User requirements and preferences for ITS Chapter | 4 51• Are difficult to be described in detail.
• Are not suitable for all project types.
4.3.4 Cognitive techniques
Cognitive requirements’ elicitation techniques build on the much broader
knowledge elicitation methods (Cooke, 1994) that allow critical collection and
analysis of information. They involve in-depth problem understanding and span
across multiple disciplines including psychology, business, cognitive science,
etc. They comprise of card sorting, repertory grids, and laddering.
- Card sorting
Card sorting sessions are a popular technique for collecting and prioritizing
requirements. During the sessions, the various user groups are asked to sort
a set of cards using various criteria. Cards correspond to domain entities and
the prioritization must be done considering different criteria. The technique is
more effective when all the essential domain entities are included, and requires
adequate domain knowledge for producing correct results. Both analysts, who
define the entities and the criteria, and stakeholders, who take the prioritization
decisions, must have this domain knowledge. Collaborative techniques can be
employed to improve the effectiveness of card sorting (Zowghi & Coulin, 2005).
- Repertory grids
Similar to card sorting, the repertory grids technique requests stakeholders
to evaluate specific domain entities by providing values according to various
attributes. They require stakeholders to categorize entities, order the categories,
and assign suitable attributes and corresponding values which are stored in a
matrix. The resulting grids can be the basis for identifying similar requirements
across the various domains and for locating differences between requirements.
They can produce an initial elicitation that will be further refined using well-
established refinement techniques (Zowghi & Coulin, 2005).
- Laddering
Laddering is a structured-interviewing technique, which builds on the rep-
ertory grid and creates a concepts’ hierarchy. It begins with the elicitation of
stakeholders’ goals and values and of the identification of the main attributes
of the product. Working on the main product attributes it is possible to extract
more information about users’ criteria and preferences. The stakeholders are
interviewed using different ladders each time following a hierarchy that maps
the order of questions to users’ priorities. It can be used in the first steps of the
elicitation process for constructing the initial ladders and then integrated with a
protocol editor that defines the conceptual input for laddering and further refines
the ladders. Brief and concrete requirements are preferred to lengthy ones, since
they can be added or removed more easily at the ladder (Shams-Ul-Arif &