Table 7.

Estimated Effects for Puberty Models: High Variability, Slow Sampling.

	Cohort Design						Accelerated Design
	Mean Est.	Std. Err.	Min	Max	Std. Bias	Prop. Sig.(adj.)	Mean Est.	Std. Err.	Min	Max	Std. Bias	Prop. Sig.(adj.)
Puberty Model
${\rho }_{growth}$	.599	.025	.521	.679			.615	.025	.502	.701
VIF	1.57	.073	1.38	1.86			1.61	.079	1.34	1.97
${\gamma }_{age}$(mis)	.456	.025	.382	.537	18.50	1 (1)	.415	.026	.329	.486	16.01	1 (1)
${\gamma }_{{age}^2}$(mis)	−.106	.024	−.176	−.037	16.72	.995 (.987)	−.084	.014	−.129	−.041	29.96	1 (1)
$\sigma$(mis)	.666	.020	.606	.730	8.12		.658	.020	.590	.702	7.56
$\sqrt{{\tau }_{00}}$(mis)	.736	.042	.609	.873	5.64		.744	.041	.604	.889	5.98
${\gamma }_{age}$	.000	.025	−.083	.072	−.009	.056 (.018)	.000	.026	−.084	.084	.001	.065 (.023)
${\gamma }_{puberty}$	2.00	.084	1.71	2.26	.027	1 (1)	2.00	.085	1.73	2.28	−.031	1 (1)
${\gamma }_{int}$	−.498	.048	−.660	−.337	.042	1 (1)	−.501	.045	−.640	−.346	−.028	1 (1)
$\sigma$	.499	.015	.456	.557	−.039		.499	.016	.447	.546	−.076
$\sqrt{{\tau }_{00}}$	.497	.032	.387	.589	−.082		.501	.031	.402	.613	.046
Tanner-Stage Model
${\rho }_{growth}$	.606	.024	.534	.677			.623	.025	.510	.707
${\gamma }_{age}$	.381	.046	.207	.546	8.20	1 (1)	.382	.044	.233	.521	8.76	1 (1)
${\gamma }_{tanner}$	.516	.023	.439	.600	.689	1 (1)	.512	.023	.447	.586	.505	1 (1)
${\gamma }_{int}$	−.128	.013	−.169	−.080	−.205	1 (1)	−.127	.012	−.167	−.092	−.211	1 (1)
$\sigma$	.499	.015	.456	.557	−.039		.499	.016	.447	.546	−.076
$\sqrt{{\tau }_{00}}$	.497	.032	.387	.589	−.082		.501	.031	.402	.613	.046

Note: Mean Est. is the expected parameter value across replicants and the Std. Err is the standard deviation of the sampling distribution. The proportion significant (Prop. Sig) represents the number of inferential tests that return a significant result for each parameter (unadjusted first, adjusted in parentheses). Std. Bias is the standardized bias estimate; values > .25 are bolded and represent unacceptable bias. The proportion significant only includes significant results where the inference is correct (e.g., significant negative effects are not included of the population parameter is positive). Rhos (${\text{ρ}}_{\text{g}\text{r}\text{o}\text{w}\text{t}\text{h}}$) represent the correlation between growth predictors, VIF represents variance inflation factor, gammas ($\text{γ}$) represent the fixed effects, sigma ($\text{σ}$) is the square-root of the residual random variance, and tau (${\text{τ}}_{00}$) is the person-level random variance (square-root shown for consistency). The (mis) denotes parameters from the mis-specified (i.e., age-only) model.