Most customers can recall a service failure experience (Lewis and Spyrakopoulos 2001). When failures occur in servicescapes (Bitner 1992), the presence of other customers seems likely to impact evaluations of the incident. Despite progress in understanding social presence effects on consumption (Zhuang et al. 2006; Luo 2004; Mangleburg, Doney, and Bristol 2004; Dahl, Manchanda, and Argo 2001; Argo, Dahl and Manchanda 2005; Fisher and Dube 2005), whether and how the presence of others affects consumer reactions to a service failure is unknown. Seeking to enhance understanding of social presence effects in the services-cape, the following studies analyze social presence effects on customer evaluations following a service failure; examine cognitive and affective pathways through which social presence impacts evaluation; and build a nomological net that features social presence and outcome attribution. Our findings are relevant to researchers interested in service management theory and practitioners looking to improve service evaluation. Investigations concerning the effects of social presence have thus far based upon two major theories: the evaluation-apprehension theory and the distraction-conflict theory. Evaluation-apprehension theory assumes that the desire to make a good impression and the fear of negative evaluation become more salient when in public (e.g., Baumeister 1982; Seta et al. 1989). Thus, individuals are more likely to view themselves from a third-party observer perspective (Duval and Wicklund 1972; Wicklund and Duval 1971) in order to either make a good impression or avoid a bad one (Green 1989). In this process, differences between the actual and ideal self may become salient and lead to enhanced affective influence on self-evaluations of the experience. Supporting this prediction, several studies report higher levels of embarrassment when a social audience is believed to be present (Costa et al. 2001; Dahl et al. 2001; Argo et al. 2005). Such strong self-directed emotion is likely to reduce negative cognitions regarding the service failure as the consumer becomes more concerned with self-impression management and less concerned with shifting blame from self to other. An alternative theory of social presence effects, distraction-conflict theory, is based on the drive theory of social facilitation (Zajonc 1965, 1980), which predicts that the presence of others increases attention to the unexpected, preparedness to respond and physiological arousal. As such, distraction-conflict theory predicts that aroused attention to others in the services-cape following a service failure will conflict with the more immediate task, for example, evaluation of the failure. Social comparison and threat monitoring may also compete for cognitive resources (Uziel 2007). As a result, high cognitive loads are likely and the directly affected customer will experience more difficulty forming a negative evaluation of the service failure. For decades, advocates have marshaled supporting empirical evidence for each theory (Argo et al. 2005; Baron 1986; Dahl et al. 2001). Most recently, researchers have argued that there may be multiple processing pathways and outcomes associated with social presence (Uziel 2007). The present research applies this integrated model and posits that each different reaction to social presence can be made salient at a given moment due to a situational factor, outcome attribution. We argue that the evaluation-apprehension theory explains social presence effects when the service failure is externally-attributed, whereas the distraction-conflict theory accounts for social presence effects when the service failure is internally-attributed. Two studies are designed to examine these predictions among individuals with low cognitive load (study 1) and with high cognitive load (Study 2). Study 1 demonstrated that compared with no social presence condition, social presence was associated with higher service evaluation pertaining to the service failure regardless of outcome attribution [M's = 3.31 versus 4.25, t (106) = 3.09, p < .01]. Further mediating analyses, however, uncovered the complexity of the above findings. In particular, the results demonstrated that anger fully mediated the relationship between social presence and service evaluation when a service failure was externally-attributed. When the service failure was internally-attributed, number of negative thoughts fully mediated the relationship between social presence and service evaluation. This indicated that when the service failure was externally-attributed, the affective process proposed by the distraction-conflict theory accounted for the social presence effects. When the service failure was internally-attributed, the cognitive process suggested by the evaluation-apprehension theory drove the social presence effects. Study 2 showed that under high cognitive load, the affective process underlying social presence effects during an externally-attributed service failure still held, that is, social presence (versus no social presence) attenuated other-directed negative emotion (i.e., anger) elicited by the service failure, and thereafter led to more positive service evaluations (M's = 5.29 versus 4.12, t [53] = 2.66, p < .02). By contrast, during an internally-attributed service failure, high cognitive load blocked the cognitive process of social presence effects, and as a result, the effect of social presence on service evaluation disappeared. To conclude, this research holds theoretical significance in that investigating social presence as a moderating factor related to consumer reactions to a service failure helps enrich the traditional service evaluation paradigm. This research also holds managerial significance as understanding the crucial role of social presence in service encounter may help service providers to be more conscientious about the social environment, which consequently leads to improved service evaluation. [ABSTRACT FROM AUTHOR]