= Solution
For the moment approach, write $m_i=\mathbb E Y_i$ and let $V_i$ be a positive-definite working <covariance> constructed from the proposed moment formulas. For identifiable mean coefficients $\theta$, let $D_i=\partial m_i/\partial\theta^T$. <Generalized estimating equations> take the form
$$
\sum_{i=1}^m D_i^TV_i^{-1}(Y_i-m_i)=0.
$$
Across independent students, correct mean specification and standard regularity conditions give consistent identifiable mean coefficients even if the working <covariance> is wrong. <Sandwich covariance matrix> <standard errors> use
$$
A=\sum_iD_i^TV_i^{-1}D_i,\quad
B=\sum_iD_i^TV_i^{-1}(Y_i-m_i)(Y_i-m_i)^TV_i^{-1}D_i,\quad
\widehat{\operatorname{Cov}}(\widehat\theta)=A^{-1}BA^{-1}.
$$
Residual second-moment estimating equations can estimate identifiable dispersion and association parameters. The marginal mean only identifies $\gamma_0=\beta_0+\log(1-\pi)$, not $\beta_0$ and $\pi$ separately. Moment restrictions beyond the mean would be needed to separate them; as shown below, the printed restrictions do not generally recover the intended latent parameters.
For the hierarchical approach, maximize the <observed-data likelihood> after integrating out the <random effect>. Set $\mu_{ij}=e^{\beta_0+\beta^Tx_i+\alpha_j}$, $r=1/\tau$, $M_i=\sum_j\mu_{ij}$ and $s_i=\sum_jy_{ij}$. The joint likelihood contribution is
$$
L_i=\pi\mathbf1_{\{s_i=0\}}+(1-\pi)
\frac{\Gamma(r+s_i)}{\Gamma(r)}
\frac{r^r}{(r+M_i)^{r+s_i}}
\prod_{j=1}^3\frac{\mu_{ij}^{y_{ij}}}{y_{ij}!}.
$$
The integral comes from multiplying three conditional Poisson mass functions by the Gamma density and integrating $b_i^{r+s_i-1}e^{-(r+M_i)b_i}$ over $b_i>0$. Independent students give $\ell=\sum_i\log L_i$. Direct numerical maximum likelihood or an <expectation-maximization algorithm> can be used; in an EM algorithm an all-zero profile has an uncertain latent component membership. Information-based likelihood <standard errors> apply under regular interior conditions, with special care for parameter boundaries.
Back to article page
