Facebook Likes and Human Behavior

Earlier in 2013, the Psychometrics Centre of the University of Cambridge conducted an impressive research about how private attributes are predictable from digital records of human behavior, like Facebook likes. Digitally mediated behaviors like Facebook likes can easily be recorded and analyzed, fueling the emergence of computational social science and new services such as personalized search engines and targeted online marketing. However, the widespread availability of extensive records of individual behavior, together with the desire to learn more about customers and citizens, presents serious challenges related to privacy and data ownership. 

The researchers (Kosinski, Stinwell & Graepel) showed that easily accessible digital records of behavior, Facebook Likes, can be used to automatically and accurately predict a range of highly sensitive personal attributes including sexual orientation,ethnicity, religious and political views, personality traits, intelligence, happiness, use of addictive substances, parental separation, age, and gender. The analysis presented is based on a dataset of over 58,000 volunteers who provided their Facebook Likes, detailed demographic profiles, and the results of several psychometric tests. The proposed model uses dimensionality reduction for preprocessing the Likes data, which are then entered into linear regression to predict individual psychodemographic profiles from Likes. The model correctly discriminates between homosexual and heterosexual men in 88% of cases, African Americans and Caucasian Americans in 95% of cases, and between Democrat and Republican in 85% of cases.

Predicting individual traits and attributes based on various cues, such as samples of written text, answers to a psychometric test, or the appearance of spaces people inhabit, has a long history. Human migration to digital environment renders it possible to base such predictions on digital records of human behavior. It has been shown that age, gender, occupation, education level and even personality can be predicted from people’s Web site browsing logs. Similarly, it has been shown that personality can be predicted based on the contents of personal web sites, music collections, properties of Facebook or Twitter profiles such as the number of friends or the density of friendship networks, or language used by their users. Furthermore, location within a friendship network at Facebook was shown to be predictive of sexual orientation.

The research 

The Psychometrics Centre, University of Cambridge

The study was based on a sample of 58,466 volunteers from the United States, obtained through the my Personality Facebook application, which included their Facebook profile information, a list of their Likes (n = 170 Likes per person on average), psychometric test scores, and survey information. Users and their Likes were represented as a sparse user–Like matrix, the entries of which were set to 1 if there existed an association between a user and a Like and 0 otherwise. The dimensionility of the user–Like matrix was reduced using singular-value decomposition (SVD). 
Numeric variables such as age or intelligence were predicted using a linear regression model, whereas dichotomous variables such as gender or sexual orientation were predicted using logistic regression. In both cases,the researchers applied 10-fold cross-validation and used the k =100 top SVD components. For sexual orientation, parents’ relationship status, and drug consumption only k = 30 top SVD components were used because of the smaller number of users for which this information was available.

Prediction of Dichotomous Variables

The Psychometrics Centre, University of Cambridge

The aforementioned figure shows the prediction accuracy of dichotomous variables expressed in terms of the area under the receiver-operating characteristic curve (AUC), which is equivalent to the probability of correctly classifying two randomly selected users one from each class (e.g., male and female). The highest accuracy was achieved for ethnic origin and gender. African Americans and Caucasian Americans were correctly classified in 95% of cases, and males and females were correctly classified in 93% of cases, suggesting that patterns of online behavior as expressed by Likes significantly differ between those groups allowing for nearly perfect classification. Christians and Muslims were correctly classified in 82% of cases, and similar results were achieved for Democrats and Republicans (85%). Sexual orientation was easier to distinguish among males (88%) than females (75%), which may suggest a wider behavioral divide (as observed from online behavior) between hetero and homosexual males.
Good prediction accuracy was achieved for relationship status and substance use (between 65% and 73%). The relatively lower accuracy for relationship status may be explained by its temporal variability compared with other dichotomous variables (e.g., gender or sexual orientation).

Predictive Power of Likes

Individual traits and attributes can be predicted to a high degree of accuracy based on records of users’ Likes. The best predictors of high intelligence include “Thunderstorms,” “The Colbert Report,” “Science,” and “Curly Fries,” whereas low intelligence was indicated by “Sephora,” “I Love Being A Mom,” “Harley Davidson,” and “Lady Antebellum.” Good predictors of male homosexuality included “No H8 Campaign,” “Mac Cosmetics,” and “Wicked The Musical,” whereas strong predictors of male heterosexuality included “Wu-Tang Clan,” “Shaq,” and “Being Confused After Waking Up From Naps.” 

Accuracy of selected predictions as a function of the number of available Likes. Accuracy is expressed as AUC (gender) and Pearson’s correlation coefficient (age and openness). About 50% of users in this sample had at least 100 Likes and about 20% had at least 250 Likes. Note, that for gender (dichotomous variable) the random guessing baseline corresponds to an AUC = 0.50. The Psychometrics Centre, University of Cambridge.

Moreover, note that few users were associated with Likes explicitly revealing their attributes. For example, less than 5% of users labeled as gay were connected with explicitly gay groups, such as No H8 Campaign, “Being Gay,” “Gay Marriage,” “I love Being Gay,” “We Didn’t Choose To Be Gay We Were Chosen.” Consequently, predictions rely on less informative but more popular Likes, such as “Britney Spears” or “Desperate Housewives” (both moderately indicative of being gay).

Conclusions

Similarity between Facebook Likes and other widespread kinds of digital records, such as browsing histories, search queries, or purchase histories suggests that the potential to reveal users’ attributes is unlikely to be limited to Likes. Moreover, the wide variety of attributes predicted in this study indicates that, given appropriate training data, it may be possible to reveal other attributes as well.
Predicting users’ individual attributes and preferences can be used to improve numerous products and services. For instance, digital systems and devices (such as online stores or cars) could be designed to adjust their behavior to best fit each user’s inferred profile. Also, the relevance of marketing and product recommendations could be improved by adding psychological dimensions to current user models. For example, online insurance advertisements might emphasize security when facing emotionally unstable (neurotic) users but stress potential threats when dealing with emotionally stable ones. 

Moreover, digital records of behavior may provide a convenient and reliable way to measure psychological traits. Automated assessment based on large samples of behavior may not only be more accurate and less prone to cheating and misrepresentation but may also permit assessment across time to detect trends. Moreover, inference based on observations of digitally recorded behavior may open new doors for research in human psychology.

On the other hand, the predictability of individual attributes from digital records of behavior may have considerable negative implications, because it can easily be applied to large numbers of people without obtaining their individual consent and without them noticing. Commercial companies, governmental institutions, or even one’s Facebook friends could use software to infer attributes such as intelligence, sexual orientation, or political views that an individual may not have intended to share. One can imagine situations in which such predictions, even if incorrect, could pose a threat to an individual’s well-being, freedom, or even life. Importantly, given the ever-increasing amount of digital traces people leave behind, it becomes difficult for individuals to control which of their attributes are being revealed.

There is a risk that the growing awareness of digital exposure may negatively affect people’s experience of digital technologies, decrease their trust in online services, or even completely deter them from using digital technology. It is our hope, however, that the trust and goodwill among parties interacting in the digital environment can be maintained by providing users with transparency and control over their information, leading to an individually controlled balance between the promises and perils of the Digital Age.

Read more »

Cross-Cultural Consumer Behavior

What are the key cultural constructs or dimension?

The constructs of individualism and collectivism represent the 
most broadly used dimensions of cultural variability for cross-cultural comparison. In individualistic (IND) cultures, people tend to prefer independent relationships to others and to subordinate the goals of their ingroups to their own personal goals. In collectivistic (COL) cultures, in contrast, individuals tend to prefer interdependent relationships to others and to subordinate their personal goals to those of their ingroups. The key distinction involves the extent to which one defines the self in relation to others. The focus is on whether the self is defined as autonomous and unique or seen as inextricably and fundamentally embedded within a larger social network. This distinction has also been referred to as egocentric versus sociocentric selves (Shweder & Bourne,1982), or independence vs interdependence.

In sum, the distinctions between IND and COL societies, and independent and interdependent self-construals, are crucial to the cross-cultural understanding of consumer behavior. Indeed, whereas the 1980s were labeled the decade of individualism/collectivism in cross-cultural psychology, similar distinctions represent the dominant structural approach in cross-cultural consumer research in 1990s and 2000s. The studies to be reviewed in this chapter offer a wealth of evidence that these cultural classifications have fundamental implications for consumption-related outcomes.

Expanding the Set of Cultural Dimensions

The nature and meaning of IND and COL (or of independent and interdependent self-construals) appear to vary across cultural, institutional, gender, and ethnic lines. Although the breadth of the INDCOL constructs lends integrative strengths, further refinement of these categories holds the potential to enhance prediction of consumer behavior.

The Horizontal/Vertical Distinction: Which additional cultural categories offer value in the prediction of cross-cultural consumer behavior? Within the INDCOL framework, Triandis and his colleagues have recently introduced a further distinction between societies that are horizontal (valuing equality) and those that are vertical (emphasizing hierarchy). The horizontal/vertical distinction emerges from the observation that American or British individualism differs from, say, Norwegian or Danish individualism in much the same way that Japanese or Korean collectivism differs from the collectivism of the Israeli kibbutz. Specifically, in vertical individualist societies (U.S.,Great Britain, France), people tend to be concerned with improving their individual status and distinguishing themselves from others via competition. In contrast, in horizontal individualist societies (HI; e.g., Sweden, Norway, Australia), where people prefer to view themselves as equal to others in status, the focus is on expressing one’s uniqueness, capability, and self-reliance. In vertical collectivist societies (Japan, Korea, India), people focus on fulfilling obligations to others, and on enhancing the status of their ingroups in competition with outgroups, even when that entails sacrificing their own personal goals. In horizontal collectivist societies (like exemplified historically by the Israeli kibbutz), the focus is on sociability, benevolence, and interdependence with others in an egalitarian context.

When such distinctions are taken into account, however, it becomes apparent that the societies chosen to represent IND and COL cultural syndromes in consumer research have almost exclusively been vertically oriented. Specifically, the modal comparisons are between the United States (VI) and any of a number of Pacific Rim countries (VC). It may be argued, therefore, that much of what is known about consumer behavior in individualistic and collectivistic societies reflects vertical forms of these syndromes and may not generalize, for example, to comparisons between Sweden (HI) and Israel (HC) or other sets of horizontal cultures. As an example, conformity in product choice, as examined by Kim and Markus (1999), may be a tendency specific to VC cultures, in which deference to authority and to ingroup wishes is stressed. Much lower levels of conformity may be observed in HC cultures, which emphasize sociability but not deference. Thus, it may be inappropriate to ascribe differences in consumers’ conformity between Korea (VC) and the United States (VI) solely to the role of IND/COL or independence/interdependence, because such conformity might not be prevalent in horizontal societies. In particular, levels of product conformity in HC contexts might not exceed those in HI contexts.

Predicting new consumer psychology phenomena

Several recent studies examining the implications of this horizontal/vertical cultural distinction have provided evidence for its value as a predictor of new consumer psychology phenomena and as a basis for refining the understanding of known phenomena. For instance, Gürhan-Canli and Maheswaran (2000) demonstrated that the tendency to favor products from one’s own country over foreign products (a country-of-origin effect) emerged more strongly in Japan (a VC culture) than in the United States (a VI culture). This fits well with a conceptualization of collectivists as being oriented toward their ingroups. However, mediational analyses using individual consumers’ self-rated cultural values indicated that only the vertical dimension of IND and COL accounted for the country-of-origin effects in Japan. In other words, the collectivistic tendency to favor one’s own country’s products appeared to be driven by cultural values that stress hierarchy, competition, and deference to ingroup wishes, not by values that stress interdependence more generally.

In line with this, research suggests that advertising messages with themes that emphasize status, prestige, hierarchy, and distinction may be more prevalent and persuasive in vertical cultural contexts. Such advertisements also appear to be generally more persuasive for those with a vertical cultural orientation, and may be inappropriate for those with a horizontal one. Shavitt, Zhang, and Johnson (2006) asked U.S. respondents to write advertisements that they personally would find persuasive. The extent to which the ad appeals that they wrote emphasized status themes was positively correlated with respondents’ vertical cultural orientation (and negatively correlated with their horizontal cultural orientation). Moreover, content analyses of magazine advertisements in several countries suggested that status-oriented themes of hierarchy, luxury, prominence, and distinction were generally more prevalent in societies presumed to have vertical cultural profiles (Korea, Russia) than a horizontal cultural profile (Denmark).

Last but not least,the horizontal/vertical distinction provides a basis for refining our understanding of individualism/collectivism effects. Their studies showed that individualism/collectivism differences in socially desirable responding appear to be mediated at the individual level by horizontal (but not vertical) IND and COL values. These findings shed light on the motivational drivers linking culture with socially desirable response styles. The main point here is that these dimensions of cultural comparison have multiple implications for advertising and marketing processes. Attention to a broader set of cultural dimensions will not only expand the range of independent variables in our research, but will also prompt consideration of cultural consequences hitherto unexamined in cross-cultural studies (based on publication by Sharon Shavitt-University of Illinois & Angela Lee, Northwestern University).

Read more »

Behavioral marketing

Emotional cues that work magic for customers

 Marketers have long understood that emotions play an important role in consumer decision making. But, as the latest scientific evidence suggests, their influence is much more nuanced and complex than many are aware. Subtle, rather than intense, emotional reactions are often more persuasive. Short-lived emotions can have lasting effects. The experience and expression of negative emotions can sometimes be beneficial. Emotional experiences are often poorly predicted and remembered. In all these areas, a better understanding of emotions will help managers tailor their own act to give better prompts and get the desired response from consumers, in order to maximize customer satisfaction and loyalty at every stage of the encounter.

Customers tend to be quite aware of how intense emotions can affect their decisions: The overcrowding and long lines at the Ikea checkout may leave customers fuming. But as they march out of the store vowing never to shop there again, they may well indulge in an impulse buy on the way out, just to relieve the stress they’re feeling.

What just happened? We often focus on the influence of intense emotions, yet ignore how subtler feelings might also be affecting consumer actions and choices, perhaps on an unconscious level. Mild emotions, whether positive or negative, can trigger or inhibit consumer actions just as powerfully as intensely realized ones.This means that, in order to ensure that your customers feel happy and relaxed when dealing with your business, it is important to pay attention to small cues that can improve their mood or dispel potentially inhibiting negative emotional states.

It is possible to enhance the appeal of your business by exploring areas in which you can evoke positive emotions, which lead to favorable evaluations and increase purchasing intentions. A pleasant scent and music may lure a customer into a store. A smile from a flight attendant may encourage a passenger to buy from the duty-free cart.Simple improvements like this – to alleviate any mild discomfort, and to enhance a positive atmosphere – are easy to implement and can turn a passerby into a lifelong client.

 Creating a pleasant atmosphere with music, scent, lighting or other atmospherics has long been known to provoke positive emotional reactions. But certainly, customer service is always crucial: saying the right words at the right time, remaining calm when faced with an agitated customer or client, going the extra mile.If there are areas of your business that may contribute to a negative emotional state, evaluate how you can alter the emotion for a positive evaluation of the atmosphere. For example, convert frustration to pleasant anticipation during a long wait at a popular restaurant by offering a taste of what’s to come, such as free hors d’oeuvres or beverages.

The Lasting Impact of Short-Lived Emotions

 Many businesses have understood how they can use the mild and mundane emotional experiences of their consumers to influence decisions at that moment. But affecting emotions in the short term can have long-term consequences, as people may form an evaluation or commit to a course of action while experiencing the initial emotion, which will impact their future behavior.Using the classic social science ultimatum game, in one recent paper we explored and confirmed that, in fact, the impact of emotions on behavior can outlive the emotional experience itself.

 In this game, a “proposer” makes an offer to split a given amount of money with a “receiver.” If the person playing the “receiver” feels that the deal is unfair, then that person can reject the offer, and both participants end up with nothing.In our experiment, we began manipulating receivers’ mood by showing them movie clips that sparked either anger or happiness. Subsequently, they were asked to play two ultimatum games.

 In the first game, the proposer would offer the receiver an unfair division of the available amount of money: the proposer would get 75 percent and the receiver 25 percent. Angry receivers rejected these unfair offers at a much higher rate than happy receivers. Even if rejecting the offer meant being left with nothing, the angry participants held to their rationale that unfairness was the basis of their decision.The twist came in the latter round, once the initial emotional reaction had already dissipated. In the next ultimatum game, the once-angry receivers were asked to play the role of the proposer, the majority of whom chose to make fairer offers.

Because people tend to behave consistently with past actions, earlier choices triggered by an incidental emotion became the basis for future decisions.When angry participants made their decision to reject an unfair offer in the first instance, they acquired a self-image of being and acting as fair individuals. Exercising fairness – which resulted from the anger they felt earlier – carried over to a future scenario. Even though the initial anger that triggered the desire for fairness had disappeared, the once-angry proposer is instead guided by behavioral consistency.

Once an emotion has prompted us to choose a course of action, we run into internal and external pressures that continually push us into making similar choices in the future. The image that we have of ourselves – as fair individuals, generous or frugal, socially conscious, conservative or fun-loving – as well as the image that we hope others will have of us, compels us to act consistently in the future.

Another way in which the impact of emotions can outlive the emotional experience itself and influence consumers is through the recall of a previously biased evaluation of a product or service.

Perhaps a biased evaluation

To take a marketing example: A humorous ad can positively bias the evaluation of the product. Based on the positive emotion triggered by that ad, in the future the positive evaluation is recalled, not the current emotional state, and that once-happy moment experienced in the past guides a new decision. Thus, the key is to find ways of making the consumer spontaneously form an evaluation of the product or service while in a good mood, which will then affect subsequent recall and purchase intentions.

Applying the lesson

The phenomena described above lead to two clear suggestions: take advantage of behavioral consistency; and stimulate consumers to form evaluations while in a good mood.Businesses can harness the consistency of self-image in order to appeal to customers. One example of this is in stakeholder marketing. In many countries, retailers have replaced disposable plastic bags with reusable cloth bags. A consumer may initially be put off when they dis-cover they have to pay for a cloth bag. But they inevitably feel good about their contribution to the environment. Behavioral consistency will take over from there as they acquire the self-image of being conscientious about their role in reducing waste and pollution. In this case, the 1 euro they spend commits them to similar decisions in the future, a case in which the emotion triggered maximizes the benefit for all stakeholders (Andrade & Capizzani, Berkeley).

Coca-Cola has been famous for putting viewers of their ads in a good mood. Whether in the ’70s with “I’d like to teach the world to sing” jingle, or during the 2010 FIFA World Cup in South Africa, the adverts trigger a spontaneous, positive emotion. Yet the decision and action to buy Coke may come days or weeks after seeing the ad.

Read more »