Sample equivalence speaks to the base likelihood of specific behaviors across cultural contexts: for instance, the likelihood that people use landlines, mobile phones, or Internet varies across countries as do color preferences for products (Madden et al., 2000) or the rates of impulsive buying behavior (Kacen and Lee, 2002). Variations in these behaviors may further exist across subgroups with respect to age, gender, educational level, and so forth. Comparing communication in individual, in-group, out-of-group, and nonsocial language situations, for instance, Kashima et al. (2011) found that participants of Asian origin and women used more self-descriptions in interpersonal contexts whereas Australians did so more frequently when confronted with collective contexts. Such differences affect the base rate of behaviors to be considered average or normal in a group.

Data collection equivalence describes the accessibility of data and the willingness of people to provide data across cultural contexts (e.g., Lowry et al., 2011). For instance, Chinese users of social networks are more likely to customize their profile picture than are Americans (Zhao and Jiang, 2011), thus potentially providing more pointers to their personality. As another example, closed circuit television cameras are widely accepted in Britain and The Netherlands and are even high on public wish lists, whereas other countries such as Germany are facing serious public and media resistance against such instruments (e.g., Bayerl et al., 2013). These differences in societies’ willingness to provide information online and to accept surveillance by private and public institutions again affect the base rate of behaviors across groups, but also take into consideration differential impacts on base rates owing to different reactions to external pressure such as physical security measures or online surveillance.

Measure equivalence refers to the comparability of measurement units across contexts. For instance, the number of words people use to express an opinion; the emotional intensity of expressions; the colors used as symbols for abstract concepts; and the number of exclamation marks, question marks, or emoticons are specific per culture and mother tongue. Koreans, for instance, seem more likely to speak indirectly and to look for indirect meanings than are Americans (Holtgraves, 1997). Also, the general tendency to extreme responses, humility, or social desirability varies across contexts, influencing tendencies to agree or disagree to questions or the frequency of extreme versus moderate responses in surveys (Smith, 2011). Even the structure of academic texts (Clyne, 1987) and the type of personal pronouns (“I,” “me,” and “mine,” vs “we,” “ours,” and “us”) depend on culture, with more individualistic cultures preferring the former and more collectivistic cultures the latter (Na and Choi, 2009). Measure equivalence thus affects the base rate, but even more important, the type and intensity of behaviors or content that can be expected across cultural groups to express the same concept.

Conceptual equivalence addresses the fact that things do not mean the same everywhere. The color white, for instance, signals mourning in Japan, and purity in most Western contexts. In some cultural contexts, intelligence is indicated by slow speech, and by fast speech in others, whereas the word “family” can refer to only the core members (i.e., father, mother, children) or include the extended family of grandparents, aunts, uncles, and nieces. Similarly, not being aware that in some locations “friendship” can be used to describe a spontaneous, short-term relationship that can be terminated quickly, whereas in others it is used only for lifelong deep loyalties implying far-reaching social and financial obligations, may lead to severe misunderstandings about the disparate implications and commitments entered into by people using this term. Conceptual equivalence thus addresses the problem of giving the right meaning to behavioral traces across contexts, or rather, the challenge of accounting for the fact that disparate base rates, behaviors, or patterns essentially mean the same thing or that the identical content or pattern carries an different meaning.

In a comparable way, artifacts can perform different functions. Imagine, for instance, a family visiting a Western fast-food restaurant. In a United States (US) context this might indicate low–social status behavior, whereas in an African context this might signal a high-status family embracing symbols of freedom and the affluent lifestyle of the West. Yet, precisely this symbol of the West can change in its meaning depending on political and societal developments. Whereas in an Eastern European context Western products such as soft drinks and fast food were long considered scarce and attractive status goods, the Western signaling function now turned in some market segments into a liability: The mere fact that a product comes from the West now can be considered as pushy (Van Rekom et al., 2006). Functional equivalence thus refers to the challenge that reactions to the same triggers are influenced by the cultural contexts in which they are encountered.

The international nature of Big Data also means that it may contain multiple languages, including local dialects and group-specific lingos. Most of us are know that adequate translations of natural text or speech are far from trivial. This starts with the fact that some experiences and objects simply do not have words in all languages (e.g., Japanese knows eight common honorifics to express hierarchical differences among individuals whereas English speakers are reduced to the choice between a person’s first or last name). Some words have double meanings in their original language that are lost in translation, or vice versa, whereas proverbs, metaphors, the level of abstraction, and precision are highly language and context specific (e.g., the same person will most certainly use different vocabulary with family, at work, and at a club with friends; few people would understand a literal translation of the Greek “having a toothache” to mean “being unhappily in love”). In addition, language usage changes quickly over time (e.g., emojis are now replacing smileys, whereas “Googling” has long become an established term for searching on the Internet). By necessity, working with a language that is not the mother-language is thus fraught with misunderstandings. Translation equivalence thus draws attention to the ability to ensure that translation of texts and speech as well as pictures or drawings are actually transporting the original meaning from the original language and context.