Partitioning the Data Model

Without using this specific terminology, all prior versions of survey-like packages divide the data model into two broad areas:

• Metadata: those entities that define the structure, composition, and behaviors of a data collection form; metaphorically -- the shoebox
• Data: those entities that actually store data collected by the form; metaphorically -- the shoes

Consider these diagrams of the Complex Survey package (full details here)

and of the Questionnaire module (full details here)

This overall conceptual approach has served well so far, and it is also the approach taken by the relevant standards groups (IMS and CDISC) with which we're attempting to be as congruent as possible. So this is the approach we're taking with Assessment.

Implications for Development

The functional requirements identified for Assessment can be lumped into several main categories. Here's where interventions need to be made:

• More sophisticated types of questions: metadata elements concerning Items and Choices need enhancement
• Enhanced post-submission form processing such as data validation and navigation controls: metadata elements concerning Item Checks etc need to be defined
• Improved data retrieval and reporting: enhanced queries and display procs need to be created, but the underlying data entities are probably pretty close
• Improved interoperability with other packages: this has implications at both the metadata and data levels