教育訓練實施與績效評鑑

臺北科技大學 115-1 學期課程，課號 366699。

課程大綱

In this course, students will gain a conceptual and practical understanding of program evaluation as a problem of causal inference under measurement error. Because every evaluation makes a causal claim about a construct that is not directly observed—knowledge, skill, transfer, performance—students will learn why measurement error is the central threat to a credible evaluation and how the latent-variable (structural equation modeling) framework addresses it. The first half builds the measurement and structural machinery; the second half re-examines the standard impact-evaluation designs (randomized assignment, instrumental variables, regression discontinuity, difference-in-differences, matching) as special cases of a structural model, making explicit where each one assumes the outcome is measured without error and what to do when it is not. Students will gain practical experience specifying, estimating, and interpreting these models, and judging what an evaluation's findings can and cannot support.

課程進度

Week 1: Overview of the course; program evaluation as a causal claim about an unobserved construct Week 2: Causality and causal models Week 3: Structural models with observed variables: regression and path analysis Week 4: The consequences of measurement error: attenuation, bias, and reliability Week 5: Measurement models: the relation between latent and observed variables Week 6: Confirmatory factor analysis Week 7: Self-Study (no physical class) Week 8: The general model: measurement and structure combined Week 9: Mid-term Week 10: Causal inference and counterfactuals (I): potential outcomes and structural causal models Week 11: Causal inference and counterfactuals (II): confounding and latent confounders Week 12: Randomized assignment (I): the RCT benchmark and the latent outcome Week 13: Randomized assignment (II): pre/post designs and latent change models Week 14: Instrumental variables and model-implied instrumental variables (MIIV-2SLS) Week 15: Regression discontinuity design with measurement error in the running variable Week 16: Difference-in-differences and longitudinal measurement invariance

評量方式

Attendance 20% Mid-term work 40% Final evaluation report 40%

教材

【遵守智慧財產權觀念，請使用正版教科書，不得使用非法影印教科書】 使用外文原文書：是 Course materials will be prepared by the instructor. Optional: Bollen, K. A. (1989). Structural equations with latent variables. Wiley. Guerra-Lòpez, I. (2008). Performance evaluation: Proven approaches for improving program and organizational performance. San Francisco: Jossey-Bass. Gertler, P. J. et al. (2016) Impact evaluation in practice, 2nd ed. (Inter-American Development Bank; World Bank, 2016).