Fig 1.
Changes in non-standardized RMSE of the discrimination parameter a with increasing sample size (n).
The RMSE consistently decreased as n increased, indicating that larger sample sizes contributed to improved estimation accuracy of the discrimination parameter.
Fig 2.
Changes in RMSEa with sample size (n).
The boxplots show the distribution of estimation errors under each condition. The parameter a represents the slope parameter in the Graded Response Model (GRM), and RMSEa indicates the estimation accuracy for a. The x-axis represents the sample size (n), and the y-axis represents the RMSE of the estimated slope parameters. The number of response categories (K) is indicated by panel or grouping as appropriate.
Fig 3.
Effects of sample size (n) and number of items (J) on the .
The gradually decreased with increasing J, although the effect was modest compared to the influence of n.
Fig 4.
Transition of with sample size (n).
Boxplots show the distribution of standardized and FPC-corrected RMSE values for the discrimination parameter a. The x-axis represents sample size (n), and the y-axis represents RMSE. The number of response categories (K) is indicated by panel or grouping as appropriate.
Fig 5.
Changes in RMSEa across the number of items (J).
The line plot shows how the RMSE of the discrimination parameter (a) changes with increasing item count (J). The x-axis represents the number of items (J), and the y-axis represents RMSEa.
Fig 6.
Changes in across the number of items (J).
The line plot shows how the changes with increasing number of items (J). The RMSE values were corrected using finite population correction (FPC). The x-axis represents the number of items (J), and the y-axis represents
.
Fig 7.
Changes in by sample size (n), number of items (J), and number of response categories (K).
The line graph illustrates how the FPC-corrected RMSE of the latent trait parameter () varies across different levels of n, J, and K.
Fig 8.
Changes in by number of items (J).
The boxplots show the distribution of under each condition. The x-axis represents the number of items (J), and the y-axis represents the FPC-corrected RMSE of the estimated latent trait (
).
Fig 9.
Changes in the correlation coefficient () between
and
by sample size (n).
The line plot shows how the Fisher-transformed correlation coefficient () between true and estimated values of
varies with increasing sample size (n). The number of items (J) is indicated as appropriate.