Errors and Inference A Framework Julia Lane Coleridge

  • Slides: 11
Download presentation
Error(s) and Inference A Framework Julia Lane, Coleridge Initiative

Error(s) and Inference A Framework Julia Lane, Coleridge Initiative

Conceptual Practical Operational

Conceptual Practical Operational

Quality Framework Some considerations: • Timeliness • Closeness to core measure • Coverage •

Quality Framework Some considerations: • Timeliness • Closeness to core measure • Coverage • Geographic detail • Longitudinal Consistency How do we trade off? Conceptual Practical Operational

The aim of total quality is to optimally balance all producer and user dimensions

The aim of total quality is to optimally balance all producer and user dimensions Timeliness Accuracy Accessibility Credibility Completeness Comparability Relevance Interpretability Coherence User Producer BIEMER, P. (2017) ERRORS AND INFERENCE, CHAPTER 10 IN BD AND SOCIAL SCIENCE Conceptual Practical Operational

Past Focus -- Total Survey Error Framework (Groves et al. 2004) Conceptual Practical Operational

Past Focus -- Total Survey Error Framework (Groves et al. 2004) Conceptual Practical Operational

Multiple administrative data sources: Phase 1 Extension of the framework when using integrated administrative

Multiple administrative data sources: Phase 1 Extension of the framework when using integrated administrative sources by Zhang (2012) Cited in Reid, Zabala, Holmberg “Extending TSE to Administrative Data: A Quality Framework and Case Studies from Stats NZ “ Conceptual Practical Operational

Multiple administrative data sources: Phase 2 Extension of the framework when using integrated administrative

Multiple administrative data sources: Phase 2 Extension of the framework when using integrated administrative sources by Zhang (2012) Cited in Reid, Zabala, Holmberg “Extending TSE to Administrative Data: A Quality Framework and Case Studies from Stats NZ “ Conceptual Practical Operational

The relationship of errors to data structure Columns = measurement V 1 V 2

The relationship of errors to data structure Columns = measurement V 1 V 2 Rows = representation Record # ∙∙∙ VK total error = row error + column error + cell error BIEMER, P. (2017) ERRORS AND INFERENCE, CHAPTER 10 IN BD AND SOCIAL SCIENCE Conceptual Practical Operational

The relationship of errors to integrated administrative datasets V 1 V 2 . .

The relationship of errors to integrated administrative datasets V 1 V 2 . . . … ∙∙∙ … Rows = representation Record # Columns = measurement VK+1 ∙∙∙ VK BIEMER, P. (2020) ERRORS AND INFERENCE, CHAPTER 10 IN BD AND SOCIAL SCIENCE Conceptual Practical Operational

Understand how data are generated http: //tylervigen. com/spurious-correlations Conceptual Practical Operational

Understand how data are generated http: //tylervigen. com/spurious-correlations Conceptual Practical Operational

Questions What do the rows for the data in your project represent? - Identify

Questions What do the rows for the data in your project represent? - Identify one or two sources of possible error - Discuss the possible consequences What do the key columns for the data in your project represent? - Identify one or two sources of possible error - Discuss the possible consequences