Abstract
IN an article in the current number of the British Journal of Psychology, I have shown that two tests which fit separately into a hierarchy, but the correlation of which with one another breaks it, can under certain conditions be weighted so as to form a team of two tests correlating perfectly with g. The object of the present note is to point out an extension of this principle. If k tests each fit separately into a hierarchy, but cannot co-exist in it, their correlations with g (r1g, r2g rkg) can be separately found. A team of these k tests can then be formed, with weights proportional to () the co-factors of r1g, r2g rkg in the determinant to give an estimate of g, and this estimate will correlate perfectly with g if the value of the above determinant is zero. If the latter is the case, the k non-conforming tests may be represented as containing among them k1 group components in addition to the general component g.
Similar content being viewed by others
Article PDF
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
THOMSON, G. Measuring General Intelligence by Tests which break the g-Hierarchy. Nature 135, 71 (1935). https://doi.org/10.1038/135071b0
Issue Date:
DOI: https://doi.org/10.1038/135071b0
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.