NEWS

---

IRTest 2.1.0 (2024-10-04)

• Weighted likelihood estimates (WLE) are available for ability parameter estimation.
• adaptive_test function is added for ability parameter estimation with fixed item parameters.
• IRTest_Cont utilizes MLE in item parameter estimation for a faster estimation.

IRTest 2.0.0 (2024-03-22)

• IRT analysis for continuous item responses is now available.
  • IRTest_Cont has been added.
  • Other utility functions of IRTest are applicable to an output of IRTest_Cont.

IRTest 1.12.1

• Modified estimation functions to be utilized in ability estimation of adaptive testing.

IRTest 1.12.0 (2023-10-16)

• Developed functions for medel comparison and model selection.
• Modified GHc as an implicit object.
• Provided examples for functions.
• Changed from any(class()) to inherits().
• Provided more S3 methods.
• Updated the manual.
• Added coef_se that returns S.E. of item parameter estimates.

IRTest 1.11.0 (2023-09-19)

• Graded response model ("GRM") is now available.
• Developed MLE for mixed-format analysis.

IRTest 1.10.0

• Outputs reflect item names in an input file.
• Enhanced algorithm convergence for IRTest_Dich when there exist one or more negative slope parameters.
• Error in calculating MLE in the presence of missing data is resolved.
• summary returns the HQ criterion.
• Error in DataGeneration's model_D argument is fixed.
• plot function employs latent_distribution.

IRTest 1.9.1

• Minor errors have been corrected such as the calculation of $\rho_{\theta\theta^{'}}$ with discrete latent distribution and the calculation of $sd$ in the Davidian-cure method.
• Vignette file has been updated.

IRTest 1.9.0

• latent_distribution has been developed. This function evaluates an estimated latent density function.
• reliability calculates a coefficient on the $\theta$ scale.

IRTest 1.8.1

• Convergence of the estimation algorithm was enhanced for "PCM" by inserting latest item parameters instead of inserting starting values repeatedly.

IRTest 1.8.0

• inform_f_item and inform_f_test are developed.
• Convergence for "DC" and "LLS" was enhanced by updating density parameters every iteration.

IRTest 1.7.1

• Minor errors in the manual file are corrected.
• reliability uses the raw value in a given data for calculating the reliability coefficient.
• reliability returns both test reliability coefficient and item reliability coefficients.
• Every estimation function returns the raw score-categories of items.

IRTest 1.7.0 (2023-08-17)

• Log-linear smoothing method (latent_dist = "LLS") can be applied to estimate a latent distribution.

IRTest 1.6.0

• reliability has been developed to calculate marginal reliability coefficient of IRT.

IRTest 1.5.2

• Some of the polytomous items could be technically dichotomous when one or more of the categories are not observed in data. Now, these items can be analyised with the polytomous function.

IRTest 1.5.1

• Rasch-type analysis imposes a constraint on slope parameters to be equal across all items. Before, they were fixed to 1 regardless of the characteristics of data.

IRTest 1.5.0

• plot_item has been developed to draw an item response function.

IRTest 1.4.1

• Enhanced algorithm convergence for IRTest_Poly when there exist one or more negative slope parameters.

IRTest 1.4.0

• item_fit evaluates item fit.

IRTest 1.3.1

• Starting values (initialitem) can be automatically generated.

IRTest 1.3.0

• NA values are permitted for estimation functions.

IRTest 1.2.0

• The MLE option for ability parameter estimation is added for polytomous data.

IRTest 1.1.0

• The PCM model for polytomous data is available.

IRTest 1.0.0 (2023-07-18)

• Unit tests are more thoroughly conducted to enhance 'test coverage.'
• S3 class generic functions (i.e., print, plot, summary) can now be used for IRTest results.
• Plotting function "plot_LD" is substituted with the generic function plot.
• Additional ggplot2 arguments can be specified to plot.
• Badges are added to the Github main page in order for the current status of IRTest to be easily recognized.
• In the summary, additional information is provided; a cursory view of the latent distribution, model fit indices, and the number of parameters used.

IRTest 0.1.0 (2023-04-02)

• Ability parameter MLE is available for IRTest_Dich.
• Data Generation with fixed item and/or ability parameters is available.

IRTest 0.0.2 (2022-09-06)

• Re-submission to CRAN.

IRTest 0.0.1

• The initial release.