16
>>               Delta(nadmission1) = 1 unit
>>               Span(nadmission1)  = 16 periods
>>               (id*nadmission1 uniquely identifies each observation)
>> Distribution of T_i:   min      5%     25%       50%       75%     95%     max
>>                             1       1       1         1         1       2      16
>>         Freq.  Percent    Cum. |  Pattern
>>     ---------------------------+------------------
>>        13302     91.72   91.72 |  1...............
>>          797      5.50   97.21 |  11..............
>>          160      1.10   98.32 |  111.............
>>           97      0.67   98.99 |  1111............
>>           58      0.40   99.39 |  11111...........
>>           31      0.21   99.60 |  111111..........
>>           29      0.20   99.80 |  1111111.........
>>           12      0.08   99.88 |  11111111........
>>            8      0.06   99.94 |  111111111.......
>>            9      0.06  100.00 | (other patterns)
>>     ---------------------------+------------------
>>        14503    100.00         |  XXXXXXXXXXXXXXXX
>>     I want to compare between this two groups
>> xttab group;
>>                      Overall             Between            Within
>>        group |    Freq.  Percent      Freq.  Percent        Percent
>> ----------+-----------------------------------------------------
>>      alcohol |     275      1.64       191      1.32         100.00
>>     nonalcoh |   16443     98.36     14312     98.68         100.00
>> ----------+-----------------------------------------------------
>>        Total |   16718    100.00     14503    100.00         100.00
>>                                 (n = 14503)
>>
>>       .quietly xtreg cost duration sex age group, fe;
>> . estimates store fixed;
>> . quietly xtreg cost duration sex age group, re;
>> . estimates store random;
>> hausman fixed random;
>>                     ---- Coefficients ----
>>                 |      (b)          (B)            (b-B)     sqrt(diag(V_b-V_B))
>>                 |     fixed        random       Difference          S.E.
>> -------------+----------------------------------------------------------------
>>        duration |    874.4642     944.5754       -70.11117        84.24204
>> ------------------------------------------------------------------------------
>>                               b = consistent under Ho and Ha; obtained from xtreg
>>                B = inconsistent under Ha, efficient under Ho; obtained from xtreg
>>        Test:  Ho:  difference in coefficients not systematic
>>                      chi2(1) = (b-B)'[(V_b-V_B)^(-1)](b-B)
>>                              =        0.69
>>                    Prob>chi2 =      0.4053
>>
>>       Breusch-Pagan Lagrange multiplier (LM)test is performed as follows
>> xtreg cost duration, re;
>> xttest0;
>> Breusch and Pagan Lagrangian multiplier test for random effects
>>            cost[id,t] = Xb + u[id] + e[id,t]
>>            Estimated results:
>>                             |       Var     sd = sqrt(Var)
>>                    ---------+-----------------------------
>>                        cost |   2.27e+09       47647.13
>>                           e |   6.78e+08       26038.66
>>                           u |   1.66e+09       40752.23
>>            Test:   Var(u) = 0
>>                                  chi2(1) =    59.40
>>                              Prob>   chi2 =     0.0000
>>
>> A test for heteroskedasticity is performed which shows presence
>> xtreg  cost duration, fe
>> xttest3
>>
>> Modified Wald test for groupwise heteroskedasticity
>> in fixed effect regression model
>> H0: sigma(i)^2 = sigma^2 for all i
>> chi2 (14503)  = 2.1e+36
>> Prob>chi2 =      0.0000
>>
>>
>>
>>
>>
>>
>> Kind Regards,
>> Moumin
>>
>> Email:  [email protected]
>>
>> *
>> *   For searches and help try:
>> *  http://www.stata.com/help.cgi?search
>> *  http://www.stata.com/support/statalist/faq
>> *  http://www.ats.ucla.edu/stat/stata/
>>
>> *
>> *   For searches and help try:
>> *  http://www.stata.com/help.cgi?search
>> *  http://www.stata.com/support/statalist/faq
>> *  http://www.ats.ucla.edu/stat/stata/
> *
> *   For searches and help try:
> *  http://www.stata.com/help.cgi?search
> *  http://www.stata.com/support/statalist/faq
> *  http://www.ats.ucla.edu/stat/stata/
>
>
> *
> *   For searches and help try:
> *  http://www.stata.com/help.cgi?search
> *  http://www.stata.com/support/statalist/faq
> *  http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*  http://www.stata.com/help.cgi?searchhttp://www.stata.com/support/statalist/faqhttp://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index