Hello all,
I'm running the Conditional (fixed-effect) logistic regression for Best Worst scaling analysis. The results are as below:
1. clogit yvalue w t f s y,group(ncs)
note: y omitted because of collinearity.
note: multiple positive outcomes within groups encountered.
note: 91 groups (1,820 obs) omitted because of all positive or
all negative outcomes.
Iteration 0: Log likelihood = -3669.8944
Iteration 1: Log likelihood = -3668.3454
Iteration 2: Log likelihood = -3668.345
Iteration 3: Log likelihood = -3668.345
Conditional (fixed-effects) logistic regression Number of obs = 26,980
LR chi2(4) = 777.31
Prob > chi2 = 0.0000
Log likelihood = -3668.345 Pseudo R2 = 0.0958
------------------------------------------------------------------------------
yvalue | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
w | .3210077 .0573404 5.60 0.000 .2086226 .4333928
t | 1.468192 .0605694 24.24 0.000 1.349478 1.586906
f | .259552 .0569483 4.56 0.000 .1479354 .3711686
s | .6957712 .059394 11.71 0.000 .5793612 .8121813
y | 0 (omitted)
------------------------------------------------------------------------------
2. clogit yvalue c g e fo k n m p q,group(ncs)
note: multiple positive outcomes within groups encountered.
note: 91 groups (1,820 obs) omitted because of all positive or
all negative outcomes.
Iteration 0: Log likelihood = -2928.488
Iteration 1: Log likelihood = -2913.7449
Iteration 2: Log likelihood = -2913.7196
Iteration 3: Log likelihood = -2913.7196
Conditional (fixed-effects) logistic regression Number of obs = 26,980
LR chi2(9) = 2286.56
Prob > chi2 = 0.0000
Log likelihood = -2913.7196 Pseudo R2 = 0.2818
------------------------------------------------------------------------------
yvalue | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
c | .9311648 .0878287 10.60 0.000 .7590237 1.103306
g | 3.009064 .1145599 26.27 0.000 2.784531 3.233598
e | 3.168068 .1238748 25.57 0.000 2.925278 3.410858
fo | 2.469387 .1009579 24.46 0.000 2.271513 2.667261
k | 1.331987 .1059611 12.57 0.000 1.124307 1.539667
n | 1.073204 .1035582 10.36 0.000 .8702334 1.276174
m | .8097167 .1134727 7.14 0.000 .5873142 1.032119
p | -.091148 .104598 -0.87 0.384 -.2961564 .1138603
q | -1.115552 .1114767 -10.01 0.000 -1.334042 -.8970618
.
where w,t,f,s, and y are attributes; c,g,e,fo,k,n,m,p,q are levels in each attribute. My question is how can we interpret the meaning of Coefficient in both analyses? For example, coefficient of "W" is 0.32, "T" is 1.46.
Thanks for your help.
.
I'm running the Conditional (fixed-effect) logistic regression for Best Worst scaling analysis. The results are as below:
1. clogit yvalue w t f s y,group(ncs)
note: y omitted because of collinearity.
note: multiple positive outcomes within groups encountered.
note: 91 groups (1,820 obs) omitted because of all positive or
all negative outcomes.
Iteration 0: Log likelihood = -3669.8944
Iteration 1: Log likelihood = -3668.3454
Iteration 2: Log likelihood = -3668.345
Iteration 3: Log likelihood = -3668.345
Conditional (fixed-effects) logistic regression Number of obs = 26,980
LR chi2(4) = 777.31
Prob > chi2 = 0.0000
Log likelihood = -3668.345 Pseudo R2 = 0.0958
------------------------------------------------------------------------------
yvalue | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
w | .3210077 .0573404 5.60 0.000 .2086226 .4333928
t | 1.468192 .0605694 24.24 0.000 1.349478 1.586906
f | .259552 .0569483 4.56 0.000 .1479354 .3711686
s | .6957712 .059394 11.71 0.000 .5793612 .8121813
y | 0 (omitted)
------------------------------------------------------------------------------
2. clogit yvalue c g e fo k n m p q,group(ncs)
note: multiple positive outcomes within groups encountered.
note: 91 groups (1,820 obs) omitted because of all positive or
all negative outcomes.
Iteration 0: Log likelihood = -2928.488
Iteration 1: Log likelihood = -2913.7449
Iteration 2: Log likelihood = -2913.7196
Iteration 3: Log likelihood = -2913.7196
Conditional (fixed-effects) logistic regression Number of obs = 26,980
LR chi2(9) = 2286.56
Prob > chi2 = 0.0000
Log likelihood = -2913.7196 Pseudo R2 = 0.2818
------------------------------------------------------------------------------
yvalue | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
c | .9311648 .0878287 10.60 0.000 .7590237 1.103306
g | 3.009064 .1145599 26.27 0.000 2.784531 3.233598
e | 3.168068 .1238748 25.57 0.000 2.925278 3.410858
fo | 2.469387 .1009579 24.46 0.000 2.271513 2.667261
k | 1.331987 .1059611 12.57 0.000 1.124307 1.539667
n | 1.073204 .1035582 10.36 0.000 .8702334 1.276174
m | .8097167 .1134727 7.14 0.000 .5873142 1.032119
p | -.091148 .104598 -0.87 0.384 -.2961564 .1138603
q | -1.115552 .1114767 -10.01 0.000 -1.334042 -.8970618
.
where w,t,f,s, and y are attributes; c,g,e,fo,k,n,m,p,q are levels in each attribute. My question is how can we interpret the meaning of Coefficient in both analyses? For example, coefficient of "W" is 0.32, "T" is 1.46.
Thanks for your help.
.
