Epidemiology and Determinants of the <i>Bouba-Kiki</i> Effect in Presence and at Distance

Abstract

When the bouba-kiki test, an evolution of the primordial baluma-takete test, was administered by an expert operator to 343 participants (225 women and 118 men), the rounded figure was associated to the name bouba in 86.9% and the angular one to the name kiki in 13.1% of cases ( p < 0.0001). In logistic regression, age was the only significant determinant of this answer ( p = 0.004), while sex and years of formal school education did not enter the model. By means of the receiver operating characteristic curves, a cut-off value of age (>47 years) was found and significantly entered logistic regression having rounded = bouba as dependent variable and sex and years of school as independent covariables ( p = 0.024). When, in sensitivity analysis, the same test was administered at distance via an Internet platform to other 62 comparable participants (21 women and 41 men) in the absence of an operator, the association of bouba to the rounded pictogram and of kiki to the angular one had a prevalence of 61.3% only ( p < 0.0001 vs. test in presence, i.e. p < 0.001 in men and p = 0.013 in women). Associating the name bouba to a rounded pictogram therefore seems to depend on age and on the physical presence of an operator, while sex and age of formal school education are irrelevant.