Two-Dimensional Models of Cultural Differences: Statistical and Theoretical Analysis

East-West Differences

A research team named Chinese Culture Connection were the first to make quantitative studies of cultural differences between East Asian cultures and Western cultures. They tried to identify characteristic aspects of East Asian culture with a survey including typical Chinese values. A factor analysis of the results yielded four factors. The factor that best captured East Asian culture was named Confucian work dynamism. This factor was positively related to items like persistence, thrift, and sense of shame (Chinese culture connection, 1987). Hofstede and coworkers further developed this factor and renamed it long-term orientation vs. short-term orientation in their study of business organizations. Long-term orientation includes the virtues oriented toward future rewards, in particular perseverance, thrift, humility, and deferred gratification. Short-term orientation is related to the past and present, including values relating to tradition, respecting social codes, fulfilling social obligations, immediate gratification, self-assertion, and reciprocation of gifts (Hofstede et al., 2010). The long-term orientation of the so-called Asian Tiger countries is theorized to have contributed to their high economic growth in the late 20’th century, while short-term orientation is claimed to have led to the 2008 economic crisis in the West (Hofstede et al., 2010).

The cultural differences between East Asian countries and the West have been researched for several decades, and the relevant variables have been renamed and redefined several times. In 2011, Minkov defined a variable named monumentalism vs. flexumility that reflects these differences. 'Flexumility' is a contraction of flexibility and humility, referring to a flexible self-identity that allows mixed feelings or duality and flexible norms. Flexumility is characteristic of East Asian cultures. The opposite end of the scale, named monumentalism, refers to immutable identities, fixed norms, pride, and strong religiousness, which can be observed for example in some Arab countries (Minkov, 2011). A later revision of the theory uses the name flexibility vs. monumentalism. This is a reconceptualization of Confucian work dynamism and long-term orientation, and reflects cultural differences between East Asian cultures at one extreme and Latin American and some African cultures at the other extreme (Minkov, 2018; Minkov et al., 2018). Another study by Minkov defined a related variable that he named K factor and hypometropia. This factor refers to a short-term oriented reproductive strategy, risk acceptance, and intra-communal violence. This factor is high in sub-Saharan Africa and Latin America, and low in East Asia. This factor is linked to geography, economy, and national behavioral statistics (Minkov, 2014).

A survey of what people would do with their money if they were rich resulted in another factor named ego boosting vs. altruism that had a high correlation with East-West differences. People in East Asia tend to score at the ego boosting end of the scale which involves spending money on expensive things, ostentatious parties, gaining political power, or saving the money. The altruism end of the scale involves donating money or investing in business (Minkov et al., 2019).

A study of social axioms or world views shows that East Asian countries score high on fate control, i.e., a belief that life events are pre-determined by fatalistic forces, but that people may be able to predict and alter their fate by various means. This is not surprising since such beliefs are part of the concept of karma in Buddhist tradition (Stankov & Saucier, 2015).

Bond and Lun (2014) have studied the socialization of children. They found that a 2-factor solution was the most appropriate. Their two factors are named self-directedness vs. other-directedness and civility vs. practicality. Self-directedness is marked by socializing children for the qualities of determination, perseverance, responsibility, independence, and imagination, while other-directedness is marked by socialization for religious faith and obedience. Civility is marked by socialization for tolerance and respect for other people and unselfishness, whereas practicality is marked by socialization for thrift and saving money and things. North European countries are high on self-directedness, and East Asian countries are high on practicality.

We can improve the theoretical understanding of East-West differences by applying the theory of relational mobility. Relational mobility is the degree to which persons have freedom to choose interpersonal relationships and to leave unsatisfactory social settings (Yuki & Schug, 2012). People in societies with low relational mobility are oriented towards long-term interpersonal relationships. Therefore, they are careful to avoid social conflicts and to guard their reputation. They tend towards an externalizing mode of thinking that attributes outcomes to forces outside their control, such as fate, luck, and other people. High relational mobility, on the other hand, is associated with analytical thinking, independent self-construal, self-enhancement, internal locus of control, and interpersonal trust (San Martin et al., 2019). Relational mobility is low in societies that practice settled, interdependent subsistence styles, such as irrigated rice farming in East Asian countries. Relational mobility is high in societies that rely little on farming, including nomadic herding cultures and urban industrial cultures (Thomson et al., 2018). We can expect relational mobility to be positively correlated with self-expression, emotional display, social polarization, internal locus of control, and sociosexuality. Societies with low relational mobility can be expected to have higher belief in fate, less trust in strangers, and high values on variables that were designed to match aspects of East Asian culture.

While there has been considerable research focus on the characteristics of East Asian cultures, there has been less focus on the opposite extremes. Scientists studying the scales that are supposed to measure characteristics of East Asian cultures have often found Latin American countries and some African countries at the opposite end of these scales. Traditional theory would predict East Asian cultures to have important similarities with Latin American cultures because both are collectivistic. But these two groups of countries are in fact found at opposite extremes of many of the variables that characterize East Asian values, such as long-term orientation, flexibility vs. monumentalism, relational mobility, etc. Scientists have shown little interest in this apparent anomaly until recently.

A new study by Krys and coworkers is trying to explain the differences between East Asian and Latin American cultures based on differences in mode of subsistence, colonial history, ethnic diversity, and religion (Krys et al., 2022). Different modes of subsistence have a high influence on cultural variables. Latin American cultures have historically relied more on herding than on farming, which involves a higher relational mobility. Differences in colonial history are also important. Several East Asian cultures experienced occupation and exploitation by European colonizing nations, where indigenous rule was returned after independence in most cases. Latin American countries were more often subject to settler colonialism in which indigenous populations were partially displaced. The colonizers that were most successful in settling on new frontiers were most likely the ones that were most self-reliant and self-directed. These self-perceptions have likely influenced their culture.

Different construals of selfhood are important elements in the theory of Krys et al. (2022). The construal of selfhood is also influenced by ethnic and cultural diversity. East Asian societies are among the most ethnically and culturally homogeneous in the world, while Latin American societies are very heterogeneous. Cultural homogeneity may foster an interdependent selfhood in East Asia, while the heterogeneity in Latin America allows a more independent and unique perception of self. These differences are amplified by religion and philosophy. The Confucian tradition in East Asia is advocating interdependence and social harmony with others. Latin American culture, on the other hand, is dominated by Christianity which gives humans a dominant position over nature and a view of humans as independent, self-directed, and self-reliant (Krys et al., 2022).

The differences in self-construal are also related to the cultural logics of honor vs. face. Cultures of honor are typically found in societies with weak law enforcement where people must be able to protect themselves or rely on their family for retaliation against wrongdoing. People gain honor by proving their ability and readiness to protect themselves and their family. The situation is very different in cultures where a person’s worth is based on protecting one’s ‘face’ rather than honor. The ‘face logic’ is typical of cooperative societies with a settled hierarchy and low relational mobility. Here, it is important to maintain social harmony by obeying norms. People lose face if misbehavior is detected (Leung & Cohen, 2011; Krys et al., 2022). The concept of ‘face’ is of Chinese origin, and the importance of ‘face’ is a well-known aspect of Chinese culture (Ho, 1976). The cultural face logic is related to long-term orientation and flexibility, while honor logic is associated with short-term orientation and monumentalism (Minkov et al., 2019). Earlier literature strangely shows a strong negative correlation between saving face and East Asian cultural values (Chinese Culture Connection, 1987; Hofstede et al., 2010), but this is probably a mistake since it contradicts all that is known about saving face.

Data-Driven Versus Theory-Driven Research

Previous studies of cultural differences have often relied mainly on a data-driven approach. A typical research method is to make a factor analysis of survey responses and try to make sense of the resulting factors and build a theory that can explain the results. An alternative to the data-driven approach is a theory-driven approach (Maass et al., 2018). An example of the latter approach is the study by Fog (2021) who first predicted the factor F1 based on theory, and then made statistical tests to verify its existence.

Many of the variables in the literature are dominated by a data-driven approach. For example, the variable named “traditional vs. secular-rational values” originates from a rotated factor analysis result. The name of this variable and the associated theory came as an attempt to interpret the observed result. A few of the variables are more theory-driven. For example, the variables “relational mobility” and “regality” have their origins in causal theories, while the statistical support came second.

The mainly data-driven approach sometimes leads to functional explanations. For example, the cultural response to various threats has been explained as follows: “Nations facing these particular challenges are predicted to develop strong norms and have low tolerance of deviant behavior to enhance order and social coordination to effectively deal with such threats.” (Gelfand et al., 2011). This begs the question of how these norms have developed and why people obey strong norms only when their nation is facing particular threats.

Ecological explanations are more informative. For example, Welzel’s theory of the cool-water condition, as explained above, provides a plausible explanation of how people in the North have escaped centralized control when they do not depend on organized irrigation.

It is useful to make a distinction between proximate and ultimate causes here. A proximate cause of why norms are strict in a certain society could be, for example, that some influential leader has decided so. An ultimate explanation would look at mechanisms in cultural evolution or psychological responses to environmental conditions. Theories based on ultimate causes are of course preferred. These theories may be developed by data-driven or theory-driven research, or a combination of both.

The present study is using both theory-driven and data-driven approaches. The factor F1 has been predicted from regality theory (Fog, 2021) and this prediction is tested with statistical methods. The factor F2 was not predicted in advance but discovered by data analysis. We are now seeking theories that can explain the factor F2.

The aim of the present study is to go beyond the popular descriptive and interpretive approach and search for ultimate causes and theories that link cultural differences to differences in the social and physical environment. We will use these theories for seeking deeper theoretical insight into the causal relationships that lead to the observed factor structure.

Data and Methods

Data were collected from published cross-cultural studies of world cultures, including organizational culture. The criteria for inclusion were that the studies must report contemporary quantitative cultural data for at least twenty different countries based on population surveys. Variables were included if they attempt to measure cultural differences reflected in personal attitudes and behaviors reported by the respondents, while behaviors and characteristics reported in national statistics were not included. Variables were included only if the published study intended to measure cultural differences. Country averages of personality measures were not included because these instruments are not designed for measuring culture and they have poor reproducibility at the country level (Meisenberg, 2015). The criteria for inclusion in this study did not set any requirements for data quality, sample size, or representativeness. Theory-related selection criteria were generally avoided in order to prevent expectation bias, but studies that explicitly attempt to replicate the two-factor structure reported by Fog (2021) were excluded in order to avoid boosting the expected factor structure.

92 cultural variables from 33 published cross-cultural studies including 125 countries were found to meet the inclusion criteria. This data set includes the variables used in the study of Fog (2021) and several more. All variables are listed in Table 1. Two similar measures of religiosity are based on the same data and published in articles by the same first author (Stankov & Lee, 2015; Stankov & Saucier, 2015). These two religiosity measures were combined into one. Three variables originally published by Schwartz (1994) were instead taken from Maleki and de Jong (2014) because the latter article included more countries. Likewise, a measure of cultural tightness from the study of Gelfand et al. (2011) were replaced by similar data from Eriksson et al. (2021) because the latter study included more countries. Three cultural measures of selfhood (Owe et al., 2013) were replaced by seven revised measures by the same group of authors based on the same data (Vignoles et al., 2016).

OPEN IN VIEWER

Table 1.

Cultural Variables and Background Variables.

Short name	Full name or explanation	Number of countries	Reference
National culture
Autonomy A	Affective autonomy	31	Schwartz, 1994
Autonomy I	Intellectual autonomy	31	Schwartz, 1994
Civility/practicality	Civility vs. practicality	55	Bond & Lun, 2014
Complexity	Social complexity	33	Stankov & Saucier, 2015
Conservatism S	Conservatism	31	Schwartz, 1994
Conservatism SLV	Conservatism vs. liberalism	33	Stankov et al., 2014
Consistency	Consistency vs. variability	34	Vignoles et al., 2016
Cynicism B	Societal cynicism	42	Bond et al., 2004
Cynicism SS	Social cynicism	33	Stankov & Saucier, 2015
Discipline vs violence	Discipline vs. violence	33	Fog, 2017
Distrust	Distrust vs. trust	102	Beugelsdijk & Welzel, 2018
Egalitarian S	Egalitarian commitment	31	Schwartz, 1994
Egalitarian SDT	Egalitarian commitment vs. conservatism	43	Smith et al., 1995
Ego boosting	Ego boosting vs. altruism	35	Minkov et al., 2019
Emancipative W	Emancipative values	96	Welzel, 2013
Embeddedness SM	Embeddedness vs. autonomy	36	Schwartz, 2006, Maleki & de Jong, 2014
Emotional display	Emotional display rules, overall expressivity	37	Matsumoto et al., 2008
Externality	Dynamic externality	42	Bond et al., 2004
Familialism	Familialism	38	Minkov & Hofstede, 2012
Fate control	Fate control	33	Stankov & Saucier, 2015
Flexibility/monumentalism	Flexibility vs. monumentalism	54	Minkov et al., 2018
Gender S	Gender egalitarianism	33	Stankov, 2015
Harmony	Harmony	31	Schwartz, 1994
Harshness	Harshness vs. softness	34	Stankov et al., 2014
Hedonism	Hedonism vs. conservatism	35	Minkov et al., 2019
Hierarchy F	Hierarchical dominance values	28	Fischer, 2013
Hierarchy SM	Hierarchy vs. egalitarianism	36	Schwartz, 2006, Maleki & de Jong, 2014
Humane S	Humane orientation	33	Stankov, 2015
Individualism B	Individualism vs. collectivism	104	Beugelsdijk & Welzel, 2018
Individualism M	Individualism vs. collectivism	55	Minkov et al., 2017
Industry/indulgence	Industry vs. indulgence	43	Minkov, 2011
Joy vs duty	joy vs. duty	104	Beugelsdijk & Welzel, 2018
K factor	K factor and hypometropia (reflects differences in mating strategies, risk-acceptance, and time orientation)	71	Minkov, 2014
LOC individual	Individual-social locus of control	43	Smith et al., 1995
LOC political	Political-personal locus of control	43	Smith et al., 1995
Long term MH	Long-term orientation	38	Minkov & Hofstede, 2012
Loyal	Loyal vs. utilitarian involvement	43	Smith et al., 1995
Manhood	Precarious manhood beliefs	34	Bosson et al., 2021
Mastery SM	Mastery vs. harmony	36	Schwartz, 2006, Maleki & de Jong, 2014
Monumentalism	Monumentalism vs. flexumility (flexibility/humility)	43	Minkov, 2011
Morality SL	Morality	33	Stankov & Lee, 2015
Nastiness	Nastiness	33	Stankov & Lee, 2015
Polarization	Social polarization	47	Minkov, 2009
Power S	Power distance	33	Stankov, 2015
Regality	Regality	33	Fog, 2017
Relational mobility	Relational mobility	38	Thomson et al., 2018
Religiosity comb	Religiosity (two similar studies combined)	33	Stankov & Lee, 2015, Stankov & Saucier, 2015
Reward	Reward for Application	33	Stankov & Saucier, 2015
Secular I	Secular/rational vs. traditional values	97	Inglehart & Welzel, 2005
Secular W	Secular values	94	Welzel, 2013
Self containment	Self-containment vs. connection to others	34	Vignoles et al., 2016
Self difference	Self difference vs. Similarity	34	Vignoles et al., 2016
Self directedness B	Self-directedness vs. other-directedness	55	Bond & Lun, 2014
Self direction V	Self direction vs. reception to influence	34	Vignoles et al., 2016
Self expression	Self-expression vs. survival val	97	Inglehart & Welzel, 2005
Self expr V	Self-expression vs. harmony	34	Vignoles et al., 2016
Self interest	Self-interest vs. commitment to others	34	Vignoles et al., 2016
Self reliance	Self-reliance vs. dependence on others	34	Vignoles et al., 2016
Sociosexuality	Sociosexuality	46	Schmitt, 2005
Tightness C	Cultural tightness, combined	68	Uz, 2015
Tightness E	Tight vs. loose cultures	32	Eriksson et al., 2021
Uncertainty S	Uncertainty avoidance and future orientation	33	Stankov, 2015
Organizational variables
Assertive AI	Assertiveness, as is	58	House et al., 2004
Assertive SB	Assertiveness, should be	58	House et al., 2004
Future AI	Future orientation, as is	58	House et al., 2004
Future SB	Future orientation, should be	58	House et al., 2004
Gender AI	Gender egalitarianism, as is	58	House et al., 2004
Gender SB	Gender egalitarianism, should be	58	House et al., 2004
Humane AI	Humane orientation, as is	58	House et al., 2004
Humane SB	Humane orientation, should be	58	House et al., 2004
Individualism H	Individualism vs. collectivism	69	Hofstede et al., 2010
Indulgence H	Indulgence vs. restraint	93	Hofstede et al., 2010
Ingroup AI	In-group collectivism, as is	58	House et al., 2004
Ingroup SB	In-group collectivism, should be	58	House et al., 2004
Institutional AI	Institutional collectivism, as is	58	House et al., 2004
Institutional SB	Institutional collectivism, should be	58	House et al., 2004
Long term H	Long term orientation	92	Hofstede et al., 2010
Masculinity	Masculinity vs. femininity	69	Hofstede et al., 2010
Performance AI	Performance orientation, as is	58	House et al., 2004
Performance SB	Performance orientation, should be	58	House et al., 2004
Power H	Power distance	69	Hofstede et al., 2010
Power AI	Power distance, as is	58	House et al., 2004
Power SB	Power distance, should be	58	House et al., 2004
Src of guidance	Vertical source of guidance	36	Smith et al., 2002
Uncertainty H	Uncertainty avoidance	69	Hofstede et al., 2010
Uncertainty AI	Uncertainty avoidance, as is	58	House et al., 2004
Uncertainty SB	Uncertainty avoidance, should be	58	House et al., 2004
Variables not included in factor analysis (< 25 countries)
Confidence	Confidence in government	21	Fischer, 2013
Confucian	Confucian work dynamism	22	Chinese culture conn. 1987
Human hearted	Human heartedness	21	Chinese culture conn. 1987
Integration	Integration	22	Chinese culture conn. 1987
Moral C	Moral discipline	22	Chinese culture conn. 1987
Background variables
Action resources	Intellectual, material, and connective resources	102	Echeverría et al., 2019
Competitiveness	Global competitiveness	108	Schwab, 2017
Conflict score	log(ConflictScore+0.1)	99	Echeverría et al., 2019
Corruption	Corruption Perceptions Index 2018 (sign changed)	119	Transparency International, 2019
Democracy	Democracy index	119	The Economist Intelligence Unit, 2019
Education	Component of HDI	119	United Nations Development Programme, 2018a
Freedom C	The Human Freedom Index, 2018	117	Vásquez & Porčnik, 2018
Freedom H	Freedom score	120	Freedom House 2019
Gender development	Gender Development Index	117	United Nations Development Programme, 2018b
Gender inequality	Gender Inequality Index	113	United Nations Development Programme, 2018c
Gini	Gini index of income inequality	110	World Bank, 2019
GNI	Component of HDI log(GNI/capita)	119	United Nations Development Programme, 2018a
Happiness	Happiness	120	Helliwell et al., 2019
HDI	Human development index, 2018	123	United Nations Development Programme, 2018a
Life expectancy	Component of HDI	119	United Nations Development Programme, 2018a
Life satisfaction	Better life index	40	OECD, 2017
Pathogen stress	Combined “parasite” stress	116	Fincher & Thornhill, 2012
Territory deaths	Deaths in own territory due to organized violence, including war, civil war, and terrorism, 1993–2017. log(deaths+0.1)	58	Uppsala Conflict Data Program, 2018
Territorial threats	External territorial threats	26	Fischer, 2013

Background variables were gathered because correlations with background variables may help in the interpretation and explanation of cultural factors. An important background variable is the human development index (HDI) and its three components: Life expectancy, Education, and Gross National Income (GNI) per capita. Other economic and political indicators include Gini coefficient, competitiveness, corruption, democracy index, two measures of freedom, and two indexes of gender relations (see Table 1). Psychological measures of happiness and life satisfaction were added. Pathogen stress is measuring the prevalence of infectious diseases. The literature is often using the misleading term parasite stress, but the figures include virus infections which are not caused by parasites. Two measures of violent conflict were also included: deaths in own territory, and a compound conflict score. Territory deaths and conflict score were log-transformed because the distributions of these variables were closer to a log normal distribution than to a normal distribution. A value of 0.1 was added to the index to avoid taking the logarithm of zero. Finally, a measure of action resources (as defined by Welzel, see above) influenced by violent conflict were added to the list (Echeverría et al., 2019). The background variables are listed in the lower part of Table 1.

Two studies were made based on these data. Study 1 combined variables representing national culture and organizational culture, while study 2 separated these variables into two datasets.

Study 1

The different studies included different subsets of the 125 countries, and only two countries were represented in all the studies. A reduced data set with fewer missing values was made by removing variables reported for less than 25 countries, and removing countries that were represented in less than half of the remaining variables. The reduced data set had 87 variables and 47 countries, with 25% missing values. A traditional factor analysis cannot handle a data set with more variables than countries because the covariance matrix becomes singular. Another problem with a traditional factor analysis is that it cannot handle missing data. Instead, we made an exploratory factor analysis of the reduced data set by full information maximum likelihood factor analysis using structural equation modeling with the factors as latent variables. This is a method that finds the best fitting factor model based on all the data supplied. This method is able to handle both the missing data problem and the situation with more variables than observations (Bates et al., 2019; Hirose et al., 2016).

A scree plot of the reduced data set shows three eigenvalues above the randomness line and two additional eigenvalues right below the randomness line (Figure 2). This suggests that a solution with three to five factors is adequate. Different rotation methods were tried.OPEN IN VIEWER

Quartimax rotation was found to give the result that was easiest to interpret. This rotation method tends to emphasize dominating factors rather than divide the loadings between multiple factors. This gives a result that is close to the unrotated solution (principal axes). Factor loadings for the 87 cultural variables are shown in Table 2, using two factors. These two factors account for 34% and 15%, respectively, of the total variance in the variables. Solutions with up to five factors and other rotations are listed in an Online Appendix. The additional factors account for 11%, 6%, and 6% of the variance.

OPEN IN VIEWER

Table 2.

Factor Loadings, Quartimax Rotation.

Variable	F1	F2
Variables that load mostly on F1
Regality	−0.98	−0.01
Conservatism SLV	−0.98	0.32
Familialism	−0.94	0.09
Emancipative W	0.94	−0.13
Hedonism	0.93	−0.24
Embeddedness SM	−0.93	0.06
Religiosity Comb	−0.92	0.22
Individualism B	0.91	−0.24
Individualism M	0.91	−0.30
Morality SL	−0.90	0.28
Externality	−0.90	0.17
Self difference	0.86	0.28
Ingroup AI	−0.84	0.07
Conservatism S	−0.82	0.04
Tightness C	−0.81	−0.19
Uncertainty SB	−0.81	0.12
Hierarchy F	−0.80	−0.38
Self expression	0.78	0.33
Self expr vs harmony	0.78	0.38
Industry	−0.78	−0.38
Autonomy I	0.77	−0.16
Power H	−0.76	0.01
Gender S	0.72	0.00
Self direction V	0.71	−0.17
Future SB	−0.71	0.36
Hierarchy SM	−0.70	−0.42
Self containment	0.70	0.00
Individualism H	0.70	0.04
Self directedness B	0.69	−0.57
Egalitarian S	0.68	0.25
Egalitarian SDT	0.68	0.41
Sociosexuality	0.67	0.46
Src of guidance	−0.67	−0.03
Gender SB	0.67	0.30
Autonomy A	0.62	0.06
Nastiness	−0.61	−0.28
Joy vs duty	0.60	0.36
LOC individual	0.56	0.18
Relational mobility	0.54	0.53
Power AI	−0.53	0.17
Reward	−0.48	0.26
Humane S	−0.48	0.09
Power S	−0.47	0.26
Self reliance	−0.46	0.31
Uncertainty AI	0.44	−0.23
Harmony	0.43	0.20
Self interest	−0.43	0.27
Mastery SM	−0.42	−0.22
Tightness E	−0.41	−0.34
Humane AI	−0.40	0.04
Complexity	0.38	−0.26
Loyal	−0.35	−0.02
Gender AI	0.33	0.13
LOC political	0.32	0.24
Cynicism SS	−0.21	−0.14
Masculinity	−0.20	0.07
Humane SB	0.14	0.09
Variables that load mostly on F2
Flexibility/monumentalism	0.26	−0.86
Ego boosting	0.27	−0.84
Long term H	0.19	−0.83
K factor	−0.02	−0.77
Secular I	0.58	−0.73
Monumentalism	−0.49	0.72
Long term MH	0.03	−0.71
Discipline	0.24	−0.71
Civility vs practicality	0.44	0.70
Performance SB	−0.05	0.67
Fate control	−0.56	−0.65
Consistency	−0.27	0.65
Secular W	0.49	−0.61
Ingroup SB	−0.11	0.57
Harshness	−0.51	−0.54
Institutional AI	0.04	−0.49
Emotional display	0.19	0.49
Distrust	0.05	−0.49
Polarization	0.17	0.47
Institutional SB	−0.39	0.47
Indulgence H	0.42	0.45
Cynicism B	−0.26	−0.39
Performance AI	−0.10	−0.36
Assertive SB	−0.31	−0.36
Power SB	−0.27	−0.34
Uncertainty S	−0.16	−0.26
Future AI	0.22	−0.24
Uncertainty H	−0.04	0.23
Assertive AI	−0.08	0.22
Manhood	−0.06	−0.07
Proportion of variance	0.34	0.15

The correlations of all 92 cultural variables against the factors were calculated using pairwise complete observations. The correlation results for a two-factor solution are shown in Table 3. Correlations for three and five factors are listed in the Online Appendix.

OPEN IN VIEWER

Table 3.

Correlation of Cultural and Background Variables against Factors and HDI.

Variable	F1	F2	HDI
Regality	−0.87***	−0.08	−0.70***
	−0.73***	−0.40*
Conservatism SLV	−0.85***	0.35	−0.80***
	−0.62**	0.1
Familialism	−0.91***	0.17	−0.70***
	−0.82***	−0.12
Emancipative W	0.94***	−0.17	0.78***
	0.85***	0.21
Hedonism	0.88***	−0.21	0.82***
	0.70***	0.16
Embeddedness SM	−0.87***	−0.04	−0.55***
	−0.84***	−0.16
Religiosity Comb	−0.82***	0.19	−0.65***
	−0.65***	0.01
Individualism B	0.91***	−0.27	0.76***
	0.80***	0.02
Individualism M	0.89***	−0.3	0.89***
	0.72***	0.07
Morality SL	−0.81***	0.26	−0.69***
	−0.59**	0.09
Externality	−0.82***	0.14	−0.62***
	−0.69***	−0.12
Self difference	0.83***	0.23	0.59**
	0.76***	0.60**
Ingroup AI	−0.87***	0.14	−0.62***
	−0.78***	−0.11
Conservatism S	−0.81***	−0.05	−0.38*
	−0.80***	−0.17
Tightness C	−0.81***	−0.01	−0.64***
	−0.65***	−0.38*
Uncertainty SB	−0.85***	0.21	−0.69***
	−0.70***	−0.06
Hierarchy F	−0.71***	−0.43*	−0.58**
	−0.50*	−0.60**
Self expression	0.76***	0.28	0.50***
	0.68***	0.58***
Self expr vs harmony	0.76***	0.33	0.48*
	0.72***	0.63**
Industry	−0.77***	−0.33	−0.51**
	−0.70***	−0.63***
Autonomy I	0.67***	−0.05	0.41*
	0.58**	0.07
Power H	−0.76***	0.02	−0.63***
	−0.55***	−0.24
Gender S	0.61***	0.09	0.58**
	0.31	0.32
Self direction V	0.74***	−0.28	0.59**
	0.55*	−0.07
Future SB	−0.73***	0.42**	−0.64***
	−0.50***	0.27
Hierarchy SM	−0.72***	−0.49**	−0.58***
	−0.53***	−0.74***
Self containment	0.70***	−0.12	0.52*
	0.56*	0.09
Individualism H	0.71***	0.04	0.51***
	0.58***	0.23
Self directedness B	0.67***	−0.57***	0.78***
	0.12	−0.59***
Egalitarian S	0.69***	0.34	0.43*
	0.60***	0.52**
Egalitarian SDT	0.70***	0.42*	0.43*
	0.64***	0.63***
Sociosexuality	0.50**	0.44*	0.23
	0.49**	0.63***
Src of guidance	−0.69***	−0.06	−0.42*
	−0.63***	−0.23
Gender SB	0.68***	0.31*	0.44**
	0.59***	0.55***
Autonomy A	0.60***	0.1	0.15
	0.67***	0.14
Nastiness	−0.57**	−0.35	−0.41*
	−0.44*	−0.53**
Joy vs duty	0.57***	0.30*	0.42**
	0.43**	0.53***
LOC individual	0.60***	0.15	0.37*
	0.53**	0.29
Relational mobility	0.47*	0.55**	0.05
	0.62***	0.72***
Power AI	−0.53***	0.24	−0.50***
	−0.28	0.08
Reward	−0.47*	0.3	−0.36
	−0.32	0.22
Humane S	−0.44*	0.03	−0.25
	−0.40*	−0.04
Power S	−0.45*	0.32	−0.40*
	−0.25	0.24
Self reliance	−0.48*	0.44	−0.37
	−0.33	0.35
Uncertainty AI	0.46**	−0.33*	0.34*
	0.34*	−0.24
Harmony	0.39*	0.31	0.21
	0.34	0.39*
Self interest	−0.44	0.28	−0.4
	−0.21	0.15
Mastery SM	−0.41*	−0.3	−0.22
	−0.38*	−0.36*
Tightness E	−0.42*	−0.41*	−0.31
	−0.29	−0.55***
Humane AI	−0.44**	0	−0.36*
	−0.27	−0.15
Complexity	0.24	−0.21	0.42*
	−0.13	−0.11
Loyal	−0.38*	−0.02	−0.07
	−0.50**	−0.04
Gender AI	0.30*	0.18	0.28
	0.14	0.32*
LOC political	0.36*	0.28	0.17
	0.35	0.35*
Cynicism SS	−0.21	−0.11	−0.17
	−0.12	−0.17
Masculinity	−0.13	0.01	−0.12
	−0.06	−0.03
Humane SB	0.17	0.07	0.07
	0.17	0.11
Flexibility/monumentalism	0.22	−0.88***	0.59***
	−0.43**	−0.88***
Ego boosting	0.23	−0.88***	0.51***
	−0.25	−0.87***
Long term H	0.23	−0.83***	0.44**
	−0.16	−0.80***
K factor	0	−0.77***	0.43**
	−0.56***	−0.72***
Secular I	0.59***	−0.73***	0.67***
	0.19	−0.70***
Monumentalism	−0.57***	0.78***	−0.72***
	−0.01	0.81***
Long term MH	0.1	−0.87***	0.39
	−0.31	−0.85***
Discipline	0.2	−0.79***	0.36
	−0.13	−0.77***
Civility vs practicality	0.48**	0.69***	0.2
	0.54**	0.79***
Performance SB	−0.09	0.69***	−0.29
	0.2	0.65***
Fate control	−0.54**	−0.74***	−0.1
	−0.72***	−0.80***
Consistency	−0.28	0.74***	−0.37
	0.04	0.70***
Secular W	0.47**	−0.56***	0.63***
	−0.01	−0.45**
Ingroup SB	−0.17	0.58***	−0.16
	−0.07	0.57***
Harshness	−0.48*	−0.68***	−0.17
	−0.54**	−0.82***
Institutional AI	0.02	−0.60***	0.17
	−0.16	−0.59***
Emotional display	0.24	0.44*	−0.19
	0.59**	0.41*
Distrust	0.08	−0.56***	0.14
	−0.05	−0.55***
Polarization	0.13	0.46*	−0.19
	0.50**	0.43*
Institutional SB	−0.40**	0.54***	−0.43**
	−0.13	0.46**
Indulgence H	0.40**	0.42**	0.16
	0.43**	0.53***
Cynicism B	−0.26	−0.37*	−0.08
	−0.29	−0.43*
Performance AI	−0.07	−0.45**	0.13
	−0.25	−0.43**
Assertive SB	−0.33*	−0.42**	−0.12
	−0.37*	−0.49***
Power SB	−0.27	−0.39**	−0.03
	−0.37*	−0.42**
Uncertainty S	−0.16	−0.33	0.06
	−0.32	−0.32
Future AI	0.24	−0.34*	0.22
	0.11	−0.29
Uncertainty H	−0.02	0.29	−0.09
	0.08	0.27
Assertive AI	−0.02	0.28	−0.08
	0.06	0.26
Manhood	−0.06	−0.02	0.11
	−0.26	0.01
Variables not included in factor analysis
Integration	0.81***	0.01	0.57**
	0.70***	0.40
Confucian	−0.24	−0.73***	0.23
	−0.59**	−0.72***
Human hearted	0.15	0.02	0.12
	0.09	0.09
Moral C	−0.44*	−0.36	−0.11
	−0.52*	−0.48*
Confidence	−0.52*	−0.49*	−0.43
	−0.33	−0.60**
Background variables
HDI	0.74***	−0.37*	1.00***
	n.a	n.a
Life expectancy	0.65***	−0.34*	0.92***
	−0.17	−0.04
Education	0.79***	−0.28	0.94***
	0.33*	0.18
GNI	0.68***	−0.39**	0.96***
	−0.27	−0.20
Gini	−0.48**	0.45**	−0.58***
	−0.01	0.34*
Democracy	0.76***	−0.16	0.68***
	0.52***	0.13
Freedom C	0.73***	−0.37*	0.79***
	0.35*	−0.13
Freedom H	0.77***	−0.10	0.64***
	0.57***	0.19
Gender Development Index	0.47***	−0.01	0.55***
	0.1	0.24
Gender Inequality Index	−0.78***	0.44**	−0.91***
	−0.23	0.47**
Competitiveness	0.60***	−0.62***	0.79***
	0.03	−0.57***
Corruption	−0.78***	0.43**	−0.83***
	−0.43**	0.23
Happiness	0.83***	−0.1	0.78***
	0.60***	0.32*
Life satisfaction	0.77***	−0.05	0.68***
	0.54**	0.42*
Pathogen stress	−0.75***	0.33*	−0.88***
	−0.22	0.09
External Territorial Threats	−0.60**	−0.35	−0.58**
	−0.26	−0.47*
Territory deaths (log)	−0.44*	0.49**	−0.70***
	0.09	0.28
Conflict score (log)	−0.71***	0.10	−0.63***
	−0.46**	−0.16
Action resources	0.74***	−0.13	0.76***
	0.37*	0.23

Cultural variables with less than 25 countries were included in the correlation tests, but not in the factor analysis. Of the variables that were not included in the factor analysis, the variable named integration was strongly correlated with the first factor F1, and Confucian work dynamism was strongly correlated with the second factor F2 (Table 3).

The human development index (HDI) was used as a control variable in order to control for the possible confounding influence of general development. The correlations of the cultural variables and background variables against the factors, with and without control for HDI, are listed in Table 3. A matrix of correlations of all combinations of two variables with and without control for HDI is provided in the Online Appendix.

The position of countries along the two factors were estimated as follows. The data set used in the factor analysis was further reduced by removing variables with more than 25% missing country values, resulting in a matrix of 47 countries by 40 variables. 5.7% missing values were imputed by the expectation maximization algorithm (Schafer, 1997), and the data for each country were multiplied by the factor loadings. Approximate positions were estimated for an additional 32 countries which had known values for at least 30% of the 40 variables, again using expectation maximization. A map of the positions of the 47 + 32 = 79 countries along factors F1 and F2 is shown in Figure 3. The less precise points based on the second estimation are marked with an open circle. Country positions and rankings on factors F3, F4, and F5 are listed in the Online Appendix. The loadings of the cultural variables on factors F1 and F2 are illustrated.OPEN IN VIEWER

Study 2

In a second study, the data set of cultural variables was split into two subsets of organizational culture and national culture, respectively. The factor analysis was repeated with the two datasets separately. The factor loadings are listed in the Online Appendix for solutions with up to five factors. The factors of a two-factor solution accounted for 22% and 17% of the variance in the organizational culture subsample, and 39% and 15% of the variance in the national culture subsample. A country map similar to Figure 3 was constructed for each of the two subsets. Figure 4 shows a map of countries along the two factors resulting from a factor analysis of organizational culture. Figure 5 shows a similar map from a factor analysis of the dataset of national culture.OPEN IN VIEWER

OPEN IN VIEWER

Figure 5.

Map similar to Figure 3 based on factor analysis of national culture only.

Explaining F1

The factor F1 in study 1 is significantly correlated with 64 out of 91 cultural variables (p < .05, Table 3) and accounts for 34% of the total variance. This factor is strongly correlated with HDI (r = 0.74, p << .0001), but the correlations of cultural variables with F1 cannot be explained by development effects alone because 54 out of these cultural variables are significantly correlated with this factor when controlling for HDI. The F1 factor is significantly correlated with all the background variables. Some of the correlations with background variables disappear when controlling for HDI, but the correlations of F1 with the background variables democracy, freedom, corruption, happiness, and conflict score remain fairly strong and significant after controlling for HDI (Table 3).

Comparing the results with Inglehart and Welzel’s (2005) modernization theory, we can observe that secular values, self-expression values, and emancipative values are all strongly correlated with F1 (Table 3). F1 also aligns with Li and Bond’s (2010) single-factor solution named secularism. This is in agreement with Inglehart’s (2018) later observation that a single factor can account for a large fraction of the variation in major cultural variables. All these variables are also strongly correlated with HDI (see Online Appendix). We can conclude that F1 captures general development effects, in agreement with modernization theory (Inglehart & Welzel, 2005; Inglehart, 2018), emancipation theory (Welzel, 2013), and secularization theory (Li & Bond, 2010).

Gelfand’s measure of cultural tightness (Gelfand et al., 2011; Eriksson et al., 2021) is significantly correlated with F1 (r = −0.42, p = .014). Uz (2015) has made a different measure of cultural tightness that shows a much stronger correlation with F1 (r = −0.81, p << .0001). The tightness measure of Uz is based on the homogeneity of norms and attitudes. This measure is more objective than Gelfand’s, but possibly confounded with response style.

Pathogen stress has a strong negative correlation with F1, but this correlation becomes insignificant when controlling for HDI (Table 3). The possibility of spurious correlations of pathogen stress has been pointed out by several authors (Currie & Mace, 2012; Hruschka and Henrich, 2013; Horita & Takezawa, 2018).

Echeverría et al. (2019) have theorized that violent conflict is reducing the action resources of the population which leads to a reduction in emancipative values. Echeverría’s measure of action resources is strongly correlated with F1 (Table 3).

A theoretical explanation for the connection of F1 with violent conflict is provided by regality theory (Fog, 2017). Regality theory focuses on the psychological effects of fear and collective danger. The level of danger and insecurity is high in the poor and war-torn countries that we find at the low end of the F1 scale, while existential security is taken for granted in the highly developed North European welfare states at the high end of F1 (see Figure 3). Regality has a strong negative correlation with F1, even after controlling for HDI (Table 3). Regality theory thus provides a psychological explanation for the correlation of many ‘soft’ variables with F1. A high value of F1 is connected with emancipative values, individualism, egalitarian values, self-expression, hedonism, gender equality, etc.; while a low value of F1 is connected with conservatism, morality, religiosity, hierarchy, power distance, tightness, and vertical source of guidance (Table 3).

By combining all the theories mentioned here, we can conclude that there are many causal links between different domains of culture. It is useful here to distinguish between three main domains of culture. The first domain is material culture represented by development in economy, technology, and health. The second domain comprises social structure including modern social institutions, democracy, peace, and international cooperation. And finally, we can consider a third domain of psychological expressions and symbolic culture including opinions, values, attitudes, ideology, beliefs, religiosity, and sexuality.

The reason why a single factor F1 is correlated with so many cultural variables as well as material and technical background variables is that there are many causal links and overlaps between all three domains of culture. Links between the first two domains are well-known and described by the many different branches of development theory (Midgley, 2013). Links from the first two domains to the third domain are described by the theories reviewed above. While different scientists each have their favorite theories, it is safe to conclude that there are many causal links between all three domains. Material and economic resources in the first domain interact with social structure and institutions in the second domain in many different ways that may be mutually enforcing and provide synergistic effects that improve collective security. A feeling of security has many psychological and cultural effects in the third domain according to the theories mentioned here. Perceived collective security is typically reflected in cultural values such as tolerance, egalitarianism, etc. These values are in turn influencing the second domain with shared aspirations for justice and democracy. The number of possible interactions between variables in the three domains is almost endless. Many additional theories can be invoked to analyze links between specific variables in the different domains. Variables in all three domains are influencing each other to such a degree that they become highly correlated with each other and tend to move in parallel and collapse into a single factor. The approximately parallel movement in all three domains of culture is what we generally perceive as development.

The observation that things tend to move in parallel should not make us fall back to the old idea that cultural evolution always goes in a specific direction. The unilinear theory of cultural evolution has been rejected long ago (Steward, 1955). There is no certainty that the different aspects of culture will continue to follow parallel trajectories in the future. In fact, we have seen a recent backlash in democracy and freedom worldwide at the same time as the economy has made significant progress (Norris & Inglehart, 2019; The Economist Intelligence Unit, 2019; Abramowitz, 2018; Freedom House, 2019). Explanations for the current backlash compromising freedom and democracy may be found in an increased media focus on terrorism since 2001, growing economic inequality, recurrent economic crises, current wars, refugee crises, as well as fears of ecological collapse (Fog, 2017). The different cultural variables are not moving in perfect synchrony, but the correlations between them have been strong enough for a sufficiently long period of time to make them load on the same factor F1.

The rotation angle of the F1 axis has been adjusted to capture as much as possible of the variation in the relevant variables, but the sign of F1 is still arbitrary. We may place the negative end of the F1 axis in the African and Middle Eastern countries and the positive end in the North European countries if the theoretical framework is focused on development. But we may just as well draw it in the opposite direction if the theoretical focus is on the psychological effects of fear, danger, and insecurity.

Explaining F2

A theoretical explanation for the second factor is less obvious. The country map in Figure 3 shows that the East Asian countries are found at the low end of the F2 axis, while mostly South American countries are found at the high end. The cultural variables that are associated with a low value of F2 include (Table 3): flexibility vs. monumentalism, ego boosting vs. altruism, long-term orientation, K factor and hypometropia, secular values, Confucian work dynamism, and discipline vs. violence. A high value of F2 is associated with monumentalism vs. flexumility, civility vs. practicality, and relational mobility. Most of these variables originated as results of factor analyses. The sometimes bizarre and confusing names reflect the attempts of the respective authors to interpret and make sense of the empirical factors they found. We need a more theoretical approach to better understand the background for the cultural differences reflected by these variables and why they are correlated to such a degree that they form a common factor F2.

Several of these variables are designed specifically to tap the characteristics of East Asian culture. The variables named Confucian work dynamism, long-term orientation, monumentalism vs. flexumility, flexibility vs. monumentalism, and K factor and hypometropia are all different revisions in a series of attempts to measure typical aspects of East Asian culture, as explained above. The many related variables in the sample are partially responsible for the formation of the factor F2, but this factor also appears in studies that are not aimed at detecting East Asian values.

The theory of relational mobility is useful for explaining the cultural characteristics of East Asian countries. Irrigated rice farming makes people dependent on the local community and local leaders, which implies a low relational mobility. This explains the long-term orientation in these countries as well as the perseverance, thrift, and humility that are important aspects of flexibility vs. monumentalism, etc. (Thomson et al., 2018). A low relational mobility makes it important to avoid local conflict and maintain social harmony. The low value of social polarization can be explained by a suppression of divergent opinions (Minkov, 2009). Strong emotional expressions must also be suppressed in cultures with low relational mobility in order to maintain social harmony (Krys et al., 2022). This explains the low value of emotional display (Matsumoto et al., 2008). At the opposite end of the F2 dimension, a high relational mobility is associated with short-term orientation and a risk-accepting strategy in mating competition. This explains why the K factor is at the low end of the F2 dimension (Minkov, 2014), while sociosexuality is at the high end (see Figure 6).OPEN IN VIEWER

Relational mobility is high in Latin American countries due to a history of herding, but the theory of relational mobility cannot fully explain why Latin America countries range higher on F2 than European and English-speaking countries (see Figure 3). In fact, we would expect relational mobility to be very high in industrial and post-industrial societies. Differences in colonial history and ethnic diversity are also part of the explanation. Settler colonialism and ethnic diversity in Latin America has fostered a more independent and unique construal of selfhood (Krys et al., 2022). An ethnically diverse society requires tolerance and respect for other people. These qualities are included in the variable civility vs. practicality (Bond & Lun, 2014), which tops the F2 scale (Figure 6). Different models of selfhood are also part of the flexibility vs. monumentalism scale (Krys et al., 2022), which aligns closely with the F2 dimension. Differences in the variable named indulgence vs. restraint (Hofstede et al., 2010) or industry vs. indulgence (Minkov, 2011) can be explained by the high self-reliance in Latin America versus interdependence in East Asia (Krys et al., 2022).

The cultural logics of honor vs. protecting one’s face are among the cultural values that differentiate East Asian from Latin American cultures. This distinction is part of the long-term orientation and the flexibility vs. monumentalism scales (Krys et al., 2022). Cultures of honor are found in societies with weak law enforcement, while the face logic is associated with efficient hierarchical control. The factor named discipline vs. violence is interpreted as organized law enforcement versus violent retaliation against wrongdoing (Fog, 2017: 266). We can expect this factor to be correlated with the logics of face vs. honor.

Long-term orientation is associated with longer education and economic investment (Hofstede et al., 2010; Minkov, 2014). We can expect the education level and the level of economic investment to influence many aspects of culture.

Religion also plays an important role. The low end of the F2 axis is dominated by Buddhist countries where beliefs in fate are strong. Beliefs in fate are also quite strong in Islamic countries which tend towards the low end of the F1 axis. The variable fate control is accordingly found in the lower left quadrant of Figure 6. The theoretically related variable individual locus of control (LOC) has the opposite sign and is found in the upper right quadrant. Religious beliefs may develop and adapt to the environment, but this process is so slow that religion may be considered an independent influence in an analysis on a short time scale.

African countries south of Sahara are under-represented in the statistical studies to date, but we may expect a relatively high relational mobility in some African cultures due to a history of nomadic herding. Relational mobility has not been measured yet in sub-Saharan Africa, except for Mauritius which has a relatively high relational mobility (Thomson et al., 2018). The colonial history of African countries is mixed. Settler colonialism has taken place where the climate was most favorable, while occupation and exploitation were more common where the incidence of infectious diseases was high (Hruschka & Henrich, 2013). Ethnic diversity is high in many parts of Africa. African countries are mainly found at the positive end of the F2 dimension (figure 3).

The F1-F2 Factor Structure

We may ask if there is a right or optimal rotation angle for the factors. Kaasa and Minkov (2022) say that the rotation is arbitrary, but these authors are in fact relying on Minkov’s revised model (Minkov & Kaasa, 2021), which is a two-dimensional model that aligns almost perfectly with the F1 and F2 dimensions of the present study. The rotation is not arbitrary if we are seeking theoretical explanations for the relationship between different cultural variables. There appears to be one dominating direction in the coordinate system of the two factors. The different cultures tend to cluster around the diagonal on Inglehart and Welzel’s cultural map (Figure 1). A study of birth cohorts shows that movements in time tend to follow approximately parallel trajectories along this diagonal (Inglehart & Welzel, 2005: 112). This observation agrees with those of Li and Bond (2010) and Inglehart (2018) that the two dimensions on Inglehart and Welzel’s cultural map can be combined into one dominating factor that explains a large part of the cultural variance. Combining the two factors named traditional vs. secular-rational values and survival vs. self-expression values into one dominating factor will produce a diagonal line on Inglehart and Welzel’s map representing a large part of the total variance of cultural values, corresponding to F1 of the present study. Li and Bond (2010) call this factor secularism, and Inglehart (2018) makes it the basis of his modernization theory.

The factor F1 reflects mainly effects related to development, economy, and the psychological effects of danger versus security. F2 is related to mode of subsistence, colonial history, relational mobility, long-term vs. short-term orientation, and self-construal. Other rotations of the factor map are possible, but the solution chosen here seeks to prioritize the strongest factors first. Many cultural variables tend to cluster around F1, and the remaining variables form a somewhat weaker cluster around F2. Several attempts to clarify and simplify cultural differences have converged. Li and Bond (2010) and Inglehart (2018) have found that a single factor similar to F1 accounts for a large part of cultural variation, while Minkov has revised earlier studies and constructed a two-factor solution similar to the one presented here (Minkov, 2018; Minkov & Kaasa, 2021). This solution appears to be easier to interpret theoretically than differently rotated solutions published elsewhere. Inglehart and Welzel’s secular-rational values and self-expression values (Inglehart & Welzel, 2005; Welzel, 2013) shown in Figure 1 are the results of a differently rotated solution. These variables are correlated almost equally with both F1 and F2 due to the rotation as we can see in Figure 6. Several other variables are the results of factor analysis with rotation, combining elements that correlate with F1 as well as elements that correlate with F2. For example, the variable named self-directedness vs. other-directedness combines a variety of traits such as determination, perseverance, responsibility, independence, imagination, religious faith, and obedience (Bond & Lun, 2014). It should be no surprise that such a composite variable is loading on multiple factors. Many cultural variables are correlated with both F1 and F2, as we can see on Figure 6. When we consider the high complexity and multicausality of cultural phenomena, it is not surprising that cultural variables are influenced by multiple factors. Neither of the factors F1 and F2 are pure indicators of a single causal influence, but aggregates of many different effects that are somehow related and strongly correlated with each other.

While relational mobility is important for explaining F2, it is also correlated with F1 as Figure 6 shows. This is because a high relational mobility is associated with individualism while a low relational mobility is associated with collectivism. The concept of relational mobility also applies to sexual partnerships and marriage. This explains why sociosexuality and gender equality lie in the same quadrant as relational mobility on Figure 6. Mathematicians may prefer explanatory variables to be perpendicular to each other, but this is not possible when explaining cultural differences because almost everything is correlated with the F1 factor that covers elements from all domains of culture. Relational mobility is not perpendicular to F1, yet it provides an important contribution to the theoretical explanation of F2.

Most of the variables that are designed to tap the characteristics of East Asian culture are lying close to the F2 axis, but the variable named monumentalism vs. flexumility is deviating somewhat from the other variables. The original definition of monumentalism involves pride, immutable norms, and strong religiousness (Minkov, 2011) which are traits that we expect to find at the low end of the F1 scale. It is therefore logical that this variable has a negative correlation with F1. A later revision renamed this variable to flexibility vs. monumentalism and changed the definition to be based it on self-enhancement, self-consistency, and willingness to help people (Minkov et al., 2018). The latter revision is closer to the F2 axis and has the sign reversed as we can see in Figure 6. We may prefer the older name long-term vs. short-term orientation as it is more intelligible. The variable with this name aligns closely with the F2 axis.

Limitations

We must bear in mind that many caveats apply to cross-cultural studies. The cross-cultural validity of survey instruments is a general problem because survey questions may be interpreted differently in different cultures and different languages (Aleman & Woods, 2016). Many of the studies included here are relying on non-representative convenience samples, and some studies have small sample sizes. Most of the studies reporting national cultures ignore cultural variations within each country (Taras et al., 2016). Only few cross-cultural studies are compensating for differences in response style or other kinds of bias (Boer et al., 2018). We can expect a more acquiescent response style in cultures with low relational mobility. Therefore, the tendency of East Asians to give less extreme and less negative answers to surveys (Harzing, 2006; Guo & Spina, 2019) may have contributed to the F2 factor.

Several of the included studies rely on surveys that ask informants to evaluate their own culture without any clear frame of reference. People tend to use their own culture as frame of reference unless instructed otherwise. The reference group effect is a likely source of error in many cross-cultural studies. For example, cultural tightness has been measured by asking survey respondents to evaluate their own culture without referring to any frame of reference (Gelfand et al., 2011; Eriksson et al., 2021). We may expect a neutral or meaningless answer when people try to evaluate their own culture against the very same culture as frame of reference. While the reference group effect has been studied mainly in connection with self-evaluation (Heine et al., 2002), we can expect it to apply also to evaluation of one’s own culture. Thus, Minkov (2011) observes that it is unreliable to use people as informants about their own culture. The reference group effect may explain, for example, why Gelfand et al. (2011) find that Norway has a tighter culture than China, while we would expect it to be opposite according to Figure 3. An attempt to replicate Gelfand’s study failed to reproduce the country rankings and found that the correlations with related cultural variables were lower than expected. It was concluded that Gelfand and coworkers’ tightness measure had poor internal consistency and validity (Treviño et al., 2021). Another attempt to replicate Gelfand’s study was more successful (Gelfand, 2021).

Galton’s problem is obvious when looking at the country map in Figure 3. The effects of geographic proximity and cultural diffusion cannot easily be disentangled from the effects of environmental influences. The correlations indicated in the present study, as well as in the cited studies, should not be interpreted as simple causal relationships because any two cultural variables may be subject to many common causal influences.

The present study has a problem with missing country data. The results could not have been obtained without the estimation of missing values. The different cross-cultural studies have been carried out at different times. Different cultures often develop in the same direction. If there is a certain statistical relationship between two different cultures at a certain point in time, then there is likely to be a similar relationship at a later time if the two cultures are following approximately parallel trajectories (Beugelsdijk & Welzel, 2018). Confidence intervals are not provided here because the accuracy of the estimation methods is difficult to gauge and because the accuracy of the input variables varies. The numerical results provided by the present study should be regarded as approximate because of problems with data quality in the cited studies as well as the methodological problems of the present study. We should rely mainly on the qualitative conclusions and general patterns identified here, rather than on exact numerical results.

The overview of cultural differences presented here may seem comprehensive, but its value is limited by the fact that it includes studies of very different quality, and by the fact that certain cultural phenomena that have attracted the attention of many social scientists and led to many quantitative publications have received more weight than less popular areas of study.