Loneliness Consumers and Conformity

Like in the majority of existing research on Loneliness, we treated loneliness as a unidimensional construct that arises from perceived deficits in social relationships (Ernst and Cacioppo 1999). For example, Russell (1996) suggested that the UCLA loneliness scale should be considered as a unidimensional measure and supported his argument with the factor structure of the scale. However, Weiss (1973)

conceptualized loneliness as a bidimensional construct and proposed two distinct types of loneliness, related to different types of relational deficits. According to Weiss, emotional loneliness comes from the lack of an intimate attachment (e.g., a spouse, lover, parent, child), whereas social loneliness derives from the lack of a larger circle of friends and acquaintances that can provide a sense of belonging and companionship. One could argue that emotional loneliness is more predictive of product evaluation in the private consumption context, whereas social loneliness is more predictive in the public consumption context. Whether our results can be replicated or further moderated by different types of loneliness (social vs. emotional) can be a fruitful direction for future research.

We hypothesized and showed that given anonymity, lonely people preferred unconventional choice alternatives such as minority-endorsed options because such options fit better with their feelings of loneliness. The fit between the minority endorsement and their feelings of loneliness makes the information (i.e., endorsed by 20% of consumers) feel right. Therefore, they evaluate these products more favor- ably than those endorsed by the majority. Previous research suggests that information that fits with one's goals receives increased processing fluency and engagement (Lee, Keller, and Sternthal 2010). However, such advantages of fit in- formation could be either conscious or subconscious to the individuals. It is possible that lonely consumers naturally attend to minority endorsement information, without being consciously aware of it, because such information fits with

their feelings of loneliness and is hence more fluent to process. However, it is also possible that lonely consumers are aware of the fit between the minority endorsement information and their feelings of loneliness and hence consciously use such information as a basis of their judgment. One way to tease apart these two possibilities is to make salient other goals such as accuracy or involvement (e.g., Wang and Lee 2006). It is plausible that if consumers consciously choose minority-endorsed products to fit with their feelings of loneliness, the fit effect should disappear when they have a salient goal to be accurate. However, if consumers subconsciously prefer the minority information that fits with their feelings of loneliness> the fit effect should remain, even with a salient goal of accuracy. Such tests merit future investigation.

Our hypothesis in the public consumption contexts was partially based on findings in the social exclusion literature, in which social exclusion was commonly manipulated by rejecting participants (e.g., "I hate to tell you this, but no one chose you as someone they wanted to work with"; Twenge et al. 2007, study 2) or ignoring them (e.g., whether participants were included or ignored in an online game; Williams et al. 2000), An important question then centers on the distinction between social exclusion and loneliness. In our view, social exclusion involves an active process of

rejecting or ignoring people. In either case, there is some action engaged (rejecting or ignoring). However, loneliness refers to a passive state or chronic personality in which people feel the undesirable social isolation. The construct of loneliness does not involve any action but merely rep- resents how people feel at a particular point of time (or always). It is plausible that social exclusion may be an (important) antecedent of loneliness, but loneliness could also emerge from other factors such as a chronic personality trait, lonely people's own interpersonal behaviors, and network deficits in the social environment (Ponzetti 1990). Future research examining the relationship between social exclusion and loneliness should shed more light on the antecedents and consequences of the two constructs and whether our results based on loneliness could also be obtained with social exclusion.

Aside from the distinction between loneliness and social exclusion, our findings extend current understanding in the social exclusion literature. Previous research on social exclusion only showed that socially excluded people conform to gain social connection in public contexts (e.g., Mead etal. 2011), but it is unclear what their default preference

would be in private contexts and what mechanisms underlie such effects. We demonstrated that lonely participants preferred the majority-endorsed options only when their preferences were subject to public scrutiny. When they made decisions for private consumption, they preferred minority-endorsed options. Moreover, our article contributes to the literature by identifying the mechanism underlying the preference reversal. When lonely participants were not afraid of negative evaluations (in private contexts), the minority-endorsed products fit better with their feelings of loneliness, and such fit mediated the effect of loneliness and endorsement on evaluation. However, when negative evaluation became a concern (in public contexts), lonely participants conformed to the preference of the majority to avoid negative evaluations.

One of the reasons why loneliness has been a lingering social problem is that it seems to be trapped in a vicious circle. Lonely people tend to have biased perceptions and less accurate interpretation of emotional or social cues (Pick-

ett and Gardner 2005). Such perception leads to defensive behaviors, withdrawal, and passive coping. This passivity in turn results in reduced social interactions, which reinforces their loneliness status and confirms