Cognitive Dissonance in Technology Adoption: A Study of Smart Home Users

Drawing on the Theory of Expectation-Confirmation, the individuals’ evaluation and satisfaction of the experience with technology is the result of the comparison of expectations and the performance (Bhattacherjee, 2001, Dai et al., 2015). Expectations refer to pre-exposure beliefs about a service or product (Susarla et al., 2003). The evaluation of pre-purchase expectations with actual performance can lead either to the confirmation or disconfirmation of the expectation. Confirmation results from the match between pre-exposure expectations and actual performance, while disconfirmation is the outcome of performance which is inconsistent with expectations. Disconfirmation is positive when actual experience with the use exceeds prior beliefs about the use. Negative disconfirmation occurs when performance falls short of expectations (Oliver, 1980, Kopalle and Lehmann, 2001). The inconsistency between the degree of perceived performance and prior beliefs represents the conflict of the two types of cognition, which can be explained by Cognitive Dissonance Theory. The inconsistency causes dissonance, which is associated with psychological discomfort (Festinger, 1962). The intensity of dissonance differs depending on the degree of discrepancy between initial cognition and the cognition after the exposure to technology use. The discrepancy can be small, falling within the zone of tolerance, without triggering dissonance arousal. As the magnitude of the discrepancy increases, the probability and the magnitude of dissonance arousal increases too (Szajna and Scamell, 1993). Dissonance can arise not only due to the discrepancy between expectation and performance, but the comparison of pre-service and post-service performance of technology or IS systems (Park et al., 2015). Pre-service performance may include the quality of pre-service customer service or website design. Post-service performance includes the evaluation of the object’s attributes related specifically to the use of the technology or IS system (Park et al., 2015). Based on the above, our first hypothesis is put forward:

Dissonance is associated with discomfort and uneasiness, which reflect negative emotions. The strength of emotions demonstrates the degree of dissonance arousal (Festinger, 1962). Past research has identified three emotions that can be associated with cognitive dissonance. The first is anger, occurring when people feel they are not responsible for the situation causing dissonance and/or are incapable of fulfilling the task (Harmon-Jones, 2004, Harmon-Jones et al., 2017). Anger is defined as a basic emotion, holding a number of other underlying similar, yet different emotions, like frustration, irritation or bitterness (Shaver et al., 1987). It has been reported that people who experience stronger cognitive dissonance have a stronger perception of anger and aggression (Soutar and Sweeney, 2003). The relationship between dissonance and anger explains the negative outcome of service performance and use of technology. For example, failure in technology performance raises anger and withdrawal behaviour, such as boycotting the retailer of the product (Donoghue and de Klerk, 2013). The use of technology contributes to the experience of anger and anxiety in situations when people have low self-efficacy in the use of computers (Wilfong, 2006). Self-efficacy represents the state when technology users feel incapable of realising the expected services (Bandura, 1977). Therefore, dissonance reflecting a disconfirmed belief about personal technical competence is more likely to be associated with anger.

The second emotion is guilt (Gosling et al., 2006, Turel, 2016). Guilt is associated with a feeling of shame and self-disappointment and can explain the psychological state between cognitive dissonance and the intention to discontinue the use of technology. Guilt is a response to the behaviour that causes moral dilemmas, such as the inconsistency with personal norms, values and self-standards (Harmon-Jones et al., 2017). Guilt can be experienced when a person feels responsible for the failure of technology performance, causing inconsistency with internal norms. The higher the control over the behaviour, the higher is the perception of guilt (Burnett and Lunsford, 1994). For example, IT addiction raises self-attributed negative emotion (i.e. guilt), which reflects the perception that a person is not capable of rationally utilising the technology and realising desired goals (Vaghefi and Qahri-Saremi, 2017). Other incidents with technology inducing guilt may include the excessive use of technology at the expense of important tasks (Turel et al., 2011) or ethical implications of the use of technology (Harrington, 1996).

The third emotion related to dissonance is regret (Roese and Summerville, 2005, Gilovich et al., 1995b). This is one of the negative outcomes of purchase decisions resulting in disconfirmed expectations (Oliver, 2014). Regret reflects self-blame for the behaviour that should not have been performed (Connolly and Zeelenberg, 2002, Gilovich et al., 1995b). Regret can be experienced when individuals choose a particular technology out of similar alternatives. In post-purchase situations, the strength of regret is conditioned by the degree to which non-selected alternatives represent the value for the individual. The experience of regret is stronger when the evaluation of the foregone alternative is increasing (Croyle and Cooper, 1983). For example, regret is experienced when the use of technology causes problems. Negative implications devalue the chosen technology and induce considerations about alternatives that could have been acquired instead (Dhir et al., 2016). Regret may occur not only as a result of issues with the utilisation of technology, but after the exposure to positive information about the services of an alternative technology (Kang et al., 2009). Also, individuals can feel regret when they realise that an alternative product could have been acquired at a lower cost (McConnell et al., 2000). Based on the above, the next hypothesis is proposed:

Emotions mediate the dissonance arousal and reduction processes (Festinger, 1962) because emotions are able to motivate and organise cognitions and actions (Izard, 2010). Emotions help interpret the signals of social interaction, communication and feeling states which underpin cognitive appraisals (Izard, 2010). Prior research has examined negative emotions in dissonance reduction as a unidimensional construct, embracing anger, fear, regret and anxiety (Jean Tsang, 2019, Gosling et al., 2006). However, this approach can be questioned given that emotions represent a complex process that guides people differently in various situations (Izard, 2010). Emotions can be differentiated by three aspects: a) affective valence, b) motivational direction and c) arousal. Affective valence refers to the degree to which people are positive or negative about the felt emotion and the state (Harmon-Jones et al., 2011, Harmon-Jones et al., 2017). Motivational direction refers to the role that the emotion plays in approach (behaviour aimed at reaching the goal) or avoidance (aversion from the goal achievement) behaviour (Harmon-Jones et al., 2017). The commitment to the behaviour by changing attitude and strengthening positive attitudes through the exposure to consonant information falls into approach behaviour. The lack of commitment, such as a change of behaviour as a result of dissonance, refers to withdrawal behaviour (Harmon-Jones, 2004). Arousal is the intensity of the feeling and psychological response to it (Harmon-Jones et al., 2017). In terms of affective valence, anger, guilt and regret refer to negative emotions. It is considered that negative emotions inhibit behaviour, which indicates withdrawal motivation (Harmon-Jones, 2004, Watson, 2000). However, when it comes to motivational direction, these types of emotions have a distinctive role in the cognitive dissonance strategies and the commitment to the behaviour causing dissonance (Harmon-Jones, 2004, Harmon-Jones et al., 2017). The distinctive motivational role of emotions is explained by different conditions in which emotions are manifested. The conditions include the degree of control over behaviour, the extent of responsibility for behavioural outcome, the justifiability of behaviour, the availability of better behavioural alternatives and the degree to which behaviour violates personal or social norms (Smith and Lazarus, 1993, Harmon-Jones et al., 2017, Harmon-Jones et al., 2003, Amodio et al., 2007, Connolly and Zeelenberg, 2002, Gilovich et al., 1995a).

Evidence suggests that anger resulting from the use of technology negatively affects its continuous use (Beaudry and Pinsonneault, 2010), which indicates the role of emotion in motivating avoidance behaviour. However, the findings on the motivational direction of anger are conflicting (Harmon-Jones, 2004, Harmon-Jones et al., 2004, Harmon-Jones et al., 2017, Carver, 2004). This inconsistency may be rooted in two reasons. First, the feeling of anger is often associated with other related emotions (e.g. irritation, shame, anxiety), motivating approach or avoidance behaviour. The interrelationship with other emotions affects the motivational direction of anger. For example, anger coupled with anxiety facilitates behaviour withdrawal (Harmon-Jones et al., 2017). Secondly, the motivational direction of anger depends on whether individuals feel responsible for the anger-inducing event and whether they have opportunities to undo the event. In situations of being intentionally harmed by another party, anger activates an approach-behaviour (Harmon-Jones et al., 2017). The common response in such situations is to punish the responsible party (Smith and Lazarus, 1993). However, the motivation to initiate any response is mitigated when there is no opportunity to ameliorate the situation causing anger (Harmon-Jones et al., 2003). Such anger is associated with feeling incapable of achieving the initial goal. It triggers the desire to change the goal orientation and switch to alternative options (Harmon-Jones, 2004, Carver, 2004). For example, anger is manifested when the use of technology inflicts security threats (Liang et al., 2019, Beaudry and Pinsonneault, 2010). When security threats occur, reduced commitment to technology and a subsequent behaviour withdrawal represent a defensive mechanism to avoid similar negative outcome in the future (Beaudry and Pinsonneault, 2010). Anger has a pro-active role in users’ behaviour, as it encourages individuals to seek out external means to cope with the emotion, which leads to the derogation of the behaviour causing anger (Liang et al., 2019). Also, it has been found that the failure of an appliance induces different levels of anger. The highest level of anger correlates with the intention to redress the experience and discontinue behaviour (Donoghue and de Klerk, 2013). As such, users who experience anger induced by an unexpected and unsatisfactory result of the use of technology are more likely to switch to another behaviour, rather than try to justify the negative outcome.

Guilt is considered to be a self-regulatory emotion (Amodio et al., 2007). There are two theoretical perspectives on the role of guilt in motivating behaviour (Turel, 2016, Amodio et al., 2007, Harmon-Jones et al., 2017). There is evidence that guilt motivates avoidance mechanisms, namely, the discontinuation of technology usage (Turel, 2016). Another perspective postulates an opposite role of guilt in behaviour (Harmon-Jones et al., 2017, Amodio et al., 2007). In morally violating situations, people tend to look for the means to resolve guilt, which contributes to approach-motivational orientation (Harmon-Jones et al., 2017, Amodio et al., 2007). The cognitive dissonance reduction through attitude change and consonant information-seeking represent the means to resolve guilt. They reflect the way to justify an action retrospectively and continue the behaviour by subduing negative emotions (Ghingold, 1981b, Kelman, 1979). For example, people who engage in conversations to reduce psychological tension have a lower level of regret than people who do not try relief dissonance through communication (Stice, 1992). Such conversations represent a form of cognitive adjustment. Given the above, it is hypothesised that:

The literature provides evidence about the effect of regret on avoidance motivation (Gilovich et al., 1995b, Davvetas and Diamantopoulos, 2017). There are two reasons to suggest that regret has a negative effect on the continuous use of technology. The motivation for withdrawal behaviour stems from cognitive processes associated with regret, such as weak self-esteem and strong self-blame. Regret is a painful feeling, since it implies a personal fault in the negative outcome and raises counterfactual thinking (Connolly and Zeelenberg, 2002, Gilovich et al., 1995b). Counterfactual thinking refers to the ruminations about alternative decisions and potential consequences (Roese, 1997). Counterfactual thinking is conditioned by the availability of alternative options and opportunities accordingly. When an individual does not have any opportunities or opportunities imply inevitable negative consequences, the individual either mitigates or terminates the feeling of regret through attitude change, thus maintaining behaviour. In contrast, alternative decisions entailing a positive outcome facilitate the feeling of regret (Roese and Summerville, 2005) and predict switching behaviour (Lee and Lee, 2012). Feeling regret often results in corrective actions, a change of decision and behaviour, such as switching service providers if they fail to meet service requirements (Zeelenberg and Pieters, 1999). For example, a scenario-based experiment found that regretful decisions positively affect the intention to discontinue the use of technology and negatively affect satisfaction (Davvetas and Diamantopoulos, 2017). Regret experienced after the appraisal of service performance positively correlates with switching behaviour (Mattila and Ro, 2008, Sánchez-García and Currás-Pérez, 2011). Regret is a stronger predictor of behaviour modification when individuals compare the actual outcomes with better alternatives (Roese and Morrison, 2009). In addition, prior research provides evidence on the correlation between information-seeking and regret. It is suggested that the exposure to information about alternatives increases experienced regret (Keaveney et al., 2007), which potentially leads to a higher dissonance and the motivation to avert behaviour in order to reduce dissonance. Hence, this study proposes that:

According to Cognitive Dissonance Theory, the behaviour that arouses dissonance is associated with a negative affective state (e.g. dissatisfaction). In conditions when dissonance is aroused, satisfaction with the behaviour can be achieved if the psychological discomfort caused by disconfirmed expectations is eliminated. That happens by reducing the discrepancy between prior expectations and perceived performance using one of the dissonance reduction strategies (Shahin Sharifi and Rahim Esfidani, 2014, Dutta and Biswas, 2005, Festinger, 1962). Although all three dissonance reduction mechanisms (attitude change, consonant information search and behaviour change) reduce psychological tension, they trigger different levels of satisfaction with the behaviour causing dissonance. Specifically, attitude change and consonant information search refer to the cognitive dissonance reduction mechanisms that change the cognition (i.e. reinforcing positive beliefs about the behaviour), thus encouraging individuals to carry on the behaviour that initially caused dissonance (Harmon-Jones and Mills, 2019). By changing attitude and seeking consonant information, users increase the likelihood of experiencing satisfaction and perceived wellbeing (Festinger, 1962). In contrast, behaviour change reduces the psychological tension by eliminating the source causing dissonance (i.e. behaviour). That means that although the psychological tension is eliminated, the individual stays dissatisfied with the behaviour (Festinger, 1962). Such a theoretical explanation of the relationship between behaviour change and dissatisfaction is different from the stream of research which focuses on the relationship between disconfirmation – dissatisfaction – switching behaviour (Fan and Suh, 2014, Zhang et al., 2016, Lu et al., 2012, Nam et al., 2018) and overlooks the role of dissonance and dissonance reduction strategies. In this research, given the established dissonant state, withdrawal behaviour is one of the measures that people employ before evaluating satisfaction. The supporting arguments can be drawn from prior research, which found that users who are more committed to the behaviour are more likely to view the selected choice favourably and in turn experience higher satisfaction (Brehm and Cohen, 1962). For example, when individuals are engaged in interactive reflection on the behaviour, they change their cognition by strengthening their positive attitude to the behaviour and improving self-perception (e.g. self-confidence, self-awareness and self-knowledge) (Jones and Oswick, 2007). Sparks et al. (2012) examined a correlation between personality traits, reduction strategies and perceived satisfaction. They found that people who tend to maximise outcomes (maximisers) tend to withdraw behaviour, which results in less satisfaction. In contrast, non-maximisers tend to change the attitude towards the choice and perceive a stronger level of satisfaction. The study by Vroom and Deci (1971) provides evidence about the positive effect of the cognitive adjustment on satisfaction. The findings of the research postulated that when people do not engage in dissonance reduction through the change of cognition following the perception of discrepancy between expectation and the actual outcome, they show stronger dissatisfaction (Vroom and Deci, 1971). Given that satisfaction is a predictor of perceived wellbeing (Lee et al., 2002), discontinuous behaviour can be negatively associated with perceived wellbeing. Hence, the following hypotheses are put forward in relation to the cognitive dissonance coping mechanisms users may deploy:

Perceived wellbeing is a perceived impact on important life domains, which underpins the evaluation of the overall quality of life (El Hedhli et al., 2013). Perceived wellbeing reflects experiences with consumer goods and services (Lee et al., 2002). It is the result of satisfaction with the acquisition, consumption, possession and disposition of a product or service. Satisfaction in the consumer life domain has a spill over effect on other life domains (Lee et al., 2002). Wellbeing captures the cumulative satisfaction with the product and the positive experience that it has on user life, social life, leisure life and community life (El Hedhli et al., 2013). In other words, wellbeing is predicted by the satisfaction experiences, such as family relationships, the status in society, material possessions and education (Lee et al., 2002). This is of particular importance to the empirical setting of this work, namely smart homes. Smart homes aim to deliver individual and societal benefits by assisting in daily routines, delivering comfort, decreasing natural resource consumption (energy and water) and in turn reducing utility bills (Marikyan et al., 2019). Given that the aim of technology is to satisfy users’ needs, which tackle different aspects of life, a strong perception of fulfilled needs can contribute to user perceived wellbeing. The hypothesised relationships in the research model are provided in Fig. 2.