Dear community members,
With reference to thread https://www.statalist.org/forums/for...-of-i-category within https://www.statalist.org/forums/for...nality-not-met.
Each student_ID is repeating multiple times.I have done xtset student_ID before running the regressions.
My dataset: It is a natural data of a college. In which students have taken many tests( within and across the semesters) in the different courses (majoring courses). I want to compare the score of students by their social category.
The dependent variable is a dummy i.e., Positive_disc01 ( 1= positive score, 0=negative score)
The core independent variable is, students' particular socio-religious group i.e., Student_Caste_New
In my research, I am mainly interested to know if students who are in a particular socio-religious group (i.e., stud_SCST), do they have a lower chance of getting a positive score (i.e., Positive_disc01: which is, a certain type of socioeconomic index).
Hausman test is failing to suggest if I should go for fixed or random effect model.
However, an expert had suggested Hausman test will run through (https://www.statalist.org/forums/for...11#post1696711)
How can one know if she needs to apply the fixed or random model for analysis in such circumstance.
regards,
ajay
With reference to thread https://www.statalist.org/forums/for...-of-i-category within https://www.statalist.org/forums/for...nality-not-met.
Each student_ID is repeating multiple times.I have done xtset student_ID before running the regressions.
My dataset: It is a natural data of a college. In which students have taken many tests( within and across the semesters) in the different courses (majoring courses). I want to compare the score of students by their social category.
The dependent variable is a dummy i.e., Positive_disc01 ( 1= positive score, 0=negative score)
The core independent variable is, students' particular socio-religious group i.e., Student_Caste_New
In my research, I am mainly interested to know if students who are in a particular socio-religious group (i.e., stud_SCST), do they have a lower chance of getting a positive score (i.e., Positive_disc01: which is, a certain type of socioeconomic index).
Hausman test is failing to suggest if I should go for fixed or random effect model.
However, an expert had suggested Hausman test will run through (https://www.statalist.org/forums/for...11#post1696711)
Code:
. xtset collegerollno
Panel variable: collegerollno (unbalanced)
. doedit "D:\PROJECT_IntExt_Internal external project stata files\New_Dofile_Deepak.do"
. xtlogit Positive_disc01 i.Student_Caste_New i.T_jati_new attendence_percent i.T_nature i.course1 semester
> i.T_gender i.division , fe
note: multiple positive outcomes within groups encountered.
note: 8 groups (122 obs) omitted because of all positive or
all negative outcomes.
note: 2.Student_Caste_New omitted because of no within-group variance.
note: 3.Student_Caste_New omitted because of no within-group variance.
note: 2.course1 omitted because of no within-group variance.
note: 3.course1 omitted because of no within-group variance.
note: 4.course1 omitted because of no within-group variance.
note: 5.course1 omitted because of no within-group variance.
note: 6.course1 omitted because of no within-group variance.
note: 7.course1 omitted because of no within-group variance.
note: 2.division omitted because of no within-group variance.
note: 3.division omitted because of no within-group variance.
Iteration 0: log likelihood = -4711.5243
Iteration 1: log likelihood = -4697.4242
Iteration 2: log likelihood = -4697.4069
Iteration 3: log likelihood = -4697.4069
Conditional fixed-effects logistic regression Number of obs = 9,791
Group variable: collegerollno Number of groups = 647
Obs per group:
min = 6
avg = 15.1
max = 16
LR chi2(6) = 49.84
Log likelihood = -4697.4069 Prob > chi2 = 0.0000
------------------------------------------------------------------------------------
Positive_disc01 | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------------+----------------------------------------------------------------
Student_Caste_New |
SC/ST | 0 (omitted)
OBC | 0 (omitted)
|
T_jati_new |
2 | -.0521505 .0663246 -0.79 0.432 -.1821443 .0778433
3 | .0123199 .0711494 0.17 0.863 -.1271304 .1517703
|
attendence_percent | .0109843 .0017903 6.14 0.000 .0074753 .0144932
2.T_nature | .0078302 .0550874 0.14 0.887 -.100139 .1157995
|
course1 |
eco | 0 (omitted)
eng | 0 (omitted)
hindi | 0 (omitted)
history | 0 (omitted)
maths | 0 (omitted)
pol | 0 (omitted)
|
semester | .0873339 .0210808 4.14 0.000 .0460162 .1286516
2.T_gender | .1974969 .0731768 2.70 0.007 .054073 .3409208
|
division |
2 | 0 (omitted)
3 | 0 (omitted)
------------------------------------------------------------------------------------
. estimates store fixed
. xtlogit Positive_disc01 i.Student_Caste_New i.T_jati_new attendence_percent i.T_nature i.course1 semester
> i.T_gender i.division , re
Fitting comparison model:
Iteration 0: log likelihood = -6795.0358
Iteration 1: log likelihood = -6100.0401
Iteration 2: log likelihood = -6095.8839
Iteration 3: log likelihood = -6095.8825
Iteration 4: log likelihood = -6095.8825
Fitting full model:
tau = 0.0 log likelihood = -6095.8825
tau = 0.1 log likelihood = -6081.9125
tau = 0.2 log likelihood = -6093.8851
Iteration 0: log likelihood = -6081.9125
Iteration 1: log likelihood = -6081.6228
Iteration 2: log likelihood = -6081.6227
Random-effects logistic regression Number of obs = 9,913
Group variable: collegerollno Number of groups = 655
Random effects u_i ~ Gaussian Obs per group:
min = 6
avg = 15.1
max = 16
Integration method: mvaghermite Integration pts. = 12
Wald chi2(16) = 941.07
Log likelihood = -6081.6227 Prob > chi2 = 0.0000
------------------------------------------------------------------------------------
Positive_disc01 | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------------+----------------------------------------------------------------
Student_Caste_New |
SC/ST | -.2598846 .0763338 -3.40 0.001 -.409496 -.1102731
OBC | -.1290695 .0698694 -1.85 0.065 -.266011 .0078719
|
T_jati_new |
2 | -.0658054 .0658859 -1.00 0.318 -.1949394 .0633286
3 | .0219436 .0698121 0.31 0.753 -.1148857 .1587728
|
attendence_percent | .011776 .0013962 8.43 0.000 .0090396 .0145125
2.T_nature | -.0126113 .0543952 -0.23 0.817 -.119224 .0940014
|
course1 |
eco | .8948999 .0890632 10.05 0.000 .7203392 1.069461
eng | -.3307983 .0965467 -3.43 0.001 -.5200264 -.1415703
hindi | 1.293322 .0993397 13.02 0.000 1.09862 1.488025
history | .1819405 .1195191 1.52 0.128 -.0523127 .4161936
maths | .7044516 .0912066 7.72 0.000 .5256899 .8832133
pol | 2.043114 .0932743 21.90 0.000 1.860299 2.225928
|
semester | .0906821 .0196079 4.62 0.000 .0522512 .1291129
2.T_gender | .2185489 .072893 3.00 0.003 .0756812 .3614166
|
division |
2 | .0818407 .0719516 1.14 0.255 -.0591819 .2228634
3 | .4876127 .1682825 2.90 0.004 .1577851 .8174403
|
_cons | -1.532339 .1544492 -9.92 0.000 -1.835053 -1.229624
-------------------+----------------------------------------------------------------
/lnsig2u | -2.247357 .231178 -2.700457 -1.794256
-------------------+----------------------------------------------------------------
sigma_u | .3250818 .0375759 .259181 .407739
rho | .0311226 .0069709 .0200101 .0481034
------------------------------------------------------------------------------------
LR test of rho=0: chibar2(01) = 28.52 Prob >= chibar2 = 0.000
. estimates store random
. hausman fixed
---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| fixed random Difference Std. err.
-------------+----------------------------------------------------------------
T_jati_new |
2 | -.0521505 -.0658054 .0136549 .0076156
3 | .0123199 .0219436 -.0096236 .0137299
attendence~t | .0109843 .011776 -.0007918 .0011207
2.T_nature | .0078302 -.0126113 .0204415 .008705
semester | .0873339 .0906821 -.0033482 .0077415
2.T_gender | .1974969 .2185489 -.021052 .0064381
------------------------------------------------------------------------------
b = Consistent under H0 and Ha; obtained from xtlogit.
B = Inconsistent under Ha, efficient under H0; obtained from xtlogit.
Test of H0: Difference in coefficients not systematic
chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= -86.36
Warning: chi2 < 0 ==> model fitted on these data
fails to meet the asymptotic assumptions
of the Hausman test; see suest for a
generalized test.
How can one know if she needs to apply the fixed or random model for analysis in such circumstance.
regards,
ajay
Comment