请问Unit Specific 模型和Population Average 模型的区别
请问老师,为什么在Bernoulli模型中有两个模型的系数,而在前面所学的一般模型中却没有呢?Overall Population是什么呢?两个系数的差别能说明什么?这里我们为什么要选择PA模型,如果他并不说明如何影响某一具体单元的均值。我看完讲义和书还是不太明白,恳请老师能否指导一下。谢谢 Hlm2例题分析讲义中的17页,有一句说明,“有学前教育的孩子的留级发生比只相当于没有学前教育的孩子的0.59倍,也就是说,只有没有受到学前教育的孩子的留级发生比的41%”,请问是不是错了,应该是59%吧? 按照您的要求作了线形模型。
----------------------------------------------------
Random level-1 coefficient Reliability estimate
----------------------------------------------------
INTRCPT1, B0 0.791
----------------------------------------------------
Bernoulli模型的可靠性估计如下:
---------------------------------------------------
Random level-1 coefficient Reliability estimate
----------------------------------------------------
INTRCPT1, B0 0.682
截距的可靠性更高,但在系数估计上差别很大,
Final estimation of fixed effects
(with robust standard errors)
----------------------------------------------------------------------------
Standard Approx.
Fixed Effect Coefficient Error T-ratio d.f. P-value
----------------------------------------------------------------------------
For INTRCPT1, B0
INTRCPT2, G00 0.154129 0.010910 14.128 354 0.000
MSESC, G01 -0.033970 0.024716 -1.374 354 0.170
For MALE slope, B1
INTRCPT2, G10 0.054292 0.008463 6.415 7512 0.000
For PPED slope, B2
INTRCPT2, G20 -0.064580 0.010707 -6.031 7512 0.000
----------------------------------------------------------------------------
Final estimation of variance components:
-----------------------------------------------------------------------------
Random Effect Standard Variance df Chi-square P-value
Deviation Component
-----------------------------------------------------------------------------
INTRCPT1, U0 0.14210 0.02019 354 1742.96220 0.000
level-1, R 0.31814 0.10121
-----------------------------------------------------------------------------
Bernoulli模型系数如下:
Final estimation of fixed effects
(Population-average model with robust standard errors)
----------------------------------------------------------------------------
Standard Approx.
Fixed Effect Coefficient Error T-ratio d.f. P-value
----------------------------------------------------------------------------
For INTRCPT1, B0
INTRCPT2, G00 -1.748402 0.082158 -21.281 354 0.000
MSESC, G01 -0.283620 0.196005 -1.447 354 0.149
For MALE slope, B1
INTRCPT2, G10 0.446546 0.062788 7.112 7512 0.000
For PPED slope, B2
INTRCPT2, G20 -0.536378 0.082221 -6.524 7512 0.000
----------------------------------------------------------------------------
但是检验显著与否的结果相同,请问老师,我还有什么没有看到吗? 您的讲义27页说层1的方差为1/L,可是输出的结果不是这样,请问为什么?
Level-1 variance = sigma_squared/L
The value of the likelihood function at iteration 6 = -2.539444E+003
RESULTS FOR NON-LINEAR MODEL WITH THE LOG LINK FUNCTION: Unit-Specific Model
(macro iteration 8)
Sigma_squared = 0.62419 上面的结果是没有选择过离散的输出结果,即假设不存在过离散现象,那么向上面这种情况,是否应该检验过离散呢,哪种结果更可信呢?另外我始终不明白非线性模型层一的方差为什么都要取倒数? 讲义29页的最下面,“实际上是说总体平均水平为前一年中平均每人平均旷课的0.147次”,应该是0.147倍吧?
页:
[1]