Correcting Model Fit Criteria for Small Sample Latent Growth Models With Incomplete Data

Area of Science:

• Statistics
• Developmental Psychology
• Quantitative Psychology

Background:

• Latent growth models (LGMs) and mixed-effect models (MEMs) are related statistical frameworks.
• LGMs offer unique criteria for assessing global data-model fit.
• Existing small sample corrections for LGMs perform poorly with incomplete data.

Purpose of the Study:

• To investigate the performance of existing small sample corrections for LGMs with missing data.
• To propose and evaluate a novel correction method for small sample LGMs accommodating missing data.
• To demonstrate the impact of ignoring missing data on global data-model fit assessment.

Main Methods:

• Simulation studies were conducted to assess the performance of small sample corrections under various missing data conditions.
• A new missing data correction was developed for small sample correction equations.
• The proposed correction was validated through simulation and an applied example.

Main Results:

• Existing small sample corrections demonstrated inadequacy when applied to LGMs with missing data.
• The proposed missing data correction significantly improved the performance of small sample correction equations.
• Disregarding missing data in correction equations led to substantial alterations in global data-model fit assessment.

Conclusions:

• Small sample corrections for LGMs require adjustments to account for missing data.
• The proposed missing data correction offers a reliable method for assessing global data-model fit in LGMs with incomplete datasets.
• Accurate assessment of global data-model fit in developmental psychology requires addressing missing data in small sample correction methods.