Men and women are from Earth Examining the latent structure of gender.

Understanding Dimensions and Taxa
One reason why the underlying nature of gender differences has
been difficult to address is that although biological sex is clearly a
categorical variable, the variables commonly of interest to re-
searchers and laypersons alike tend to be dimensional (e.g., mas-
culinity, femininity, school achievement, depression, aggression),
varying along a continuum. The statement that men are more
aggressive than women, for example, implicitly assumes that there
is one group of people who are high in aggression (men) and
another group of people who are low in aggression (women). This
assumption treats an observed mean difference between men and
women as a special kind of category called a taxon. Examples of
taxa include animal species (gophers vs. chipmunks), certain phys-
ical illnesses (e.g., one either has meningitis or not), and biological

no it doesnt. “men are more aggressive than women” has what logicians call a missing quantifier, meaning that one has to infer it from context. in this case it is pretty clear that the meant quantifier is “usually” or “typically”, which makes this sentence equivalent in meaning with “the average aggressiveness of men is higher than the ditto of women”. another quantifier cud be “all”, but no one seriously thinks that all men are more aggressive than all women. there is a difference in the average. i think that most people agree with this.


Although gender differences on average are not under dispute,
the idea of consistently and inflexibly gender-typed individuals is.
That is, there are not two distinct genders, but instead there are
linear gradations of variables associated with sex, such as mascu-
linity or intimacy, all of which are continuous (like most social,
psychological, and individual difference variables). Thus, it will be
important to think of these variables as continuous dimensions that
people possess to some extent, and that may be related to sex,
among whatever other predictors there may be. Of course, the term
sex differences is still completely reasonable. In a dimensional
model, differences between men and women reflect all the causal
variables known to be associated with sex, including both nature
and nurture. But at least with regard to the kinds of variables
studied in this research, grouping into “male” and “female” cate-
gories indicates overlapping continuous distributions rather than
natural kinds.

they seem confused. it does not follow that genders are not distinct just becus they indicators of the genders are dimensional rather than taxonomic. altho one cud think of the personality of people as being on a continuum from archtypical male to archtypical female.


This research also adds further evidence to the current debate
about whether it is more profitable to focus this literature on
gender differences or gender similarities (Hyde, 2005). “The gen-
der similarities hypothesis states, instead, that males and females
are alike on most—but not all—psychological variables” (Hyde,
2005, p. 590). Our research shows, moreover, that even those
variables on which males and females are not alike may be
evidence of variations along a continuous dimension rather than
categorical, and as Hyde terms them, “overinflated claims of
gender differences” (Hyde, 2005, p. 590). Clearly, if differences
between men and women are conceptualized as variations along a
continuum, there is little reason to reify these differences with the
sorts of extremities typically mentioned. Instead, these differences
would be seen as reflecting all the influences that are brought to
bear on an individual’s growth, development, and experience, and
would be relatively amenable to modification.

no such thing follows. gender differences can be small with lots of overlapping variation and still be 100% genetic, and thus not changeable with the usual socialization tools.


If gender is dimensional, why do categorical stereotypes of men
and women persist in everyday life? Although our research does
not speak to this issue, several explanations seem relevant. One
reason is that people tend to think categorically (Medin, 1989), or
as Fiske (2010) put it, referring to both laypeople and researchers,
“we love dichotomies” (p. 689). People use easily accessible
categories to help organize the abundance of information that the
social world presents, a mental shortcut that has come to be known
as the “cognitive miser” hypothesis (Fiske & Taylor, 1991). Be-
cause sex is one of the most readily observed human traits, it forms
an easy and common basis for categorizing other persons. As a
result, because other qualities tend to be accommodated to acces-
sible categories, and because men and women do differ in myriad
ways, category-based generalizations maximize the difference be-
tween the sexes while minimizing differences within them (e.g.,
Fiske & Neuberg, 1990; Taylor et al., 1978). Furthermore, as
Krueger, Hasman, Acevedo, and Villano (2003) showed, it may be
rational to accentuate intergroup differences whenever these dif-
ferences are easy to learn, fairly accurate, and helpful for action.

there are patterns in experience and in nature, and one sign of intelligence is to spot those patterns and use them to make decisions. stereotypes are useful for this.


It may be fruitful to consider how our findings are bound to the
cultural and historical context within which the data were col-
lected. With a few exceptions, most of these data were collected
from young Americans in the last quarter of the 20th century. This
is a time and setting in which differences between men and women
were shrinking, reflecting societal, economic, and educational
circumstances that contributed to the increasing liberalization of
gender roles (Brooks & Bolzendahl, 2004). Indeed, it seems likely
that were we to examine new data sets collected in 2012, they
would, if anything, be even more likely to be dimensional. This
point suggests two important implications. First, to the extent that
our data sets are outdated, they should have been more likely to
reveal a taxonic structure (which they did not), making our support
for dimensionality more compelling. Second, if suitable data sets
can be found, historical comparisons of underlying structures may
prove revealing of the impact of societal trends.

some things are shrinking, others are apparently increasing with increasing HDI. see


in a happy coincidence, i recently learned about, which is a site that guesses (stereotypes) about various things from one’s profile text on dating sites. they must have data that can create a bayesian probability distribution like those in the article.