A posteriori novelty metrics are often used in design research, in order to extract important information about creativity. However, different assessment approaches can be found in the literature, each of them with related pros and cons. In particular, weighted uncommonness, overall uncommonness and uncommonness across groups are the three main families of a-posteriori novelty metrics identified in this paper. Each of the considered literature metrics can provide specific types of information about the uncommonness of ideas, but in certain experimental circumstances, it could be difficult to rapidly identify the best-suited approach. This paper proposes an integrated procedure where the advantages offered by the three families of metrics can be applied concurrently. A generic case study is used for a first application of the proposal, and the obtained results show that a more comprehensive set of information about a-posteriori novelty can be extracted. In particular, novelty data from the three families of metrics are extracted in a single assessment process.

Different variants of a-posteriori novelty metrics can be found in the literature. Indeed, such a kind of assessment procedures is often used to extract useful information about creativity and/or idea generation effectiveness. In particular, the metric proposed by Shah et al. in 2003, is one of the most used and discussed in the literature. However, scholars highlighted some flaws for this metric, and some variants have been proposed to overcome them. This paper argues about the variants proposed for the a-posteriori metric of Shah et al., and proposes a selection framework to support researchers in selecting the most suited for their experimental needs. The proposed selection framework also highlights important research hints, which could pave the way for future activities. More specifically, it is still necessary to support the identification of the best-suited abstraction framework to assign weights to attributes, and the assignment of weights should be better supported as well. Moreover, this paper highlights the presence of “uncommonness of key attributes”, which needs to be investigated for experimental cases where ideas missing some key attributes are present.