This report introduces the Global Collectivism Index (GCI) – a measure covering 99.9% of the earth's population; collectivism is very high in Sub-Saharan Africa, very low in Western Europe, and intermediate in most other regions

A Truly Global, non-WEIRD Examination of Collectivism: The Global Collectivism Index. Brett Pelham et al. Current Research in Ecological and Social Psychology, December 29 2021, 100030. https://doi.org/10.1016/j.cresp.2021.100030

Highlights of Global Collectivism Index Report

• Despite recent growth in research on collectivism, there are no truly global measures of collectivism

• Our measure (the GCI) updates past measures and covers 99.9% of earth's population

• Our measure eliminates the strong WEIRD biases of past research

• Our measure is mostly behavioral and relies on representative national sampling

• Our measure compares favorably with existing measures

• Our measure is associated with important outcomes (e.g., suicide rates, human rights)

• This association is statistically independent of both national wealth (GDP) and modernity

• Our measure uncovers two important drivers of cultural evolution: farming and pathogen load

Abstract: This report introduces the Global Collectivism Index (GCI) – a measure covering 99.9% of the earth's population. The GCI includes six sub-scores (e.g., household living arrangements, ingroup favoritism). Collectivism is very high in Sub-Saharan Africa, very low in Western Europe, and intermediate in most other regions. Even after controlling for both national wealth and technological sophistication, national collectivism scores predict variables such as suicide rates, alcohol consumption, agricultural employment, and valuing child obedience. Further, this was true after directly pitting the GCI against several competing predictors of the major cultural outcomes examined in this report – from national wealth (GDP) and modernization to a seven-factor conceptualization of interdependence. The GCI is a much-needed, well-validated, historically updated measure that eliminates previous WEIRD biases and offers greatly increased statistical power in cross-cultural research.

A Final Look at Japan's Low Collectivism Scores

Many readers have been puzzled by our finding that Japan received a very low collectivism score. In fact, Japan's collectivism score on the GCI is identical to that of the United States. We are not the first to document low levels of collectivism in Asia. Oyserman et al. (2002) found that Japanese, South Korean, and U.S. citizens had similar collectivism scores. Such findings need not invalidate the bulk of the empirical observations that have been made in past comparisons of Japan (or South Korea) and the United States. But they do suggest that collectivism is an unlikely explanation for some of the recent cross-cultural differences observed between Japan and the United States. They also suggest that some of the past cross-cultural differences observed between these two nations might no longer appear in modern samples of young adults.

The assumption that Japan is a collectivistic nation is so widely accepted that many readers may question the validity of the GCI. But readers who reject our measure because of Japan's cross-national rankings would have to reject previously developed measures of collectivism as well. On Hofstede et al.’s (2010) popular measure of individualism, 39 out of 70 nations were more collectivistic (less individualistic) than Japan in 1970. On the GLOBE measure, Japan ranks 44th out of 55 nations in collectivism. Minkov's (2017) modern update on Hofstede's individualism measure places Japan one notch higher than the United States in global individualism. The WEIRD analyses presented here virtually guarantee that if existing measures of collectivism had included many more nations, Japan would have scored very low in collectivism. This is exactly what the GCI reveals. Further, as shown in Study 4, the fact that Japan has become much more individualistic in the past few decades is theoretically predictable, from drivers of cultural change.

Some avid believers in Japan's collectivism have argued that the GCI must be missing subscales that capture crucial aspects of collectivism. Such critics have yet to name any indicators that would radically change Japan's global rankings. For example, if one were to treat attitudes about child obedience as a subscale in our collectivism Index (rather than an outcome), this would lower Japan's collectivism score. There does appear to be one very good candidate for distinguishing Japan and other east Asian nations from the United States and Western Europe – in ways that are pretty consistent with much prior research on culture. This is Minkov et al.’s (2017) concept of flexibility-monumentalism. This cultural dimension has to do – at least in part – with the degree to which people behave very differently from one situation to another. It appears to overlap greatly with what Vignoles and colleagues call self-consistency (see Vignoles et al., 2016; especially Table 8). In fact, in Minkov's (2017) study of 56 nations, Japan had the world's highest score on this monumentalism-flexibility dimension (scoring strongly in the flexible direction). Minkov and colleagues specifically argue, however, that this dimension is largely independent of the dimension of individualism versus collectivism.

Brewer and Chen (2007) might argue that the present measure of collectivism – like other past measures of collectivism, allocentrism, and interdependence – conflates relational and group processes. These reflect two distinct forms of collectivism (Gabriel & Gardner, 1999). As applied to identity, relational identities refer to motives and self-evaluations that are grounded in close relationships and personal roles (e.g., aunt, grandfather). In contrast, group identities connect people to larger but less intimate social factions (e.g., one's fellow Moroccans, Cubs fan). Gabriel and Gardner (1999) found, for example, that men tend to privilege collective identities whereas women tend to privilege relational identities. From this viewpoint, the GCI seems to emphasize relational identities. This is especially true for total fertility, family living arrangements, and marriage to divorce ratios. But it is presumably much less true for religiosity – which connects people to large, extra-family groups. As operationalized in the GCI, ingroup bias is a roughly 2:1 blend of relational to collective concerns because Van de Vliert's measure includes familism and nepotism (which are heavily relational) and nationalism (which is about an extremely large group). Motor vehicle ownership rates may have more to do with truly individualistic motivations. Such motivations can be at odds with both relational and group motivations – and may differ less across modern cultures than do relational or collective motivations (Kreuzbauer, Chiu, & Lin, 2009).

General Discussion

In this report, we introduce the GCI – a theoretically-derived, empirically-validated national-level measure of collectivism. This single-factor measure has several desirable features. First, rather than relying on self-reported attitudes, the GCI consists mainly of behavioral measures, many of which can be assessed objectively. Second, unlike previous national-level measures of collectivism, the GCI covers virtually the entire planet. Moreover, most of the more than 100 nations that are covered by the GCI and not covered by existing measure are understudied, non-WEIRD cultures. The GCI also appears to be less strongly confounded with GDP and with modernity than are existing measures. Finally, taking nothing away from Minkov's (2017) update, the GCI represents a much-needed global update on levels of collectivism across the globe.

The GCI also has the advantage of being conceptually similar to Vandello and Cohen's (1999) measure of regional variation in collectivism across the 50 U.S. states. With this in mind, national scores on the GCI are reliably associated with national scores on both the likely origins and the likely consequences of collectivism. The GCI almost always predicted cross-national variation in social, political, and health outcomes even after we controlled for both GDP and “modernity” (technological and economic innovation).

Future studies should carefully address the likely origins and consequences of collectivism. The GCI will make that job much easier than it has been in the past. For example, one could use the GCI and Hofstede's original, 50-year-old, collectivism scores to model changes in collectivism over time. Doing so might help resolve the current debate about whether national wealth or pathogen loads better predict cultural evolution in the direction of individualism. Of course, this model would be limited to only about 55 nations, but we can think of no reason why such an analysis would not be at least as informative as the current studies that focus on many fewer nations (sometimes a single nation). We hope that even critics of this particular idea will agree that, at a bare minimum, the present data on the GCI show that the often-cited findings of Vandello and Cohen (1999) replicate very well at the global level.

Another way to appreciate the utility of the GCI is to see what would have happened if we had tested the hypotheses examined here using only the existing measures of collectivism. In such cases, we would have often observed incorrect or ambiguous findings. In a series of supplemental analyses, we assessed whether any of the existing measures of collectivism predicted how strongly parents valued obedience. All of these measures were correlated in the expected direction with this measure. However, only two of the four zero-order correlations were significant at p < .05 (those for the GLOBE and for Suh's measure). Further, after we statistically controlled for GDP, none of the measures was uniquely associated at p < .05 with valuing child obedience. Further, in four of the five cases, there was a significant unique effect for GDP. The main reason for this problem is clear. There was a median of 34 cases per regression analysis. This, combined with the limited statistical range for GDP and collectivism in heavily WEIRD nations, made it hard to separate collectivism and GDP. A nearly global pool of nations offers researchers much-needed statistical power when it comes to separating GDP and collectivism.