Skip to main content

Table 3 Average model standard error (and Monte Carlo standard error) relative to empirical standard deviation

From: Two-stage sampling in the estimation of growth parameters and percentile norms: sample weights versus auxiliary variable estimation

  Fixed intercept Fixed Age R.Intercepts R.Coefficients R.Covariance
Complete 0.5694 (2.2505) -0.4001 (2.2288) -0.6407 (2.2251) -1.6948 (2.2007) 3.3954 (2.3145)
Naive 2.1402 (2.2864) -1.3941 (2.2074) 2.6297 (2.3006) -1.6499 (2.2035) -1.0186 (2.2171)
Weighted 8.5475 (2.4301) -2.9089 (2.1738) 0.1789 (2.2503) -0.4633 (2.2335) -1.3902 (2.2119)
PoAux -0.0666 (2.2366) -3.0192 (2.1711) -0.4454 (2.2310) -1.1911 (2.2138) -1.2034 (2.2128)
TPoAux 1.7493 (2.2775) -3.1518 (2.1681) -2.7690 (2.1791) -1.5976 (2.2046) -1.3179 (2.2104)
PoMiss 3.3083 (2.3130) -4.2109 (2.1448) 1.2974 (2.2726) -1.4011 (2.2091) -1.1374 (2.2158)
TPoMiss -3.3571 (2.1646) -4.4398 (2.1396) -4.8465 (2.1341) -1.6610 (2.2033) -1.5355 (2.2064)
NbAux 0.6920 (2.7424) -1.9741 (2.6677) -0.9206 (2.6994) -0.6157 (2.7084) -0.6002 (2.7096)
  1. The higher the value in absolute terms the larger the discrepancy between the model standard error and the empirical standard deviation. The models are abbreviated as follows: Complete the complete data model, Naive model 1 fitted to incomplete data, Weighted the Weighted model, PoAux the Poisson auxiliary model, TPoAux the transformed Poisson auxiliary model, PoMiss the Poisson/missingness model, TPoMiss the transformed Poisson/missingness model, NbAux the negative binomial auxiliary model