period和loan一起做回归就出现了这种情况
. logit schedule period loan
Iteration 0: log likelihood = -65818.578
Iteration 1: log likelihood = -64067.481
Iteration 2: log likelihood = -62353.435
Iteration 3: log likelihood = -60212.509
Iteration 4: log likelihood = -59835.765
Iteration 5: log likelihood = -59829.737
Iteration 6: log likelihood = -59829.737 (backed up)
Iteration 7: log likelihood = -59829.737 (backed up)
Iteration 8: log likelihood = -59829.737 (backed up)
Iteration 9: log likelihood = -59829.737 (backed up)
Iteration 10: log likelihood = -59829.737 (backed up)
Iteration 11: log likelihood = -59829.737 (backed up)
Iteration 12: log likelihood = -59829.737 (backed up)
Iteration 13: log likelihood = -59829.737 (backed up)
Iteration 14: log likelihood = -59829.737 (backed up)
和其他变量在一起则完全没问题,如下表。loan变量和其他变量一起也完全没问题
. logit schedule period inscore outscore
Iteration 0: log likelihood = -65818.578
Iteration 1: log likelihood = -48557.048
Iteration 2: log likelihood = -47195.366
Iteration 3: log likelihood = -47058.916
Iteration 4: log likelihood = -47051.595
Iteration 5: log likelihood = -47051.583
Iteration 6: log likelihood = -47051.583
Logistic regression Number of obs = 97224
LR chi2(3) = 37533.99
Prob > chi2 = 0.0000
Log likelihood = -47051.583 Pseudo R2 = 0.2851
------------------------------------------------------------------------------
schedule | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
period | -.051859 .0024789 -20.92 0.000 -.0567175 -.0470005
inscore | .0809126 .0005939 136.23 0.000 .0797485 .0820767
outscore | .0001233 4.76e-06 25.91 0.000 .0001139 .0001326
_cons | -1.496561 .0235419 -63.57 0.000 -1.542702 -1.45042
------------------------------------------------------------------------------
本人计量原理学的不是很好,求高人指点 这不停的Backed up是怎么了?