Test designs and modeling under the general nominal diagnosis model framework.

Area of Science:

• Psychological assessment
• Psychometrics
• Educational measurement

Background:

• Traditional psychological questionnaires often suffer from response bias and limitations in criterion-referenced testing.
• Estimating numerous latent traits with existing scales can be challenging and inaccurate.
• Existing methods may not adequately address respondent-reported scale issues.

Purpose of the Study:

• To introduce situational tests within a general nominal diagnosis model framework as a superior alternative to traditional questionnaires.
• To demonstrate how this framework can mitigate response bias and improve latent trait estimation.
• To explore flexible item design within the situational testing framework.

Main Methods:

• Utilizing a general nominal diagnosis model framework for situational tests.
• Designing situationally nominal items to enhance flexibility in test development.
• Employing saturated and reduced forms of the model to verify test design expectations.
• Conducting two simulation studies to evaluate model and design effectiveness.

Main Results:

• Situational tests within the general nominal diagnosis model framework effectively address response bias.
• The framework allows for flexible test development and accurate latent trait estimation.
• Simulation studies confirmed the efficacy of the proposed models and situational test designs.
• Appropriate test designs can further mitigate remaining response biases.

Conclusions:

• Situational tests combined with the general nominal diagnosis model provide a robust solution for psychological assessment challenges.
• This approach offers enhanced flexibility and validity compared to traditional respondent-reported scales.
• The proposed methods are effective in improving the quality and reliability of psychological testing instruments.