Gender schematicity, identity salience, and -linked language use
18
such that as the function increases references to emotion increase. The function is also moderately
correlated with judgmental adjectives (canonical coefficient = -.495), MLS (canonical coefficient =
-.479) and references to quantity (canonical coefficient = .494). As the function increases references
to emotion increase, while judgmental adjectives and MLS decrease. As previous research indicates
that references to emotion and long MLS are prototypically female language variables, and
references to quantity and judgmental adjectives are prototypically male language variables (Mulac
et al, 2001a), high scores on the discriminant function represent prototypical female language, while
low scores on the discriminant function represent prototypical male language. As the function
increases, prototypical female language increases and prototypical male language decreases.
Examination of the discriminant function score centroids for each of the eight groups helps to
determine the relative extent to which each group used the language variables that define the
discriminant function. Figure 1 plots the discriminant function score centroids for each of the eight
groups, using two different graphs for GS, Figure 1(a), and NGS individuals, Figure 1(b),
respectively. Table 6 provides the F tests for the pairwise group comparisons for the relative distance
of the eight groups’ discriminant function score centroids. As can be determined from Table 6 and
Figure 1(a), the language use of GS women with high GIS (group centroid = 1.15) and GS men with
low GIS (group centroid = .76) evidences prototypical female language use. On the other hand, the
language use of GS women with low GIS (group centroid = -.65) and NGS men with a low GIS
(group centroid = -.99) is strongly characterized by prototypical male language. In other words, GS
women with low GIS and NGS men with low GIS use the most prototypical male language, whereas
GS women with high GIS and GS men with low GIS use the most prototypical female language.
Those groups with language use in-between the two prototypical extremes are NGS women whose
gender identity is not salient (group centroid = .12), NGS women whose gender identity is highly
salient (group centroid = .003), GS men whose gender identity is highly salient (group
centroid = -.09), and NGS men whose gender identity is highly salient (group centroid = -.40).
Convention