Create a Plan for the Integration of New Technologies

Computers in Human Behavior 53 (2015) 381–395

Contents lists available at ScienceDirect

Computers in Human Behavior

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / c o m p h u m b e h

Full Length Article

Understanding the most influential user experiences in successful and unsuccessful technology adoptions

http://dx.doi.org/10.1016/j.chb.2015.07.012 0747-5632/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. E-mail address: [email protected] (T. Partala).

Timo Partala ⇑, Timo Saari Human-Centered Technology, Department of Pervasive Computing, Tampere University of Technology, P.O. Box 589, FI-33101 Tampere, Finland

a r t i c l e i n f o

Article history: Received 4 September 2014 Revised 8 July 2015 Accepted 9 July 2015 Available online 23 July 2015

Keywords: Technology adoption Technology acceptance User experience Emotions User needs Well-being

a b s t r a c t

Understanding processes underlying technology adoption or non-adoption is an important research theme often addressed using the technology acceptance model (TAM) approach. The objective of this research was to investigate most influential user experiences in successful and unsuccessful technology adoptions using user experience related concepts and methods in conjunction with the TAM. Participants (N = 76) described their most influential user experiences related to one successful and one unsuccessful technology adoption process and evaluated both experiences using rating scales, including the central TAM related scales and user experience related scales probing emotions, psychological needs, user values, task load, and the impact of technology on the user’s well-being.

The results suggested that user experience and technology acceptance related viewpoints can comple- ment each other in order to gain a more holistic understanding of the factors affecting the success or fail- ure of technology adoptions, and the results showed how these variables typically behave in both contexts. The overall valence of user experience was significantly affected by perceived usefulness, the fulfillment of psychological needs, and the salience of negative emotions in the most influential user experiences of successful adoptions, and by perceived usefulness, output quality, and the salience of neg- ative emotions in the unsuccessful adoptions.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Technology adoption is a growingly important research theme, as new technologies are being rapidly introduced, and it is cur- rently typical for a user to have many personal devices and interact with multiple information technology services including different commercial and public services. Different information technology devices and services are now also increasingly used in the less developed countries. From the user’s viewpoint, successful tech- nology adoption is important in order to fully participate in the rapidly changing modern society and it is also crucial in many pro- fessions. Those who cannot adopt new technology are limited in their ability to become productive members of their communities. Understanding the factors influencing technology adoption helps us predict and manage, which technologies are successfully adopted, by whom, and under what kind of conditions. Designers of technology and technology-based services can also use that information toward building systems, which can overcome the most common barriers in technology adoption.

The most popular approach for studying technology adoption has been the Technology Acceptance Model (TAM) and the related questionnaire-based research methods. Davis (1989) developed the original TAM model based on the theory of reasoned action (Fishbein & Ajzen, 1975). In the TAM model, system usage behavior is determined by behavioral intention to use, which is affected by perceived usefulness, or ‘‘the degree to which a person believes that using a particular system would enhance his or her job perfor- mance’’ and perceived ease-of-use or ‘‘the degree to which a per- son believes that using a particular system would be free from effort’’. These perceptions are in turn affected by a number of external variables The original model was later extended into TAM2 (Venkatesh & Davis, 2000), in which the external variables of subjective norm, image, job relevance, output quality, and result demonstrability were identified, affecting the perceived usefulness and behavioral intention variables of the original model. The model also introduced experience of using the system and voluntariness of use as moderating variables. Later, Venkatesh and Bala (2008) released the third version of the model, TAM3, in which further determinants of perceived ease of use were added to the model: computer self-efficacy, perceptions of external control, computer anxiety, computer playfulness, perceived enjoyment, and objective usability.

382 T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395

Despite its importance for the technology acceptance research field and practical studies, TAM has also received a great deal of criticism. For example, Legris, Ingham, and Collerette (2003) sug- gested that TAM and TAM2 models explain only about 40% of infor- mation technology use and significant factors are not included in the models. Bagozzi (2007) published a commentary article point- ing out a number of limitations with the TAM model and the related research. A major conclusion was that besides variables included in TAM, new variables are needed for understanding the users’ decision making related to technology adoption and these variables are likely to be grounded in emotional, social, and goal-directed behavior research.

Recently, these kinds of variables have been applied to the study of interactive systems in the research field of user experience (UX), which is currently one of the most significant directions in human technology interaction research. While there is plenty of literature on technology acceptance and a growing body of research on user experience, only a little research exists combining both research directions. Some of the most potential user experi- ence related concepts include emotions, user needs, and user val- ues, all of which have been already applied successfully in practical user experience studies based on well-established theo- ries from psychological research. Another promising concept is experienced well-being impact of technology, which can be assumed to be especially important when studying technologies used daily or frequently such as different work systems and equip- ment. In the following chapters, we introduce these user experi- ence related concepts in detail and present the aims and hypotheses of the current study.

2. User experience

User experience is a growing research field, which has in recent years gained considerable interest from both scholars and practi- tioners. Generally, the research field of user experience is seen to include all factors, which affect the user’s interaction with and experience of a system or a product. At the core of the concept is experience itself. The current scholarly conception of user experi- ence mostly follows the psychological line of thinking, in line with the ‘‘technology as experience’’ approach (McCarthy & Wright, 2004), which shifted the focus of product design from technologi- cal and pragmatic aspects toward subjective and emotional quali- ties of interaction with products. Many different definitions have been suggested for user experience stressing different aspects of experience or factors affecting user experience. However, most researchers and practitioners agree that user experience is subjec- tive (vs. objective), holistic (vs. instrumental), situated (vs. abstract), and dynamic (vs. static) (Hassenzahl, 2010; Law, Roto, Hassenzahl, Vermeeren, & Kort, 2009). Below we briefly introduce some of most important established and emerging concepts and methods of user experience research, all of which are also applied in the current research.

2.1. Emotions

In the study of human experiences, emotion is generally seen one of the most central and pervasive aspects and consequently they also play a central role in understanding user experiences. Most of the scientific research on user experience follows the notion is that emotions are integral to experiences and they are also intertwined with our actions (Carver & Scheier, 1989; Hassenzahl, 2010). Emotions can be studied from two main view- points: dimensional and discrete emotions, which are currently seen as complementary to each other. Empirical research using the dimensional viewpoint has shown that emotions can be

organized along two main dimensions: valence (ranging from neg- ative to neutral and positive affect) and arousal (ranging from very calm to neutral and very highly aroused). From the discrete emo- tions point of view, there are numerous theories and methods on which particular emotions are the most central and should be included in emotion research. One of the most widely used meth- ods has been the Positive Affect Negative Affect Schedule (PANAS, Watson, Tellegen, & Clark, 1988), which studies experienced emo- tions through a balanced set of ten positive and ten negative emo- tions (e.g. inspired, excited, scared, distressed). There is evidence about the reliability and validity of the method in a general popu- lation (e.g. Crawford & Henry, 2004). This method enables calculat- ing salience scores separately for positive and negative emotions, as well as an affect balance score. This is especially useful, as there is evidence that people can experience both positive and negative emotions as parts of the same experience (Russell & Carroll, 1999).

Emotion-related concepts have a history of being used in tech- nology adoption studies, but for example, Straub (2009) has noted that there is not a sufficient empirical basis for understanding the influence of emotions on the technology adoption process. The research on emotion-related constructs in technology adoptions is mostly limited to the notions included in Technology Acceptance Model 3 that perceived enjoyment and computer anx- iety are (positively and negatively, respectively) related to per- ceived ease of use (Venkatesh, 2000). Other researchers have also found these variables to be related with other TAM variables in dif- ferent contexts. For example, Van der Heijden (2004) found that perceived enjoyment can be directly related to intention to use in hedonic information systems (e.g. a movie website). There are, however, also a few studies that have used more advanced emo- tion concepts. Cenfetelli (2004) averaged a range of different posi- tive and negative emotions and found that positive emotions were positively related and negative emotions were negatively related to perceived ease of use of the TAM model in an e-business context. Kim, Chan, and Chan (2007) used the two main emotional dimen- sions: valence and arousal, in a study focusing on the continued use of mobile Internet services, and found that both dimensions were positively related to attitude toward using the services (the concept which predicts actual intention to use in TAM2). Based on appraisal theory, Beaudry and Pinsonneault (2010) developed a framework of four emotions: anger, anxiety, excitement, and happiness, and found that these emotions were either directly or indirectly related to IT use among bank account managers.

2.2. Psychological needs

Another central concept in understanding human experiences and also well-being is the concept of psychological needs. Self-determination theory (Deci & Ryan, 2014; Ryan & Deci, 2000) suggests that three needs are of central importance: autonomy (to actively participate in determining own behavior without external influence), competence (to experience oneself as capable and com- petent in controlling the environment and being able to reliably pre- dict outcomes), and relatedness (to care for and be related to others). Sheldon, Elliot, Kim, and Kasser (2001) presented a model of ten can- didate psychological needs extending self-determination theory with seven other needs: self-actualization-meaning, physical thriv- ing, pleasure-stimulation, money-luxury, security, self-esteem, and popularity-influence. They also presented a questionnaire method for studying the degrees of satisfaction for the ten needs using 30 statements (three statements for each need) and applied the method in two studies on the most and least satisfying experiences of college students in two different cultural settings. The results showed that autonomy, competence, and relatedness were consistently among the most salient needs, together with self-esteem needs.

T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395 383

Later, the model of ten psychological needs and the related methods have been found to be useful in understanding users’ experiences and behavior. Psychological needs are seen as ‘‘be-goals’’ underlying the users’ actions, or ‘‘do-goals’’ (Carver & Scheier, 1989; Hassenzahl, 2010). For example, the need for being related may motivate a phone call to a close person. Hassenzahl, Diefenbach, and Göritz (2010) studied positive user experiences with technology using the framework and method by Sheldon et al. (2001) and found a clear relationship between need fulfill- ment and positive experiences, with stimulation, relatedness, com- petence, and popularity as especially salient needs. Using similar methods, Partala (2011) was able to understand salient needs motivating the usage virtual worlds (e.g. autonomy, self-esteem, physical thriving), and Partala and Kallinen (2012) found the needs of autonomy, competence, and self-esteem as salient in underlying most satisfying user experiences over a range of different technolo- gies. In the context of technology adoptions, psychological needs have been studied only rarely. Roca and Gagné (2008) found that the needs of autonomy and competence were significantly related to the central TAM concepts of perceived usefulness and perceived ease of use in the context of predicting e-learning continuance. Also in the context of e-learning continuance, Sørebø, Halvari, Gulli, and Kristiansen (2009) found that the psychological need of competence had an important role in predicting perceived use- fulness, as well as intrinsic motivation for use.

2.3. Values

User values are a concept, which has gained increased attention in the past few years. Like psychological needs, personal values have been found to be related to emotions (e.g. Desmet & Hekkert, 2007). For example, in an early explorative study in the context of consumer experiences, Laverie, Kleine, and Kleine (1993) found that consumers’ values can be related to both posi- tive and negative emotions. The research on personal values in technology adoptions has been scarce, even though Malhotra and Galletta (1999) found in an early study that social influence (e.g. whether a system is line with the users’ values) strongly influenced the user’s attitude toward using a system. Srite and Karahanna (2006) have also found that national cultural values can have a moderating role in technology acceptance. Schwartz (1992) pre- sented a well-known model of ten universal values based on exten- sive empirical studies in multiple countries: power, achievement, hedonism, stimulation, self-direction, universalism, benevolence, conformity, tradition, and security. In a qualitative study. Isomursu, Ervasti, Kinnula, and Isomursu (2011) studied the suc- cess of the adoption of a school attendance control system in detail by using the framework of Schwartz (1992) as the basis of the anal- ysis. Recently, Partala and Kujala (in press) presented a quantita- tive approach aimed for evaluating the users’ personal values in relation to products and services systematically based on Schwartz’s model of ten universal values. This approach offers a possibility for theory-based evaluation of user values using quanti- tative methods. In their study on the role of universal values in the use of products and services, hedonism, self-direction, security, and universalism emerged as the value categories rated as the most important overall by their participants.

2.4. Well-being

There has been a recent interest in designing and evaluating information technology in relation to its impact on the users’ life well-being (e.g. Desmet & Hassenzahl, 2012; Kamp & Desmet, 2014), but this approach is not yet visible in technology acceptance literature. When the aim is to evaluate the impact of perceived long-term technology use and technology acceptance in the long

run, there is a need for more stable indicators of individuals’ well-being besides emotions, needs, and values. Subjective well-being combines the hedonic (e.g. well-being as an affectively pleasant state) and eudaimonic (e.g. well-being as living a mean- ingful life) approaches to well-being (Argyle & Martin, 1991; Keyes & Waterman, 2003). Subjective well-being is often conceptu- alized as a broad domain of interest rather than a specific construct (e.g. Diener, Suh, Lucas, & Smith, 1999), including the idea of pre- dominantly positive mood and emotional states and a cognitive evaluation of the person’s life circumstances and conditions, such as satisfaction with one’s life. One effective approach for studying well-being is through two complementary viewpoints: psycholog- ical flourishing and lack of depressive symptoms. Psychological flourishing assesses psychological wealth and well-being which goes beyond positive emotions and life satisfaction (Diener et al., 2009). It measures key aspects of psychological wealth such as social relationships, self-respect, competence, engaging work and spirituality as well as life purpose and meaning. Additionally, we are interested in the role of mental health in well-being via exam- ining the lack of depressive symptoms as a secondary indicator of well-being (e.g. Argyle & Martin, 1991; Elliot & Sheldon, 1998).

2.5. Task load

Finally, the concept of task load can be helpful in studying the perceived difficulty of the task or goal-directed activity at hand, and the associated effort, including mental, physical, and temporal effort invested by the user (Hart & Staveland, 1988). Understanding these concepts has been highlighted as an important challenge in usability studies by Hornbæk (2006). It is generally thought that any activity with interactive products will have demands on the user, and in design it is advisable to minimize extraneous task load to achieve better usability and more positive user experiences (e.g. Hollender, Hofmann, Deneke, & Schmitz, 2010).

3. Aims and hypotheses

Most of the existing research on technology adoption has con- centrated on studying the interrelations of the technology accep- tance model variables (and possibly additional variables) in the context of single systems or products. In the current study, we adopt a basic research oriented approach, in which technology adoptions are studied by means of analyzing the most influential single user experiences related to adoptions over a range of different products. Each participant described and rated the most influential user experience of one successful technology adoption and one unsuccessful technology adoption in a questionnaire. Using this approach allows within-subject comparisons between the most influential experiences of successful and unsuccessful adoptions. To our knowledge, a similar design drawing from both user expe- rience and technology adoption research has not been used before.

Thus, in the current research, the aim was to study the most influential user experiences in successful and unsuccessful tech- nology adoptions in order to gain a systematic understanding of those experiences. We aimed at forming an overall understanding of which variables are the most central in studies combining the technology acceptance and user experience paradigms. User expe- rience concepts described in the previous chapter were studied in conjunction with the central variables from the technology accep- tance model to investigate, whether there are significant relation- ships between those variables.

In line with many advances from previous literature focusing on user experience (e.g. Desmet & Hekkert, 2007; Hassenzahl, 2010; Hassenzahl et al., 2010; Partala & Kallinen, 2012) we assumed that the participants’ most influential experiences of successful and

384 T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395

unsuccessful technology adoptions would have clear differences on the level of emotional experiences. In line with this notion, we were especially interested in understanding the variables affecting the central dimensions of emotional valence and arousal in both of these contexts. Two additional, broader emotion-related variables, flourishing and depression were also of special interest in order to be able to study the perceived effect of the technology to be adopted to one’s life well-being. Thus, four emotion-related vari- ables (valence and arousal of user experience, flourishing, and depression) were selected as dependent variables, and the rest of the main variables were treated as independent variables. Following this arrangement, hypotheses for the relationships of the main variables in this study are presented in Table 1.

It was noted before that user experiences can involve varying levels of both positive and negative emotions, both of which can contribute to overall valence of the experience (Partala & Kallinen, 2012; Russell & Carroll, 1999). In line with this, we pre- dicted that the amount of positive emotions experienced would be positively associated with user experience valence in the con- text of both successful and unsuccessful technology adoptions and the amount of negative emotions would have a negative rela- tion with user experience valence in both cases. Riva, Banos, Botella, Wiederhold, and Gaggioli (2012) have suggested in the context of designing for positive technologies that emotional qual- ities and experiences can be linked to well-being. In line with this notion, we also predicted that positive emotions would have a pos- itive relation to flourishing and a negative relation to depression and that negative emotions would have exactly opposite effects.

For predicting arousal, typical interaction patterns of valence and arousal from existing literature were taken as a starting point (see e.g. Bradley & Lang, 1994). In line with this, it was hypothe- sized that positive emotions would enhance positive arousal in the context of most influential user experiences of successful tech- nology adoptions and alleviate negative arousal in the context of unsuccessful adoptions. In contrast, negative emotions could decrease the level of positive arousal in the context of successful adoptions and increase negative arousal in unsuccessful adoptions.

The central technology acceptance model variables (usefulness, ease of use, output quality and intention to use) were predicted to be positively associated with the valence of the most influential user experiences and perceived impact on positive well-being (flourishing), and negatively associated with perceived depression. For perceived enjoyment, similar effects were predicted to those of positive emotions. Task load was expected to contribute toward negative valence and high arousal, as well as less flourishing and more depression. In addition, based on previous findings, we expected that need fulfillment would have a positive effect on valence of user experience (Hassenzahl et al., 2010) and flourishing

Table 1 Hypothesized relationships between variables. Before the slash: relationships in the most influential user experiences in successful technology adoptions. After the slash: relationships in the most influential user experiences in unsuccessful technology adoptions.

Variable Valence Arousal Flourishing Depression

Positive emotions +/+ +/� +/+ �/� Negative emotions �/� �/+ �/� +/+ Usefulness +/+ 0/0 +/+ �/� Ease of use +/+ 0/0 +/+ �/� Output quality +/+ 0/0 +/+ �/� Intention to use +/+ 0/0 +/+ �/� Task load �/� +/+ �/� +/+ Enjoyment +/+ +/� +/+ �/� Need fulfillment +/+ 0/0 +/+ �/� Value concordance +/+ 0/0 +/+ �/�

+ positive relationship between variables; � negative relationship between vari- ables; 0 no significant relationship between variables.

(Ryan & Deci, 2000), and a negative effect on perceived depression (cf. Ryan & Deci, 2000) in the context of both successful and unsuc- cessful technology adoptions. Similar effects were predicted for value concordance based on corresponding findings from the field of psychology (e.g. Sheldon et al., 2004). Finally, it was predicted that need fulfillment and value concordance, as well as the tech- nology acceptance model variables would have no direct effects on perceived arousal.

Besides finding relationships between these variables, a second aim was to analyze the ratings of individual variables (e.g. specific emotions, psychological needs, and universal values) in the context of most influential user experiences of successful and unsuccessful technology adoptions. By doing so, it is possible to analyze, whether the specific variables, which have been suggested to be of central importance in previous research (e.g. the psychological needs of autonomy, competence, and relatedness), also turn out to be important based on the results of the current study. It is also possible to analyze, which of the variables differ significantly between most influential user experiences in successful and unsuccessful technology adoptions. Based on previous results uti- lizing somewhat similar basic research settings (e.g. Partala & Kallinen, 2012), we expected that the participants’ ratings on most of the studied individual variables would differ significantly between these two extreme conditions.

4. Materials and methods

4.1. Participants

76 participants (48 male and 28 female university students) completed all parts of the questionnaire with satisfactory responses. The mean age of the participants was 26.6 years (range 20–48 years). The participants consisted of 37 participants origi- nally from Finland and 39 international participants from different ethnic backgrounds. The participants also had different academic backgrounds ranging from technical to humanistic disciplines. All the participants answered to the questionnaire in English, which was also the official language on the courses, on which this study was arranged. The participants were instructed to use a dictionary, if they encountered unfamiliar terms in English. The participants were familiar with the basic concepts in the fields of usability and user experience and during the course they had also been introduced to the specific user experience related concepts studied in the current study (i.e. emotions, needs, values, etc.). It was assumed that at this point of their academic careers, the students had the required skills for analyzing their experiences using intro- spective methods and also capabilities for understanding all the scales used in the current study.

At the time of the reported most influential experiences of suc- cessful technology adoptions, the participants had been the using the evaluated product or system on average for 26.5 days, and at the time of the reported most influential experiences of unsuccess- ful technology adoptions on average 24.9 days (no significant dif- ference). The participants reported that they were using the system more frequently around the time of most influential expe- rience in a successful adoption when compared to the most influ- ential experience in unsuccessful adoption (respective average ratings 1.8 and 3.3 on the scale of 1 = several times each day, 2 = about once each day, 3 = several times a week, 4 = about once each week, and 5 = less than once each week).

4.2. Procedure

This research was carried out as part of course requirements for two user experience related courses at Tampere University of

T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395 385

Technology, Finland, and the participants consisted of Master’s stu- dents and doctoral students enrolled for the courses. If a student had enrolled for both of the courses, the assignment was carried out only once. The assignment was introduced for the participants during course lectures. Before answering to the questionnaire, the participants were asked to think of two kinds of technology adop- tions: one successful technology adoption (defined as a process, in which a participant had adopted some kind of interactive technol- ogy, for example, a system, an application or an interactive pro- duct, successfully into her/his long-term usage during the past year) and one unsuccessful technology adoption (a process, in which the participants had tried to adopt some kind of interactive technology, again, a system, an application or an interactive pro- duct, into their voluntary long-term usage, but – for any reason – the adoption process has been unsuccessful and the use of the technology has not become part of the participant’s routines). The participants were also instructed not to report adoptions of faulty technologies. As the most influential user experiences, the participants were instructed to report ‘‘the single experiences, which had the largest influence on the success (or failure) of the technology adoption’’.

The participants signed up for the study in the Moodle online learning environment, where they were identified using their uni- versity network IDs. The participants were provided a link to the questionnaire in Moodle, and they had more than two weeks to respond to the questionnaire, before it was closed. The actual ques- tionnaire was filled in the web questionnaire system Webropol. Individual questionnaire responses were anonymous, but in order to give course credit for participation, the researchers had access to the list of names of the students, who responded to the survey. This was achieved in practice so that after the participants had sub- mitted responses to both phases of the research anonymously, they were automatically forwarded to a different Webropol survey, in which they were prompted to enter their names and student num- bers. This way it was possible to keep track of the participants who had responded to the survey, but individual responses were still anonymous.

4.3. Tasks and materials

On the first page of the questionnaire, the participants were again told that in the study their task is to focus on one successful and one unsuccessful technology adoption of interactive technol- ogy and focus on the most influential single user experience in both technology adoption processes. They were also told that their responses will be confidential and anonymous and that the results of the research will be presented so that no single participants can

Table 2 The main variables, methods, and scales used in this study.

Concept Variable Scale M

Emotions Valence Pictorial 1–9 S Arousal Positive emotions 1–5 P Negative emotions

Technology acceptance Perceived usefulness Likert 1–7 T Perceived ease of use Output quality Intention to use

Enjoyment Perceived enjoyment Semantic differential 1–7 Task load Task load 1–7 N Psychological needs Need fulfillment Likert 1–7 User values Value concordance Likert 1–7 C Well-being Flourishing Likert 1–7 F

Depression Likert 1–7 C

a Items or scales were slightly modified to fit the current study.

be identified based on the results. On the second page, the partic- ipants’ task was to describe the most influential user experience in a successful technology adoption process in a text field in about 30–70 words. The definition for the most influential user experi- ence in a successful technology adoption and the instructions given earlier were repeated on this page. The participants were also asked to estimate, how many days they had been using the tech- nology, when the reported experience took place (days since first time of use, ‘‘0’’ if the reported experience happened on the first day of use). Further, they were asked to choose from five alterna- tives, how often they were using the system at the time of the experience (see section 2.1).

On the next three questionnaire pages (pages 3–5), the partici- pants were asked to analyze the most influential user experience in a successful technology adoption using blocks of rating scales, which are presented in Appendix A. A summary of the methods and scales used in this research is presented in Table 2.

On page three, the participants evaluated their emotions during the user experience using two different methods. First, they evalu- ated their experienced emotions using the two main affective dimensions: valence (unhappiness – happiness) and arousal (calm – highly aroused) using the self-assessment manikin method (SAM; Bradley & Lang, 1994), which contained two pictorial scales consisting of nine pictures each. On both scales, the user selected the picture, which best described his/her emotions by clicking on the picture. Second, the participants evaluated the extent of 20 specific emotions during the experience using the PANAS method (Watson et al., 1988) with the original set of ten positive and ten negative emotions and the original 1–5 scale (see Appendix A).

On page four, the participants rated their perceptions of the technology using the Technology acceptance model 3 scales (Venkatesh & Bala, 2008) for perceived usefulness, perceived ease of use, output quality, and intention to use. These scales were pre- sented to the participants as one block of 14 statements using the original 1–7 Likert scale ranging from ‘‘strongly disagree’’ to ‘‘strongly agree’’. On the same page, there were six NASA Task Load Index scales (Hart & Staveland, 1988) probing mental demand, physical demand, temporal demand, effort, frustration, and performance.

On page five, perceived enjoyment was first studied using four semantic differential 1–7 scales taken directly from Van der Heijden (2004). After that, psychological needs were studied using 16 statements based on the questionnaire items from Sheldon et al. (2001). The self-determination theory related concepts of auton- omy, competence, and relatedness were studied using three scales, while the other seven candidate needs suggested by Sheldon et al. (2001) were studied by one selected scale. User values were

ethod Source

elf-assessment manikin (SAM)a Bradley and Lang (1994)

ositive Affect Negative Affect Schedule (PANAS) Watson et al. (1988)

echnology Acceptance Model 3 (TAM3) Venkatesh and Bala (2008)

Van der Heijden (2004) ASA-TLXa Hart and Staveland (1988)

Sheldon et al. (2001)a

ustom statements based on Partala and Kujala (in press) lourishing scalea Diener et al. (2009) ES-Da Radloff (1977)

386 T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395

studied with a customized set of ten items inspired by a more detailed value questionnaire developed by Partala and Kujala (in press). The salience of each of the ten universal values (Schwartz, 1992, 2006) was studied with one item. Finally, the perceived well-being impact of technology was studied using 16 scales. The first eight scales were the items from Flourishing Scale (Diener et al., 2009) and the scales from nine to sixteen were selected from the CES-D scale (Radloff, 1977), studying flourishing and depres- sion, respectively.

It should be noted that some of the methods (marked with an asterisk in Table 2) were slightly adjusted to fit the current aim of analyzing single user experiences. The pictures used in assessing emotional valence and arousal were enhanced versions of the orig- inal pictures from the Self-Assessment Manikin (Bradley & Lang, 1994). To study task load, the original NASA-TLX scale anchors were retained, but instead of the original 21-point scale, we used a more convenient seven point scale, which was also consistent with most of the other evaluations. The statements were also slightly modified to fit the current aim of studying user experi- ences relative to using a particular interactive system or product. For example, the question ‘‘How mentally demanding was the task?’’ was adjusted to ‘‘How mentally demanding was your use of this system?’’. The statements probing psychological needs and user values had to be somewhat similarly modified in order to study user experiences. For example, the item ‘‘During this event I felt free to do things my own way’’ from Sheldon et al. (2001) studying the fulfillment of the autonomy need became ‘‘Using this system made me feel free to do things my own way.’’ Similar adjustments were also made to the items studying flourishing and depression. For example, the item ‘‘I felt that everything I did was an effort.’’ of the original CES-D method was changed to ‘‘Using this system made me feel that everything I did was an effort.’’ The original CES-D scale with four options was also chan- ged to a seven point Likert scale in line with most of the other evaluations.

On pages six to nine, the participants carried out the second phase of the questionnaire. In the second phase, the participants reported and analyzed their selected most influential user experi- ences related to unsuccessful technology adoptions using the same methods as in the first phase (writing a qualitative description of 30–70 words and filling in three questionnaire pages with the scales described in Appendix A). The definition for the most influ- ential user experience in an unsuccessful technology adoption, as well as the instructions given before the study were repeated on page six. On page ten, the participants filled in a brief personality survey using quantitative scales. After all, the scales on page ten were not informative enough to be included in the analyses reported in this paper. On the final page of the questionnaire (Page 11), the participant could report her/his gender and age, and also give feedback about the questionnaire. When the partici- pants had submitted the questionnaire, they were forwarded to the second brief questionnaire, in which only their names and student numbers were asked in order to give them course credit.

4.4. Data analysis

First, the data of selected single variables were averaged to form new variables as follows: positive emotions (average rating of the ten positive PANAS emotions), negative emotions (average rating of the ten negative PANAS emotions), task load (average of the NASA-TLX ratings), need fulfillment (average rating of the ten psy- chological needs), value concordance (average rating of the ten user values), flourishing (average rating of items 1–8 in the well-being block), and depression (average rating of items 9–16 in the well-being block, with item 12 reverse scored). In the anal- ysis of psychological needs, autonomy, competence, and

relatedness, which were studied using three items, were first aver- aged over those items, and after that, the scores for all the ten needs were averaged to form the need fulfillment score. The tech- nology acceptance model related variables were analyzed in line with the analyses in Venkatesh and Bala (2008) to form the con- cepts of perceived usefulness, perceived ease of use, output quality, and intention to use. Items studying perceived enjoyment (Van der Heijden, 2004) were reverse scored and then averaged.

Friedman’s rank tests were used to compare the participants’ ratings across multiple categories (e.g. the ten psychological needs) for significant differences and Wilcoxon’s matched pairs signed ranks tests were used in pairwise comparisons. These tests were selected because not all individual scales followed a normal distri- bution. Bonferroni corrected significance levels are presented in pairwise comparisons. Cronbach’s a scores were calculated to esti- mate the reliability of the scales consisting of multiple items.

The composite variables (and valence and arousal) followed normal distribution more closely. An examination of the skewness and kurtosis of the composite variables showed a skewness range from �1.35 for Intention to use (successful adoptions) to 1.51 for Enjoyment (successful adoptions) and a kurtosis range of �1.24 for Intention to use (unsuccessful adoptions) to 1.64 for Intention to use (successful adoptions). For regression analyses, assumptions of linearity, homoscedasticity, and lack of outliers were visually confirmed from P–P plots. Based on these findings, it was decided that multiple regressions can be used in the analysis of the data from the composite variables, valence, and arousal.

Hierarchical multiple regressions with two blocks of variables (method: ‘‘Enter’’) were used to analyze the ability of a set of inde- pendent variables (block 2) to predict the dependent variables (valence, arousal, flourishing, and depression) after controlling for three demographic variables: age, gender, and technology skills (block 1). The independent variables (usefulness, ease of use, inten- tion to use, output quality, task load, enjoyment, positive emotions, negative emotions, need fulfillment, and value concordance) entered the regression models as predictors, and separate regres- sions were run for each of the four dependent variables and for both most influential user experiences in successful and unsuc- cessful adoptions. Indicators of multicollinearity (tolerance and variance inflation factors) were well within acceptable limits (tol- erance > .25; VIF < 10) in all reported regression analyses.

5. Results

5.1. Technology adopted by the participants

The main focus of this research was on the quantitative data; however, brief qualitative data were also collected to understand the objects of the successful and unsuccessful technology adop- tions reported. In line with many related studies (e.g. Hassenzahl et al., 2010; Sheldon et al., 2001), the user experience descriptions collected in the current study were also too varied in many respects (e.g. length, contents) to warrant a full qualitative analy- sis. In line with previous user experience studies (e.g. Partala & Kallinen, 2012), the descriptions were also mainly very pragmatic, and as a consequence classifying user experiences on experiential dimensions was not feasible.

However, the objects of evaluation (the technology that the suc- cessful or unsuccessful adoption experiences were related to) were analyzed and categorized to aid in the interpretation of the current results. The technologies adopted by the participants were first categorized to devices and software technologies. In case of suc- cessful technology adoption, 44 out of 76 technologies were pri- marily about adopting devices, while the remaining 32 descriptions were about adopting software technology. When the

Table 3 Overview of the descriptive results (mean ratings). Scale 1–7 unless otherwise indicated.

Successful technology adoption

Unsuccessful technology adoption

Emotions Valence (1–9) 7.7 2.3 Arousal (1–9) 6.3 5.8 Positive emotions (1–5) 3.3 2.3 Negative emotions (1–5) 1.4 2.4

Technology acceptance

Perceived usefulness 5.8 2.5 Perceived ease of use 5.5 2.9 Output quality 5.6 2.6 Intention to use 6.2 3.4 Perceived enjoyment 5.8 3.2

Task load Task load 2.6 4.1 Psychological needs Need fulfillment 4.0 2.7 User values Value concordance 4.1 2.8 Well-being Flourishing 4.2 3.2

Depression 2.4 3.7

T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395 387

44 devices were further categorized, it was found that 22 descrip- tions were about mobile phones, six descriptions about tablet or e-reader devices, four about laptops, and three about accessories (e.g. a wireless mobile battery charger). The remaining nine descriptions were about other devices such as a digital camera, a smart TV, and a game console. Of the 32 adopted software technologies, 12 descriptions were about computer software applications (e.g. image processing software, a web browser), 7 were about mobile applications (e.g. a calendar app), 6 were about operating systems, and 5 were about software services (e.g. location-aware and social media services).

In case of unsuccessful technology adoption, 28 out of 76 tech- nologies were primarily about adopting devices, while the remain- ing 48 descriptions were about adopting a software technology. Out of the 28 devices, seven were mobile phones, four were tablet or e-reader devices, four were accessories (e.g. a game controller), and nine descriptions were about other kinds as devices such as digital cameras, a digital amplifier, or an ATM. Of the 48 adopted software technologies, 22 descriptions were about computer soft- ware applications (e.g. a video editing software, a CAD application), 14 were about mobile apps (e.g. calling and messaging apps, apps related to exercising, and an app for creating and sharing shopping lists), nine were about software services (e.g. different social media services and web services), and three were about operating sys- tems. In general, most of the technologies adopted successfully or unsuccessfully by the participants included both hardware and software components.

The technologies adopted by the participants were also classi- fied according to the purpose of usage to personal technologies (with a voluntary technology adoption context), technologies used for work or studies (with a more compulsory technology adoption context), and technologies explicitly stated to be used for both per- sonal and work/study purposes. The results of this analysis showed that 57 out of the 76 descriptions given in the context of successful technology adoptions were about personal technologies (e.g. vol- untary adoptions of personal mobile phones, tablets, or laptops with no mention of using them for work/study purposes). 13 descriptions expressed a work or study context (e.g. a project man- agement tool used at the workplace and a collaboration tool used in a school project), and six descriptions expressed both kinds of usage (e.g. a tablet computer used for reading both personal and study materials), or did not explicitly specify the context of use. For the unsuccessful technology adoptions, 56 descriptions were about personal technologies (again different personal devices and, for example, personals apps such as a sports tracker or a sleep cycle alarm clock). 17 descriptions expressed a work or study con- text (e.g. a data analytics program), and three descriptions did not explicitly specify the context of use.

1

2

3

4

5

Successful technology adop�on Unsuccessful technology adop�on

*** ***

***

***

*** ***

*** *** ***

Fig. 1. Average ratings of positive emotions during the evaluated user experiences (1 = very slightly or not at all – 5 = extremely; error bars denote standard errors of the means).

5.2. Quantitative results

5.2.1. Overview An overview of the quantitative results is presented in Table 3

below. For a detailed description of the methods and scales used, see Table 2 above.

Pairwise statistical tests for the emotion ratings confirmed that the participants experienced more happiness (higher emotional valence) Z = 7.5, p < .001, more positive emotions Z = 7.2, p < .001, and less negative emotions Z = 7.1, p < .001 in the context of most influential user experiences related to successful technology adop- tions, when compared to unsuccessful technology adoptions. For arousal, there was no statistically significant difference between experiences related to successful and unsuccessful technology adoptions. Overall, the participants also reported more need fulfill- ment Z = 7.0, p < .001 and value concordance Z = 6.8, p < .001 in the

context of successful technology adoptions, when compared to unsuccessful technology adoptions.

5.2.2. Experienced emotions Average ratings for the ten positive and ten negative PANAS

emotions are presented in Figs. 1 and 2 for most influential expe- riences related to both successful and unsuccessful technology adoptions. Statistical testing confirmed that there were significant differences between ratings of the ten positive emotions for both successful technology adoption vF

2 = 231.7, p < .001 and unsuccess- ful technology adoption values vF

2 = 91.5, p < .001. Similarly, varia- tions in the ratings of negative emotions differed significantly for both successful technology adoption vF

2 = 115.9, p < .001 and unsuccessful technology adoption values vF

2 = 220.1, p < .001. When the participants’ ratings for the positive emotions (Fig. 1)

were compared pairwise between successful and unsuccessful technology adoption experiences, it was found that for nine out of ten emotions, the participants gave higher ratings in the context of successful technology adoption (Interested Z = 6.8, p < .001; Excited Z = 6.2, p < .001; Strong Z = 4.1, p < .001; Enthusiastic Z = 7.0, p < .001; Proud Z = 6.0, p < .001; Inspired Z = 5.8, p < .001; Determined Z = 3.2, p < .01; Attentive Z = 4.7, p < .001; Active Z = 5.9, p < .001). For Alert, there was no statistically significant difference.

The ratings of negative emotions (Fig. 2) were also significantly higher in the reported unsuccessful adoption experiences, when compared to corresponding ratings for successful experiences, in the case of nine of the ten emotions (Distressed Z = 5.3, p < .001;

1

2

3

4

5

Successful technology adop�on Unsuccessful technology adop�on

*** *** ***

** ***

* ***

* *

Fig. 2. Average ratings of negative emotions during the evaluated user experiences (1 = very slightly or not at all – 5 = extremely; error bars denote standard errors of the means).

1

2

3

4

5

6

7

Successful technology adop�on Unsuccessful technology adop�on

** **

*** ***

Fig. 4. Average ratings for the NASA Task Load Index questions (1 = very low – 7 = very high, except for performance: 1 = success, 7 = failure; error bars denote standard errors of the means).

388 T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395

Upset Z = 7.2, p < .001; Guilty Z = 3.1, p < .05; Hostile Z = 5.5, p < .001; Irritable Z = 6.8, p < .001; Ashamed Z = 3.7, p < .01; Nervous Z = 3.0, p < .05; Jittery Z = 3.1, p < .05; Afraid Z = 4.6, p < .001). For Scared, there was no statistically significant differ- ence. Significant pairwise differences between the most influential user experiences in successful and unsuccessful technology adop- tions are also marked to all figures (asterisks denote significance levels: ⁄p < .05; ⁄⁄p < .01; ⁄⁄⁄p < .001).

5.2.3. Technology acceptance and perceived enjoyment Average ratings for the technology acceptance model variables,

as well as enjoyment, are presented in Fig. 3 for both successful and unsuccessful technology adoption experiences. The statistical analyses showed significant variation among the ratings given for the four technology adoption model variables (usefulness, ease of use, output quality, and intention to use) both in the context of successful vF

2 = 21.3, p < .001 and unsuccessful vF 2 = 8.3, p < .05

technology adoption experiences. As expected, pairwise statistical tests for the technology accep-

tance model related ratings showed that in the context of success- ful technology adoption experiences, the participants gave significantly higher ratings in terms of usefulness Z = 7.4, p < .001, ease of use Z = 7.2, p < .001, output quality Z = 7.4, p < .001, and intention to use Z = 7.0, p < .001, when compared to unsuccessful technology adoption experiences. Perceived enjoy- ment was also rated higher in the context of successful technology adoption experiences Z = 7.0, p < .001.

5.2.4. Task load Average ratings for the NASA Task Load Index variables are pre-

sented in Fig. 4 for both successful and unsuccessful technology

1

2

3

4

5

6

7

Perceived usefulness

Perceived ease of use

Output quality Inten�on to use Enjoyment

Successful technology adop�on Unsuccessful technology adop�on

*** *** *** ***

***

Fig. 3. Average ratings for the main Technology acceptance model variables and perceived enjoyment (1 = strongly disagree – 7 = strongly agree; error bars denote standard errors of the means).

adoption experiences. The statistical analyses showed significant variation among the NASA TLX ratings both in the context of suc- cessful vF

2 = 180.0, p < .001 and unsuccessful vF 2 = 95.6, p < .001

technology adoption experiences. Overall task load scores (see Table 3) were rated as significantly

higher in most influential experiences of unsuccessful (vs. success- ful) technology adoption Z = 3.3, p < .01. Pairwise statistical tests comparing successful and unsuccessful adoption experiences for each of the task load dimensions showed that the participants reported more mental demand Z = 3.9, p < .01, effort Z = 3.7, p < .01, and frustration Z = 7.3, p < .001 in the context of unsuccess- ful technology adoption experiences. As expected, they also reported more successful performance in the context of successful technology adoption experiences Z = 7.2, p < .001. The differences in reported physical demand and temporal demand were not sta- tistically significant between successful and unsuccessful technol- ogy adoptions.

5.2.5. Psychological needs Fig. 5 illustrates the average ratings for the fulfillment of the ten

psychological needs for both successful and unsuccessful technol- ogy adoption experiences. The statistical analyses showed signifi- cant differences among the ratings for the ten psychological needs both in the context of successful vF

2 = 197.3, p < .001 and unsuccessful vF

2 = 34.3, p < .001 technology adoption experiences. Pairwise comparisons between successful and unsuccessful

technology adoption experiences showed that need fulfillment rat- ings for seven of the ten needs were higher in the context of suc- cessful adoption experiences (Autonomy Z = 7.4, p < .001;

1 2

3

4 5

6

7 Successful technology adop�on Unsuccessful technology adop�on

*** *** ***

*** *** ** ***

Fig. 5. Average ratings for the statements probing the fulfillment of psychological needs (1 = strongly disagree – 7 = strongly agree; error bars denote standard errors of the means).

1

2

3

4

5

6

7

Successful technology adop�on Unsuccessful technology adop�on

*** ***

*** ***

*** *** **

**

*

Fig. 6. Average ratings for the statements probing concordance with personal values (1 = strongly disagree – 7 = strongly agree; error bars denote standard errors of the means).

T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395 389

Competence Z = 6.5, p < .001; Relatedness Z = 5.5, p < .001; Self-esteem Z = 5.3, p < .001; Self-actualization Z = 5.7, p < .001; Pleasure-stimulation Z = 3.9, p < .01; Money-luxury Z = 5.3, p < .001). For physical thriving, popularity-influence, and security, no statistical differences were found.

5.2.6. User values Fig. 6 illustrates the average ratings for the concordance with

the ten universal values for both successful and unsuccessful tech- nology adoption experiences. The statistical analyses showed sig- nificant differences among the ratings for the ten universal values both in the context of successful vF

2 = 132.6, p < .001 and unsuccessful vF

2 = 64.4, p < .001 technology adoption experiences. Pairwise comparisons showed that the participants reported

higher concordance with nine of the ten values in the context of experiences related to successful technology adoption, when compared to unsuccessful adoption experiences (Achievement Z = 5.3, p < .001; Power Z = 4.2, p < .001; Hedonism Z = 7.1, p < .001; Stimulation Z = 6.7, p < .001; Self-direction Z = 5.0, p < .001; Universalism Z = 3.2, p < .05; Benevolence Z = 4.3, p < .001; Tradition Z = 3.8, p < .01; Conformity Z = 3.9, p < .01). For the security value, no statistically significant difference was found.

5.2.7. Well-being impact The average score for flourishing (questions 1–8 in the

well-being question block) was 4.2 for the most influential user experiences in the context of a successful technology adoption

Table 4 Correlations (r) between the main variables. Lower left diagonal: most influential expe experiences in unsuccessful technology adoptions.

Variable 1 2 3 4 5 6

1. Valence 1 �.33** .26* �.42*** .50*** .39*** 2. Arousal .20 1 .15 .28* �.21 �.31** 3. Positive emotions .28* .13 1 �.03 .30** .10 4. Negative emotions �.49*** .05 .02 1 �.10 �.33** 5. Usefulness .35** �.05 .24* �.15 1 .41*** 6. Ease of use .36** .04 .06 �.15 .10 1 7. Output quality .31** �.01 .10 �.05 .09 .40*** 8. Intention to use .37** .08 .14 �.29* .16 .26* 9. Task load �.20 .00 .08 .24* .04 �.34** 10. Enjoyment .30** .27* .27* �.03 .02 .15 11. Flourishing .10 �.08 .24* .05 .25* �.04 12. Depression �.26* .08 �.09 .30** �.14 �.10 13. Need fulfillment .20 .08 .27* .16 .20 .12 14. Value concordance .06 �.04 .18 .06 .20 .05

* Correlation significant at p < .05. ** Correlation significant at p < .01.

*** Correlation significant at p < .001.

and 3.2 in the context of an unsuccessful adoption. The difference was statistically significant Z = 5.8, p < .001. The average depression score (questions 9–16 in the well-being question block) was 2.4 for the most influential user experiences in the context of a successful technology adoption and 3.7 in the context of an unsuc- cessful adoption. This difference was also statistically significant Z = 6.5, p < .001.

5.2.8. Reliability analysis Cronbach’s a scores were calculated for the all the constructs,

which were measured using multiple scales. In the context of user experiences related to successful technology adoption, the reliabil- ity scores for the technology acceptance model variables were as follows: Usefulness a = .81, Ease of use a = .82, Output quality a = .81, Intention to use a = .76. The corresponding scores in the context of unsuccessful technology adoption were: Usefulness a = .96, Ease of use a = .84, Output quality a = .91, Intention to use a = .94. The reliability score for the scale used for measuring enjoyment was for successful technology adoption experiences a = .86 and for unsuccessful technology adoption experiences a = .82.

For the self-determination theory related needs, Cronbach’s a scores were as follows in the context of successful technology adoption experiences: Autonomy a = .71, Competence a = .80, and Relatedness a = .94. The corresponding scores in the context of unsuccessful technology adoption were: Autonomy a = .78, Competence a = .77, and Relatedness a = .93. Finally, in the context of successful technology adoption, the reliability scores were a = .93 for the Flourishing Scale, and a = .77 for the modified Depression Scale. In the context of unsuccessful technology adop- tion the scores were a = .90 for flourishing and a = .70 for depres- sion. Thus, all of the reliability scores calculated were above or equal to .7, indicating satisfactory internal reliability. These scores were also compared to the reliabilities reported in the original publications (see Table 2 for references), or in the case of psycho- logical needs, for which alphas were not reported in the original publication, to user experiences studies applying the methods (Hassenzahl et al., 2010; Partala, 2011; Partala & Kallinen, 2012). In the current study, the alphas were on the same level than in these publications, except for depression, in which the original CES-D scale consisting of 16 items had higher internal consistency (a = .84 � .90; Radloff, 1977). The alpha for intention to use in the context of successful adoptions was also slightly lower in the cur- rent study (.76) than in the TAM literature (a = .82 � .97; Venkatesh & Davis, 2000).

riences in successful technology adoptions. Upper right diagonal: most influential

7 8 9 10 11 12 13 14

.61*** �.06 �.11 .18 .16 �.33** .32** .30* �.19 .04 .15 .04 .08 .21 .04 �.06 .28* .23 �.03 .40*** .30** �.18 .28* .33** �.20* .12 .32** �.13 �.06 .58*** �.08 �.09 .54*** .18 �.01 .31** .46*** �.16 .54*** .60*** .46*** .00 �.37** .28* .31** �.43*** .40*** .43***

1 �.05 �.02 .26* .28* �.43*** .45*** .45*** .43*** 1 .14 .23* .18 .18 .11 .20 �.28* �.22 1 �.10 �.01 .32** �.05 .00 .32** .24* �.08 1 .15 �.24* .20 .26* .03 �.02 .16 �.08 1 �.13 .78*** .82*** �.12 �.25* .26* .14 .00 1 �.20 �.13 .14 �.05 .20 �.16 .64*** .23 1 .85*** .10 .06 .06 �.13 .75*** .00 .62*** 1

Table 5 Summary of hierarchical multiple regression analyses predicting valence of user experience and flourishing in the context of most influential user experiences of successful technology adoptions after controlling for age, gender, and technology skills. For the other dependent variables (arousal and depression), the F values did not become significant.

Variable Valence of user experience Flourishing

B Std. error b B Std. error b

Positive emotions .283 .196 .129 .120 .151 .066 Negative emotions �1.091 .227 �.439*** �.112 .175 �.054 Usefulness .335 .124 .245** .082 .095 .073 Ease of use .146 .110 .128 �.095 .085 �.101 Output quality .046 .135 .035 �.007 .104 �.007 Intention to use .069 .131 .053 �.010 .101 �.009 Task load �.017 .136 �.012 .016 .104 .014 Enjoyment �208 .113 .167 .023 .087 .022 Need fulfillment .383 .158 .274* .311 .122 .269*

Value concordance �.221 .140 �.167 .619 .107 .567*** DR2 .469 .614 F (13, 62) 6.85*** 8.62***

* Significant at p < .05. ** Significant at p < .01.

*** Significant at p < .001.

390 T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395

5.2.9. Regression analyses Pearson correlations between the main variables were first cal-

culated to aid in the interpretation of the regression analyses. They are presented in Table 4 below with significance levels denoted by asterisks.

In order to identify significant relationships among the studied variables, hierarchical regression analyses were performed. The results are presented in Tables 5 and 6 below for most influential experiences of successful and unsuccessful technology adoptions, respectively. These results showed that the overall valence of the user experience was significantly affected by perceived usefulness, the fulfillment of psychological needs, and the PANAS score for negative emotions in the context of successful technology adop- tions. Somewhat similarly, overall user experience valence was sig- nificantly affected by perceived usefulness, output quality, and the PANAS score for negative emotion in the context of unsuccessful adoptions. Depression was significantly affected by negative emo- tions and output quality in the context of unsuccessful technology adoptions (Table 6), but not in the context of successful technology adoptions. In both most influential user experiences of successful and unsuccessful technology adoptions, flourishing (positive well-being) was largely determined by need fulfillment and value concordance. In both analyses more than half of the variance in flourishing could be explained by need fulfillment and value

Table 6 Summary of hierarchical multiple regression analyses predicting the four dependent variab of most influential user experiences of unsuccessful technology adoptions after controlling for

Variable Valence of user experience Arousal of user experienc

B Std. error b B Std. error b

Positive emotions .331 .225 .146 .633 .401 Negative emotions �.637 .201 �.302** .519 .358 Usefulness .323 .117 .324** �.391 .209 � Ease of use �.020 .125 �.019 �.249 .223 � Output quality .447 .123 .421** �.122 .220 � Intention to use �.017 .076 �.020 �.063 .136 � Task load �.101 .180 �.059 .367 .321 Enjoyment .121 .127 .095 �.212 .227 � Need fulfillment .098 .256 .066 .804 .457 Value concordance �.263 .242 �.196 �.247 .432 � DR2 .513 .230 F (13, 62) 6.14*** 2.90**

* Significant at p < .05. ** Significant at p < .01.

*** Significant at p < .001.

concordance together. For predicting arousal, no significant results emerged in this analysis, even though the regression model was significant for most influential experiences in unsuccessful tech- nology adoptions.

6. Discussion

Overall, the results of the current research highlighted the importance of user experience related variables focusing on the user’s felt experience of technology use in the most influential technology adoptions. In the current study, we asked our partici- pants to rate most influential user experiences related to successful and unsuccessful technology adoptions. The participants were not directed in any way regarding, which aspects (e.g. pragmatic vs. hedonic aspects) of user experience they should consider before deciding which experiences to report. Nevertheless, the results showed that the reported experiences were highly emotional: the overall emotional valence of user experience was given very high ratings in the context of the most influential experiences of successful adoptions and very low ratings in the most influential experiences of unsuccessful adoptions. The results on specific emo- tions further highlighted the dominant role of emotions in the reported user experiences. These results are in line with the holis- tic notion of user experience Hassenzahl (2010), according to which ‘‘feelings are integral to experiences (maybe even its core)’’.

The next step was to understand the antecedents of these pos- itive or negative experiences, which were critical to the success of the reported technology adoptions. Contrary to our hypotheses presented in chapter three, only a few significant predictors emerged in the multiple regression analysis explaining valence of user experience, flourishing, and depression. Of the TAM variables, perceived usefulness and output quality were most directly related to experienced emotions. Perceived usefulness had a significant relationship with experienced valence in both successful and unsuccessful technology adoptions, which suggests that it may have a central role in predicting overall user experience evalua- tions across the full range of felt user experiences. Higher output quality also predicted higher valence and it was inversely related with depressive symptoms in the most influential experiences related to unsuccessful adoptions.

The results on specific emotions suggest that users may form their overall valence evaluations of user experience valence more closely based on the presence of negative emotions than positive emotions, at least when analyzed using the model of ten positive and ten negative emotions in the research instrument used in the current research (PANAS). This was indicated by the finding

les (valence and arousal of user experience, flourishing, and depression) in the context age, gender, and technology skills.

e Flourishing Depression

B Std. error b B Std. error b

.187 .101 .136 .057 �.158 .146 �.109

.165 .037 .122 .023 .535 .130 .397***

.262 .021 .071 .027 .054 .076 .085

.159 �.009 .076 �.011 �.108 .081 �.162

.077 �.102 .075 �.123 �.211 .080 �.310*

.052 .013 .046 .021 .053 .049 .100

.143 �.019 .109 �.015 .077 .117 .070

.111 .059 .077 .059 .082 .083 .100

.361 .409 .155 .352* �.228 .167 �.239

.123 .633 .147 .605*** .214 .158 .249 .717 .521

13.30*** 5.82***

T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395 391

that the PANAS negative emotions score was a significant predictor of experienced overall valence in the context of most influential experiences of both successful and unsuccessful technology adop- tions. In the context of successful adoptions, need fulfillment was also a significant predictor of overall valence, which is in line with the results by Hassenzahl et al. (2010), who also found a clear rela- tionship between need fulfillment and positive affect in their study on positive experiences with interactive products. In order to achieve positive user experiences, the perceived pragmatic value of a product or system has to be high, but the product or system has to be also fulfilling in respect to hedonic aspects such as psy- chological needs. Psychological needs also seem to be more rele- vant in the context of positive than negative experiences, which are instead more directly caused by pragmatic aspects instead of hedonic or psychological aspects. A similar observation has been made, for example, by Partala and Kallinen (2012) in the context of participants’ most satisfying and most unsatisfying recent user experiences. This might also explain why output quality was in the current study inversely related to valence and depression only in the context of unsuccessful adoptions: outcome, which is con- sidered inappropriate by the user, may be alone enough to lead to an unsuccessful adoption, especially if it is associated with a negative emotional response from the user.

The current results also suggested that the concepts of need ful- fillment and value concordance are related to positive long-time well-being (flourishing) in the most influential experiences of both successful and unsuccessful technology adoptions. These results suggest that in order to design for technologies, which can bring everyday value to people and improve the quality of life (cf. life-based design), psychological needs and especially values are of central importance. On the other hand, negative well-being (depres- sion) seems to be related to negative emotional responses caused by unusable system outcomes in the context of unsuccessful adoptions. It is possible that the failure to adopt a system with desired out- comes itself affected the reported depression ratings in the context of the reported unsuccessful adoptions in the current study.

Overall, the descriptive results paint a picture, in which most influential experiences of successful technology adoptions are per- ceived as hedonically stimulating, enjoyable, and highly fulfilling of the most important psychological needs such as autonomy. The experiences are also accompanied by high levels of motivation-related positive emotions such as interest and enthusi- asm, and almost complete lack of experienced negative emotions. They are also associated with pragmatic concepts such as usefulness and output quality, as well as ease of use and lack of effort. In con- trast, the most influential experiences of unsuccessful technology adoptions are short on perceived pragmatic aspects such as useful- ness and output quality. This also leads to pronounced levels of neg- ative emotions such as being irritable, upset, or frustrated.

When the results for the successful and unsuccessful adoptions were compared to each other, statistically significant pairwise dif- ferences were found for most of the studied variables. These find- ings suggest that overall the current selection of variables was very relevant for the current study. Of the variables that did not differ significantly between the two conditions, physical demand and temporal demand seem to be more contextual than experiential variables and the equal fulfillment of the security need may also be due to similar use contexts in this respect. There were also no differences in experienced emotional arousal, which is in line with the common notion that both highly positive and negative experi- ences are often associated with more than average arousal (see e.g. Bradley & Lang, 1994). The participants may have also associated their ratings of the specific emotion ‘‘alert’’ (also no significant dif- ference) with their general level of arousal. The insignificant differ- ences between the ratings for both the security need and the security value in successful and unsuccessful adoptions suggest

that security-related feelings did not play a large role at in the cur- rent results.

The results of the current study are well in line with previous research with one clear exception. In the majority of previous stud- ies and also in the TAM3 model, the emotion-related concepts (most typically enjoyment and anxiety) are directly related to per- ceived ease of use, but not to perceived usefulness. In contrast, the current result showed that perceived usefulness (and also output quality in unsuccessful adoptions) was significantly related to emotional valence of the reported user experiences. These results suggest that in fact, there are usefulness (and output quality) related emotions, which may become important mainly in the con- text of single (decisive) user experiences, possibly as evaluative feelings contributing to technology acceptance or rejection.

The other results are much more consistent with previous research. As one would expect, all the TAM related variables – per- ceived ease of use, perceived usefulness, intention to use, output quality, and perceived enjoyment – were all significantly higher in the context of most influential experiences of successful than unsuccessful technology adoption. The mean ratings for these cen- tral TAM variables were about on the same level compared to each other, as in many existing TAM studies (see e.g. Hu, Clark, & Ma, 2003; Park & Chen, 2007), and the mean ratings in the context of most influential experiences of successful and unsuccessful adop- tions were higher and lower, respectively, than in average TAM studies, in which mean ratings for the TAM variables typically set- tle around 4 or 5 on the 1–7 scale. The emotion ratings of both dimensional and specific emotions also followed typical patterns obtained in previous literature using those methods (e.g. Bradley & Lang, 1994; Partala & Kallinen, 2012).

The results on psychological needs are also mostly in line with previous research (e.g. Hassenzahl et al., 2010; Partala & Kallinen, 2012), as the self-determination theory related needs of autonomy, competence, and relatedness, as well as self-esteem were rated among the most salient needs. On the other hand, however, the ful- fillment of self-actualization-meaning and money-luxury needs were rated as more salient in the current study compared to previ- ous research, especially in the context of most influential user experiences related to successful technology adoptions. These results may be due to the current emphasis on the acquisition of personal technologies in the descriptions of the participants. Similarly, the results on values are in line with previous research, (e.g. Partala & Kujala, in press), as hedonism, self-direction, achievement, and tradition were rated among the most influential user values. In contrast, stimulation and conformity were also among the most salient values in the current study. This might be because of the large number of experiences reported on per- sonal technologies, which the participants found personally stimu- lating and not interfering with their social relationships. Security values were also not emphasized in the current results. Otherwise, the results on user values also give support to findings by Partala and Kujala (in press) that all ten universal values are potentially relevant in user experiences.

In previous research, flourishing has been shown to increase with positive emotions and has been linked to fulfilled psycholog- ical needs (e.g. Diener et al., 1999). The flourishing scale used in the study is based on theories of universal human needs, such as com- petence, relatedness and self-acceptance, so it would be expected that flourishing increases while key psychological needs are being filled (Ryan & Deci, 2000). This was confirmed in the current data in both successful and unsuccessful technology adoption contexts, and the results also suggested that higher reported value concor- dance increased flourishing in both contexts, which is also in line with notions from the field of psychology (e.g. Sheldon et al., 2004). However, in the current data, the relationship between pos- itive emotions and flourishing did not become statistically

392 T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395

significant. Depression is increased by negative emotions and has been shown to decrease when psychological needs are filled (e.g. Meifen, Schaffer, Young, & Zkalik, 2005). The current findings also suggest that negative emotions affect depression in the context of user experiences, but this relationship was statistically significant only in the context of unsuccessful technology adoptions, in which the participants reported much more negative emotions. Again, this relationship is in line with prior findings from general psychol- ogy (e.g. Gentile et al., 2009). In contrast, the relationship between need fulfillment and depression was not statistically significant in the current research design focusing on most influential user expe- riences, even though a line psychological research suggests that level of fulfillment of psychological needs underlies both positive and negative mental health (e.g. Ryan & Deci, 2000).

When interpreting the current results, one has to keep in mind that the current research was carried out using a basic research approach, in which a phenomenon is studied through its extremi- ties, in this case, the most influential experiences in successful and unsuccessful technology adoptions. It is reasonable to believe that the current approach was highly sensitive in identifying the most important variables, which have practical value on the user’s deci- sion on technology adoption or non-adoption. This was an impor- tant goal of the current research. On the other hand, the limitation of this approach is that the data does not represent the full range of all situations and user experiences, for example, typical technology use situations, which often are quite neutral in many respects. The current results also did not warrant a fully successful structural modeling or path analysis by taking the central TAM concepts as the basis and structuring the UX variables around these concepts, and this is why the analysis presented is limited to studying direct effects between the selected independent and dependent variables. Overall, the correlations between the TAM concepts were also lower than in previous studies. This was probably because the users reported highly positive and negative user experiences and also gave the TAM evaluations in the context of those experiences instead of the general impression of the system like in most TAM studies. It may be that the TAM model loses some of its explana- tory power in the context of evaluating single highly influential positive and negative user experiences, in which hedonic variables become more pronounced. However, this notion should be con- firmed in further research. In more neutral settings, user experi- ence variables (e.g. the variables significantly related to TAM concepts in this study) might also be successfully augmented to the entire TAM model (e.g. TAM3). Further research is also needed to study, for example, to what extent the current regression based findings would hold in more neutral everyday user experiences of technology use. In addition, future research could aim at more detailed understanding of the qualitative contents of central user experience variables such as emotions, psychological needs and values, and also of behavioral patterns, which lead to negative user experiences and unsuccessful adoptions.

We aimed at including all the central user experience variables in the current research to understand the nature of most influential user experiences in technology adoptions in detail. In fact, the cur- rent study is distinguished from most user experience studies by its aim of understanding user experience holistically. However, one might argue there are still other factors influencing user expe- riences, which could have been added to this research. For exam- ple, it is known that user experiences are greatly affected by the purpose of use, as well as the use context more generally. Within the resources of the current study, it was only possible to take con- text into account briefly through some of the statements to study task load (see Appendix A) and also by classifying the qualitative descriptions to voluntary vs. more compulsory contexts of technol- ogy adoption. The current study also did not concentrate on cul- tural factors affecting technology use, which have been studied

in previous user experience studies mostly using qualitative methods.

Another limitation of the current study is that the current par- ticipants were undergraduate and graduate university students, partly from different cultures, which may mean that some partici- pants understood some of the statements somewhat differently than others. There were also more male than female participants. These are limitations in the generalizability of the results to the wider population, which typically has less education and familiar- ity with technology acceptance and user experience related issues. However, according to a meta-analysis, there is previous evidence of ‘‘the value of using students as surrogates for professionals in some technology adoption studies, and perhaps more generally’’ (King & He, 2006). Indeed, it is reasonable to believe that the cur- rent sample understood the different scales and assignments well and the participants were also sensitive to the scales used in the current study, which might have enhanced the validity of the results. The experiences described were unique for each partici- pant and the anonymity of the participants was protected, which gives reason to believe that a valid enough data set could be obtained using the current methods.

From the practical perspective, the most important implication of the current research seems to that besides traditional attributes (e.g. usefulness, ease of use, output quality), designers have to pay attention to the users’ emotional responses to products in order to promote successful technology adoptions. In the current study, the objects of technology adoptions were mostly personal technolo- gies, but they were not especially hedonic in nature. Still, the par- ticipant reported highly emotional positive or negative experiences as the most influential user experiences in technology adoptions. From this perspective, it seems that the designers of technology products may be overlooking the current knowledge on emotional design (e.g. Norman, 2004). The results also give support to the important role of the concept of psychological needs (Hassenzahl, 2010; Sheldon et al., 2001; Deci & Ryan, 2014) especially in design- ing for positive experiences, which can in turn contribute to suc- cessful technology adoptions. Systematic knowledge how psychological needs should be considered in design is, however, still emerging and it is mostly up to individual designers to utilize their insight in this respect. Finally, designers and technology man- agers should pay attention to the wider context of the technology use, including also activities more generally. The technologies and activities have to be in line with the users’ values (cf. value-sensitive design) and they also have to fit into the use’s life so that the user thinks that any changes contribute positively to overall well-being (flourishing; cf. life-based design).

In all, the results add to the current knowledge on the exact nat- ure of most influential user experiences of successful and unsuc- cessful technology adoptions from a psychological perspective. The results also highlighted relationships between user experience and technology acceptance concepts, which have not been system- atically explored in previous studies. The results suggest that con- cepts and variables from the user experience research field complement the existing TAM methods in a beneficial way, espe- cially in understanding the complexity of positive user experiences associated with successful technology adoptions. The concepts and methods currently available in the research field of user experience are mainly based on well-established theories and methods from psychological sciences, and using them in technology adoption studies is likely to result in a more comprehensive and holistic account of all the factors affecting usage behavior, as well as tech- nology adoption and non-adoption. The current results suggest that an integrated approach focusing on both the users’ percep- tions related to technology use and most influential felt user expe- riences would be an improvement in broadening the scope of current technology adoption research.

T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395 393

Acknowledgements

The authors would like to thank all the participants for their thoughtful responses. This research was supported by Finnish Work Environment Fund (grant 113103) and Tampere University of Technology.

Appendix A. Statements and scales used in the study

See Section 2.3 for explanation of the origin of the scales.

A.1. Emotions

Dimensional emotions:

1. Unhappiness – Happiness

2. Calmness – Arousal

Specific emotions:

1. Interested

11. Irritable

2. Distressed

12. Alert

3. Excited

13. Ashamed

4. Upset

14. Inspired

5. Strong

15. Nervous

6. Guilty

16. Determined

7. Scared

17. Attentive

8. Hostile

18. Jittery

9. Enthusiastic

19. Active

10. Proud

20. Afraid

Scale:

Very Slightly or Not at All

A Little

Moderately

Quite a Bit

Extremely

1

2

3

4

5

A.2. Technology acceptance model

1. Using the system improved my performance. 2. Using the system increased my productivity. 3. Using the system enhanced my effectiveness. 4. I found the system to be useful. 5. My interaction with the system was clear and

understandable. 6. Interacting with the system did not require a lot of my men-

tal effort.

7. I found the system to be easy to use. 8. I found it easy to get the system to do what I want it to do. 9. The quality of the output I got from the system is high.

10. I had no problem with the quality of the system’s output. 11. I rated the results from the system to be excellent. 12. Assuming I had access to the system, I intended to use it. 13. Given that I had access to the system, I predicted that I

would use it. 14. I planned to use the system in the next few months.

Scale:

Strongly disagree agree

Neither agree nor disagree

Strongly

1

2

3

4

5

6

7

A.3. Task load

1. How mentally demanding was your use of this system? 2. How physically demanding was your use of this system? 3. How hurried was your use of this system? 4. How hard did you have to work to accomplish your level of per-

formance when using this system? 5. How insecure, discouraged, irritated, stressed, and annoyed

were you when using this system? 6. How successful were you in accomplishing what you were try-

ing to do when using this system?

Scale in questions 1–5:

Very low

Very high

1

2

3

4

5

6

7

Scale in question 6:

Failure

Success

1

2

3

4

5

6

7

A.4. Enjoyment

When using this system, I thought that the use was:

Enjoyable

1

2

3

4

5

6

7

Disgusting

Exciting

1

2

3

4

5

6

7

Dull

Pleasant

1

2

3

4

5

6

7

Unpleasant

Interesting

1

2

3

4

5

6

7

Boring

394 T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395

A.5. User needs

1. Using this system made me feel that my choices were based on my true interests and values.

2. Using this system made me feel free to do things my own way.

3. Using this system made me feel that my choices expressed my ‘‘true self’’.

4. Using this system made me feel that I was successfully com- pleting difficult tasks and projects.

5. Using this system made me feel that I was taking on and mastering hard challenges.

6. Using this system made me feel very capable in what I did. 7. Using this system made me feel a sense of contact with peo-

ple who care for me, and whom I care for. 8. Using this system made me feel close and connected with

other people who are important to me. 9. Using this system made me feel a strong sense of intimacy

with the people I spent time with. 10. Using this system made me feel a sense of deeper purpose in

life. 11. Using this system made me feel a strong sense of physical

well-being. 12. Using this system made me feel that I was experiencing new

sensations and activities. 13. Using this system made me feel that I had nice things and

possessions. 14. Using this system made me feel safe from threats and

uncertainties. 15. Using this system made me feel quite satisfied with who I

am. 16. Using this system made me feel that I strongly influenced

others’ beliefs and behavior.

Scale:

Strongly disagree

Neither agree nor disagree

Strongly agree

1

2

3

4

5

6

7

A.6. User values

1. Using this system made me feel that I was an ambitious and successful individual.

2. Using this system made me feel that I had a high social sta- tus in my community.

3. Using this system made me feel that my use was pleasurable.

4. Using this system made me feel that I was getting exciting stimulation.

5. Using this system made me feel that I was a creative and independent individual.

6. Using this system made me feel that my actions were well in line with global values such as social justice and protecting the environment.

7. Using this system made me feel that I could help other peo- ple near me.

8. Using this system made me feel that my actions were well in line with traditions I consider as important.

9. Using this system made me feel that I was not causing any harm or disorder in my community.

10. Using this system made me feel that I could advance the safety and stability of myself and my environment.

Scale:

Strongly disagree

Neither agree nor disagree

Strongly agree

1

2

3

4

5

6

7

A.7. Psychological well-being

1. Using this system made me feel I lead a purposeful and meaningful life.

2. Using this system made me feel that my social relationships are supportive and rewarding.

3. Using this system made me feel that I am engaged and inter- ested in my daily activities.

4. Using this system made me feel that I actively contribute to the happiness and well-being of others.

5. Using this system made me feel that I am competent and capable in the activities that are important to me

6. Using this system made me feel that I am a good person and live a good life.

7. Using this system made me feel optimistic about my future. 8. Using this system made me feel that people respect me. 9. Using this system made me feel that I was bothered by

things that usually don’t bother me. 10. Using this system made me feel that I could not ‘‘get going’’. 11. Using this system made me feel sad. 12. Using this system made me feel that I was just as good as

other people. 13. Using this system made me feel that I had trouble keeping

my mind on what I was doing. 14. Using this system made me feel that everything I did was an

effort. 15. Using this system made me feel depressed. 16. Using this system made me feel that my sleep was restless.

(statements 1–8 indicate flourishing; statements 9–16 indicate depression, statement 12 was reverse scored).

Scale:

Strongly disagree

Neither agree nor disagree

Strongly agree

1

2

3

4

5

6

7

References

Argyle, M., & Martin, M. (1991). The psychological causes of happiness. In F. Strack, M. Argyle, & N. Schwartz (Eds.). Subjective well-being. An interdisciplinary perspective (pp. 77–100). Oxford: Pergamon Press.

Bagozzi, R. P. (2007). The legacy of the technology acceptance model and a proposal for a paradigm shift. Journal of the Association for Information Systems, 8(4), 244–254.

Beaudry, A., & Pinsonneault, A. (2010). The other side of acceptance: Studying the direct and indirect effects of emotions on information technology use. MIS Quarterly, 34(4), 689–710.

Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The self-assessment manikin and the semantic differential. Journal of Behavioral Therapy and Experimental Psychiatry, 25, 49–59.

Carver, C. S., & Scheier, M. F. (1989). On the self-regulation of behavior. New York: Cambridge University Press.

Cenfetelli, R. T. (2004). Getting in touch with our feelings towards technology. In Proceedings of the academy of management conference 2004 (pp. F1–F6). Academy of Management.

Crawford, J. R., & Henry, J. D. (2004). The positive and negative affect schedule (PANAS): Construct validity, measurement properties and normative data in a large non-clinical sample. British Journal of Clinical Psychology, 43, 245–265.

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340.

T. Partala, T. Saari / Computers in Human Behavior 53 (2015) 381–395 395

Deci, E. L., & Ryan, R. M. (2014). The importance of universal psychological needs for understanding motivation in the workplace. In R. M. Ryan (Ed.), The oxford handbook of work engagement, motivation, and self-determination theory (pp. 13–32). New York: Oxford University Press.

Desmet, P., & Hassenzahl, M. (2012). Towards happiness: Possibility-driven design. In Human–computer interaction: The agency perspective (pp. 3–27). Berlin Heidelberg: Springer.

Desmet, P., & Hekkert, P. (2007). Framework of product experience. International Journal of Design, 1(1), 57–66.

Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125, 276–302.

Diener, E., Wirtz, D., Tov, W., Kim-Prieto, C., Choi, D., Oishi, S., et al. (2009). New measures of well-being: Flourishing and positive and negative feelings. Social Indicators Research, 39, 247–266.

Elliot, A. J., & Sheldon, K. M. (1998). Avoidance personality goals and the personality-illness relationship. Journal of Personality and Social Psychology, 75, 1282–1299.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.

Gentile, B., Grabe, S., Dolan-Pascoe, B., Twenge, J. M., Wells, B. E., & Maitino, A. (2009). Gender differences in domain-specific self-esteem: A meta-analysis. Review of General Psychology, 13(1), 34–45.

Hart, S. G., & Staveland, L. E. (1988). Development of NASA-TLX (Task Load Index): Results of empirical and theoretical research. Advances in Psychology, 52, 139–183.

Hassenzahl, M. (2010). Experience design: Technology for all the right reasons. Morgan and Claypool Publishers.

Hassenzahl, M., Diefenbach, S., & Göritz, A. (2010). Needs, affect, and interactive products – facets of user experience. Interacting with Computers, 22(5), 353–362.

Hollender, N., Hofmann, C., Deneke, M., & Schmitz, B. (2010). Integrating cognitive load theory and concepts of human–computer interaction. Computers in Human Behavior, 26(6), 1278–1288.

Hornbæk, K. (2006). Current practice in measuring usability: Challenges to usability studies and research. International Journal of Human–Computer Studies, 64(2), 79–102.

Hu, P. J. H., Clark, T. H., & Ma, W. W. (2003). Examining technology acceptance by school teachers: A longitudinal study. Information and Management, 41(2), 227–241.

Isomursu, M., Ervasti, M., Kinnula, M., & Isomursu, P. (2011). Understanding human values in adopting new technology—A case study and methodological discussion. International Journal of Human–Computer Studies, 69(4), 183–200.

Kamp, I., & Desmet, P. (2014). Measuring product happiness. In CHI’14 extended abstracts on human factors in computing systems (pp. 2509–2514). New York: ACM Press.

Keyes, C. L. M., & Waterman, M. B. (2003). Dimensions of well-being and mental health in adulthood. In M. H. Bornstein, L. Davidson, C. O. M. Keyes, & K. A. Moore (Eds.), Well-being. Positive development across the life course (pp. 477–497). Mahwah, NJ: Lawrence Erlbaum.

Kim, H. W., Chan, H. C., & Chan, Y. P. (2007). A balanced thinking–feelings model of information systems continuance. International Journal of Human–Computer Studies, 65(6), 511–525.

King, W. R., & He, J. (2006). A meta-analysis of the technology acceptance model. Information and Management, 43(6), 740–755.

Laverie, Debra A., Kleine, Robert E., & Kleine, Susan Schultz (1993). Linking emotions and values in consumption experiences: An exploratory study. Advances in Consumer Research, 20(1), 70–75.

Law, E. L. C., Roto, V., Hassenzahl, M., Vermeeren, A., & Kort, J. (2009). Understanding, scoping and defining user experience: A survey approach. In Proceedings of the 27th international conference on Human factors in computing systems, CHI 2009 (pp. 719–728). New York: ACM Press.

Legris, P., Ingham, J., & Collerette, P. (2003). Why do people use information technology? A critical review of the technology acceptance model. Information and Management, 40(3), 191–204.

Malhotra, Y., & Galletta, D. F. (1999). Extending the technology acceptance model to account for social influence: Theoretical bases and empirical validation. In Proceedings of the 32nd annual Hawaii international conference on system sciences (pp. 1006–1020). IEEE.

McCarthy, J., & Wright, P. (2004). Technology as experience. Cambridge MA: MIT Press.

Meifen, W., Schaffer, P. A., Young, S. K., & Zkalik, R. A. (2005). Adult attachment, shame, depression, and loneliness: The mediation role of basic psychological needs satisfaction. Journal of Counseling Psychology, 52(4), 591–601.

Norman, D. A. (2004). Emotional design: Why we love (or hate) everyday things. New York: Basic books.

Park, Y., & Chen, J. V. (2007). Acceptance and adoption of the innovative use of smartphone. Industrial Management and Data Systems, 107(9), 1349–1365.

Partala, T. (2011). Psychological needs and virtual worlds: Case Second Life. International Journal of Human–Computer Studies, 69(12), 787–800.

Partala, T. & Kujala, S. (in press). Exploring the role of ten universal values in using products and services. Interacting with Computers.

Partala, T., & Kallinen, A. (2012). Understanding the most satisfying and unsatisfying user experiences: Emotions, psychological needs, and context. Interacting with Computers, 24(1), 25–34.

Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.

Riva, G., Banos, R. M., Botella, C., Wiederhold, B. K., & Gaggioli, A. (2012). Positive technology: Using interactive technologies to promote positive functioning. Cyberpsychology, Behavior, and Social Networking, 15(2), 69–77.

Roca, J. C., & Gagné, M. (2008). Understanding e-learning continuance intention in the workplace: A self-determination theory perspective. Computers in Human Behavior, 24(4), 1585–1604.

Russell, J. A., & Carroll, J. M. (1999). On the bipolarity of positive and negative affect. Psychological Bulletin, 125(1), 3–30.

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68–78.

Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. Advances in Experimental Social Psychology, 25(1), 1–65.

Schwartz, S. H. (2006). Basic human values: Theory, measurement, and applications. Revue française de sociologie, 42, 249–288.

Sheldon, K. M., Elliot, A. J., Kim, Y., & Kasser, T. (2001). What is satisfying about satisfying events? Testing 10 candidate psychological needs. Journal of Personality and Social Psychology, 80, 325–339.

Sheldon, K. M., Elliot, A. J., Ryan, R. M., Chirkov, V., Kim, Y., Wu, C., et al. (2004). Self- concordance and subjective well-being in four cultures. Journal of Cross-Cultural Psychology, 35(2), 209–223.

Sørebø, Ø., Halvari, H., Gulli, V. F., & Kristiansen, R. (2009). The role of self- determination theory in explaining teachers’ motivation to continue to use e- learning technology. Computers and Education, 53(4), 1177–1187.

Srite, M., & Karahanna, E. (2006). The role of espoused national cultural values in technology acceptance. MIS Quarterly, 30(3), 679–704.

Straub, E. T. (2009). Understanding technology adoption: Theory and future directions for informal learning. Review of Educational Research, 79(2), 625–649.

Van der Heijden, H. (2004). User acceptance of hedonic information systems. MIS Quarterly, 695–704.

Venkatesh, V. (2000). Determinants of perceived ease of use: Integrating control, intrinsic motivation, and emotion into the technology acceptance model. Information Systems Research, 11(4), 342–365.

Venkatesh, V., & Bala, H. (2008). Technology acceptance model 3 and a research agenda on interventions. Decision Sciences, 39(2), 273–315.

Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technology acceptance model: Four longitudinal field studies. Management Science, 46(2), 186–204.

Watson, D., Tellegen, A., & Clark, L. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070.