A group is essentially a functional operator, working in the tide of the researcher, and the questions laid out by the paper's intention. One can meaningfully construct groupings according to sex, race, age, socioeconomic status. The most prominent form of group comparison designs is the randomized control trial, where participants are assigned to different conditions to isolate the causal effect of a variable, allowing researchers to observe differences between groups while controlling for confounding factors.

​

The traditional cross-cultural comparison, which is a methodological inheritance from the anthropological tradition of ethnography and comparative fieldwork, maintains that measurable behavioral tendencies can be aggregated to characterize the culture itself. This was employed to identify patterns of cognition, values, or social behavior in populations where direct experimental manipulation was impractical or impossible. 

​

Particular societies became X and Y Cultures, and the individual persons at the lowest rung of this hierarchy are taken up into the frame as members who are enculturated and exhibit certain orientations of the culture. This translation makes it possible to speak about psychological tendencies as if they were cultural properties, so that individual data is aggregated into evidential claims about the structure and direction of Culture X in contrast to Culture Y.

​

In psychology, as it was in the anthropological tradition, there is a sustained focus on comparison and typology. The obtained measures are descriptive and comparative, highlighting differences between groups without committing to any causal inference. 

But, at any level of analysis, psychology struggles with the problem of making sense of group differences. The strategic maneuver is to transform the descriptive features of the group into a kind of causal essence, which may be labeled as “culture” or some variant thereof. A difference observed between Culture X and Culture Z, from the samples of individuals labeled as “x” and “z,” is interpreted as if the disparity itself is caused by “the culture.” Individual behaviors are treated as manifestations of Culture-X-ness, as though the culture is an operative force producing the observed traits.

​

There are some grounds to recognize that culture is a complex and distributed system, without having to commit to the view that it is an independent and uniform entity. The next direction is to point our suspicion to the mechanisms and ask whether a supra-individual entity must be static and monolithic to effectuate any functional influence, or if it is possible that the intercalating networks of emergent practices can only be sustained through local interactions. 

​

A more nominalist position is to argue that culture is merely a product of our propensity to point and single out objects, reifying the patterns through an "initial baptism" of names (to borrow a phrase from Kripke); that is, the structure and the behavior of social practices are products of our labeling, which stabilizes regularities and attaches significance to the labels and descriptions that correspond to it. These labels only point to the amorphous and distributed nature of a concept of culture, and not that culture is itself a discrete entity, but is still useful as a heuristic for understanding recurring patterns.

​

The most defensible and perhaps common conception today is to suggest that culture is largely socially constructed and instantiated by the many social institutions that structure its practices and norms. But this only becomes intelligible once we identify its mechanisms and understand its effects through the ways it manifests in behavior and social institutions at large. The strongest evidence that culture exists and operates is the folk treatment of norms and cultural practices, where norms and practices are regarded as self-evident and causally effective through repeated enactment in local practice.

​

At any rate, these "Cultures" are not located idiographically. They are assumed to hold an operative authority in a hierarchy that positions them above the social institutions embedded in the society. The individual is a receptive and malleable agent, enculturated through the practices and norms that instantiate the manifested properties of the culture. In doing cross-cultural comparisons, the cultural essence is inferred through the behavioral tendencies exhibited by the individuals in their social contexts.

​

The mean difference of the properties or behaviors between groups is read as evidence that culture, as an entity, generates the variation between individuals. In effect, the analysis turns bottom-up observations into top-level explanations, projecting individual data into a collective essence that is assumed to act uniformly on all members.

​

The causal leap occurs when researchers obtain a statistical difference in the measured trait between two groups, and then interpret this difference as evidence of an underlying cultural pattern or mechanism that is operational in the social institutions of the population. The individual performances that generated the corpus of data are transformed into statements about the cognitive styles of Culture X and Culture Y, and in effect, projecting the distribution of responses in the sample of persons upward through the hierarchy, becoming an explanatory feature attributed to the culture itself.

​

In a more formal arrangement, the inference usually looks like this:

​

The will research begin with a simple, empirically verifiable statistical observation

E[P(x) | x ∈ G1] ≠ E[P(x) | x ∈ G2]

​

That "The mean of property P in Group 1 is not equal to the mean of property P in Group 2." By all measures, this is a neutral, mathematical description of the data.

​

The leap involves two conjoined and erroneous steps:

​

The first step involves reification, or the invention of the causal essence. Understanding that a statistical difference is obtained through measurement and comparison, the researcher reifies an abstract essence E(G1) that is defined solely by the group description.

​

E[P(x)∣x∈G1]≠E[P(x)∣x∈G2]  ⇒  ∃E(G1),E(G2) such that C(E(G1), P)

​

The second step is the careless leap of universal causal attribution to the observed properties. The researcher fallaciously infers that this invented essence E(G1) is the cause of the property P for every single member of the group.

​

C(E(G1),P) ⇒ ∀x(x∈G1 → C(E(G1),P(x))) which reads that "For all individuals x, if x is in Group 1, then the essence of Group 1 causes x to have property P." 

​

The premise is a statement about group aggregates (E[...]).

 

The conclusion is a universal statement about every individual within the group (∀x). One cannot deduce necessary truths about individuals from statistics about groups. There will always be variability within a group. Some individuals in G1 may not exhibit P at all, which would falsify the universal claim that ∀x....

​

In its most compressed form, the inference looks like this:

​

An individual from Culture Z performs behavior P. Aggregating across individuals, one finds a statistical difference in the average level of P between Group 1 (Culture Z) and Group 2 (Culture Y). The researcher interprets this difference as evidence that Culture Z, as such, produces P. Individual behavior P is now explained as the effect of “Culture-Z-ness.”