ECD and Task Models
Task model is the term that ECD uses for whatever guidance might be available to item writers and content development professionals (CDPs) developing and refining items for a particular content standard/assessment target. Unfortunately, ECD has very little to say about what might or should be included in a task model.
Some writing about ECD offers an example or two of how task models have historically been used. But those task models are sparse and do not, in our view, come close to providing the kind of clarity that item writers and CDPs need. This omission from ECD is part of the inspiration for the RTD Project.
RTD’s Task Model
RTD does not have a singular task model format that is appropriate for all tests and for all item development efforts. Rather, RTD has an approach to development task models built towards the goal of producing items that elicit evidence of the targeted cognition for the range of typical test takers. This means giving item writers and CDPs the information they need to do their work. RTD task models are artifacts of learning and recording particular knowledge within an organization and making it available for individuals to further learn from.
RTD task models are used in the context of large scale standardized assessment. The items that follow from an RTD task model should equivalently assess the targeted cognition or intentionally address different facets of a standard. Task models help item developers to understand the bounds, stay within them and avoid avoidable mistakes and problems. At their best, they may even encourage innovation and improvement for future items.
Critical Elements of RTD Task Models
Though RTD task models vary from project to project, some elements are always important to include.
What the Targeted Cognition Is. Task models should unpack the meaning of the standard or the assessment target more fully and explicitly than the pithy and compact language that is usually used to write standards themselves. These explanations should focus on the aspects of the standards that are meant to be targeted with this assessment. This may effectively narrow the meaning of a standard and/or may break out different elements or implications that are assumed and/or implied by the language of a standard.
What the Targeted Cognition Is Not. It is variably useful to explain the bounds of some targeted cognition by laying out what falls beyond the targeted cognition. This may be because this assessment is intentionally not including that cognition, because those KSAs actually are part of some other standard or grade, or simply to ensure there no confusion as to what the limits of a content standard/assessment target include.
Preferred Distractors. Whenever task models are written for selected response items, it is important to include suggestions for distractors (i.e., incorrect answer options) or even preferred distractor options. It is so important that selected response items offer the most plausible and likely incorrect answers that it should not be left to to individual item writers and CDPs to attempt to figure those out themselves every time they work on an item.
Previously Unsuccessful Approaches. RTD task models should be repositories of knowledge built up by item developers over time, which includes knowledge gained through trial and error. Task models should list and perhaps explain these unsuccessful approaches so that future item developers can learn vicariously from their predecessors and avoid wasting time on dead ends.
These are not the only important elements of high quality task models, but they are the most important.
RTD Task Models are Living Documents
RTD task models are living documents, and should be editable in an ongoing basis. They should collect knowledge and lessons as it occurs, though not necessarily in the main sections of the document. Hence, it can be helpful to have an unstructured Notes or Emerging Thinking section where contributors can record notions, emerging ideas and potential additions for the more formal sections of the document. Eventually, appropriately senior and experienced CDPs — perhaps in consultation with test owners/commissioners — can change the main sections of the document.
RTD Task Model Format
RTD task models can be set up in a variety of formats, with different numbers of sections and subsections. Their level for formality and polish can vary, as befits the particular project. Of course, they can vary in the length, too.
In practice, task models often live as Microsoft Word documents or PDF files. The former are more easily editable as living documents and the latter prevent accidental edits from deleting or alternative contents inappropriately. They may even exist in both forms, with those authorized to make changes having access to the editable forms. Perhaps ideally, task models could exist in databases, thus most easily presenting relevant fields and sections, as needed. Ease of use is really what is most important.