Measuring mental capacity: Thomsonʹs Bonds model and Spearmanʹs g-model compared

A contemporary version of Thomsonʹs Bonds model for intelligence was proposed in Bartholomew et al. (2009). It was shown there that, as far as existing data was concerned, it could compete on equal terms, biologically and statistically with Spearmanʹs long-established g-model. In this paper we take the further, crucial, step of showing how to fit and interpret the Bonds model. We shall show that although the two models are statistically indistinguishable, they offer radically different interpretations of how the brain works. In particular, we show that:i) nds model suggests a way of measuring individual ability (or mental capacity) from test scores by introducing the notion of what we shall call the power or mental capacity of the brain. This is analogous to, but not equivalent to, Spearmanʹs g. del can be generalised to allow for the possibility that the brain makes several ‘passes’ through the test items. This identifies different characteristics of the items rather than different dimensions of ability as with Spearmanʹs model. nds model does not require each person to have the same number of bonds as Spearman appears to have mistakenly supposed. thods are illustrated on two data sets and an R package including functions and data sets is provided to enable the reader to explore the model in more detail.