Abstract/Description/Artist Statement
This study examined the nonparametric item response characteristics of the Teacher Formative Assessment Practice Scale (TFAPS) using Mokken scale analysis within its established two‑factor structure. The TFAPS includes a teacher‑directed subscale (items 1–6) and a student‑directed subscale (items 7–10), each intended to capture distinct dimensions of formative assessment practice. Weighted Loevinger scalability coefficients for polytomous items were computed to assess how well each subscale met the assumptions of the monotone homogeneity model. Results indicated that the teacher‑directed subscale had a scalability coefficient of H = 0.354 and Cronbach’s alpha of 0.721, suggesting weak but acceptable internal structure. In contrast, the student‑directed subscale showed considerably stronger psychometric properties, with H = 0.530 and Cronbach’s alpha of 0.800, indicating a more cohesive and reliable item hierarchy. Automated Item Selection Procedure (AISP) solutions at lower bounds of .30, .40, and .50 further highlighted these differences. Across all thresholds, the student‑directed items formed a consistently stable and well‑defined cluster, whereas the teacher‑directed items exhibited greater dispersion under stricter criteria. Monotonicity assessments revealed only minimal, localized violations on a small number of items, supporting the appropriateness of both subscales for analysis under nonparametric item response models. Overall, findings indicate that the TFAPS comprises a strong and psychometrically robust student‑directed subscale and a weaker yet interpretable teacher‑directed subscale. Several teacher‑directed items may benefit from targeted refinement to enhance scalability and overall measurement precision.
Faculty Advisor/Mentor
Prof. Kristin Gansle
Faculty Advisor/Mentor Email
kgansle@odu.edu
Faculty Advisor/Mentor Department
STEM Education and Professional Studies
College/School Affiliation
Darden College of Education & Professional Studies
Student Level Group
Graduate/Professional
Presentation Type
Oral Presentation
Included in
Educational Assessment, Evaluation, and Research Commons, Educational Psychology Commons, Secondary Education Commons
Mokken Scale Analysis of the Teacher Formative Assessment Practice Scale
This study examined the nonparametric item response characteristics of the Teacher Formative Assessment Practice Scale (TFAPS) using Mokken scale analysis within its established two‑factor structure. The TFAPS includes a teacher‑directed subscale (items 1–6) and a student‑directed subscale (items 7–10), each intended to capture distinct dimensions of formative assessment practice. Weighted Loevinger scalability coefficients for polytomous items were computed to assess how well each subscale met the assumptions of the monotone homogeneity model. Results indicated that the teacher‑directed subscale had a scalability coefficient of H = 0.354 and Cronbach’s alpha of 0.721, suggesting weak but acceptable internal structure. In contrast, the student‑directed subscale showed considerably stronger psychometric properties, with H = 0.530 and Cronbach’s alpha of 0.800, indicating a more cohesive and reliable item hierarchy. Automated Item Selection Procedure (AISP) solutions at lower bounds of .30, .40, and .50 further highlighted these differences. Across all thresholds, the student‑directed items formed a consistently stable and well‑defined cluster, whereas the teacher‑directed items exhibited greater dispersion under stricter criteria. Monotonicity assessments revealed only minimal, localized violations on a small number of items, supporting the appropriateness of both subscales for analysis under nonparametric item response models. Overall, findings indicate that the TFAPS comprises a strong and psychometrically robust student‑directed subscale and a weaker yet interpretable teacher‑directed subscale. Several teacher‑directed items may benefit from targeted refinement to enhance scalability and overall measurement precision.