The distribution of hand preference is discrete: A taxometric examination

Abstract

The dominant genetic models of human handedness (Annett, 1985; McManus, 1985a) have contrasting views on its nature. Whereas Annett's model proposes that handedness is distributed continuously (following the distribution of performance asymmetries of the hands), McManus's model proposes that handedness is distributed dichotomously. The aim of this study was to investigate the nature of the underlying distribution. The taxonomy of hand preference was analysed by two independent taxometric procedures, 'mean-above-mean-below-a-cut' and 'maximum-eigenvalue', separately in two geographically, demographically, and culturally distinct samples; one of Australian adults (N=787), and the other of Serbian high school students (N=1,224). The analyses revealed a latent categorical structure in both samples, with estimated mean base-rates of a right-handed taxon of 0.80 and 0.94 from the two analyses of the adult sample and of 0.79 and 0.93 from the two analyses of the high school sample. Subsequent analyses showed that the initial right-handed class could be subdivided into consistent and inconsistent right-handed subclasses in both samples. A supplementary analysis, including a measure of hand proficiency in a subset of the sample of Australian adults (N=596), also revealed a latent categorical structure, showing evidence for a categorical latent structure from a continuous measure of hand performance. We conclude that the distribution of hand preferences in humans is discrete and not continuous.

The dimensionality of the Edinburgh Handedness Inventory: An analysis with models of the item response theory

“…This meets the recommendation given by Ivo W. Molenaar (Fischer & Molenaar, 1995, p. 5) in the book he wrote together with Gerhard H. Fischer, Rasch Models: Foundations, Recent Developments, and Applications: ''Whenever possible, it is thus recommended to find a set of items that satisfies the RM [Rasch model].' ' Dragovic et al (2008) also used a shortened version of the EHI. In addition to the abovementioned items 8 and 10, they excluded the item Drawing hand, which is highly correlated with the writing hand.…”

“…Based on the earlier model concepts and taking into account the ''classic'' J distribution of preferred handedness, one would usually anticipate a maximum of 10% consistent left-handers and approximately 60Á70% consistent right-handers. In questionnaires containing several items, this is joined by an intermediate domain with 20Á30% socalled mixed-handers (Annett, 2002) or 9% inconsistent left-handers (Peters & Murphy, 1992) or 10Á29% inconsistent right-handers (Dragovic et al, 2008). A comparable classification into left-, right-and mixed-handers could also be confirmed with latent class analysis (Dragovic, 2004a).…”

“…However, the thresholds were unordered, indicating that participants did not use the categories in a systematic manner or that the items were not ordinal Rasch conforming. Moreover, one can also see that the class sizes of 52% (Class 1), 27% (Class 2), and 21% (Class 3) clearly deviated from other, broadly accepted handedness distributions (e.g., Dragovic, 2004a;Dragovic et al, 2008;Peters & Murphy, 1992).…”

“…Dragovic (2004a) pointed out that writing and drawing were either less informative or easier, whereas the items opening a box lid and using a broom had the highest informativeness or were more difficult (McManus, 1996). Nevertheless, the latter items have often been considered problematic (Bryden, 1977;Dragovic, 2004b;Dragovic, Milenkovic, & Hammond, 2008;McFarland & Anderson, 1980;Peters, 1990;Ransil & Schachter, 1994;Williams, 1986).…”

“…Likewise, the assumption of a J-shaped distribution of handedness also has to be rejected, because instead of a unidimensional distribution of handedness, it is necessary to define a cut-off point to distinguish between left-and right-handers on a theoretically and empirically sound basis. However, the underlying latent structure of handedness is rather categorical (Dragovic, et al, 2008;McManus, 1985) with quantitative differences of the preferred handedness and not a continuous trait (Annett, 1970).…”