15 Smartphone for social networking: methodological aspects

: In this chapter we parsed methodological aspects in studying behavior with mobile social networking sites like Facebook by using physiological measures. Current computational methods and the huge availability of sources and devices for psychophysiological recording allowed a deeper understanding of complex behaviors and an even more detailed knowledge of human emotions. These methods can be applied to understand the user navigation experience.


Introduction
Recent smart phones and mobile platforms integrate multiple advanced functions and allow individuals to share a lot of different information, such as photos and GPS data, within their social networks in real-time (Humphreys & Liao, 2011).Although a mobile device has by definition a smaller display (Ballagas, Borchers, Rohs, & Sheridan, 2006;Bulusu, Heidemann, & Estrin, 2000), which could create many problems during the experience (Chae & Kim, 2004;Codispoti & de Cesarei, 2007;De Cesarei & Codispoti, 2010;Lombard, Grabe, Reich, Campanella, & Ditton, 1996), they are easy-to-use and customizable: in particular, touch screen capabilities may improve the usability and navigation experience when browsing photos and information (Albinsson & Zhai, 2003;Chae & Kim, 2004;Vogel & Baudisch, 2007).Hence, it is easy to understand how mobile phones enable people to stay virtually connected at all times irrespective of their location.They assure "perpetual contact" (James, 2004;Katz & Aakhus, 2002) and affect every aspect of personal and professional life directly or indirectly.On the other hand, it is important to highlight that pathological experience may also arise from a pervasive use of mobile phones (Lepp, Barkley, & Karpinski, 2014;Salehan & Negahban, 2013).
Multifunctional mobile platforms allow individuals to be constantly in touch with their social network and to go beyond the main constraint of face-to-face communication, enacting the spatial-temporal contiguity (G.Riva, 2010).Moreover, mobile devices create an ubiquitous bridge between real and virtual life (Barkhuus & Polichar, 2011); they enhance the capability of social networking sites (SNSs) to create a new social space, the Interreality, marked by the fusion of real and virtual networks (Gaggioli et al., 2011;Repetto et al., in press).From a wider viewpoint, it is possible to assume that connections between millions of people via social networking sites constitute a new and powerful "social-self" where individuals can share experiences and have the "sense of being with another" (Biocca, Harms, & Burgoon, 2003), i.e., the Social Presence.The feeling of being with another person while using a social medium means, first of all, feeling part of the mediated communication (Biocca et al., 2003;Villani, Repetto, Cipresso, & Riva, in press), i.e., the Presence.According to Biocca and Colleagues (Biocca et al., 2003), presence is marked by two main aspects, telepresence, the sense of "being there" in a virtual space as the result of the possibility of executing automatic responses to spatial cues and creating the mental models of the mediated space, and social presence, the sense of "being with another" in a mediated space, including effective responses to social cues and mental models of other intentions.From this perspective, mobile platforms could be defined as social presence technologies (Biocca et al., 2003) specifically designed to increase social presence, since they assure ubiquitous (Ballagas et al., 2006;Repetto et al., in press;Villani et al., 2011) and perpetual (Katz & Aakhus, 2002) social contact via multimodal access (Brown, Green, & Harper, 2002).
As IJsselsteijn and Colleagues (IJsselsteijn, de Ridder, Freeman, & Avons, 2000;Ijsselsteijn, Freeman, & De Ridder, 2001) and Biocca and Colleagues (Biocca et al., 2003) underlined, there is not a unified model that assesses presence in mediated communication, although different approaches, such as, subjective self-report measures (Slater & Garau, 2007), continuous presence assessment (IJsselsteijn et al., 1997), analysis of postural and gestural responses (Freeman, Avons, Meddis, Pearson, & IJsselsteijn, 2000;Giakoumis et al., In press), psychophysical methods (Stanney et al., 1998), physiological indexes (Baumgartner, Valko, Esslen, & Jancke, 2006;Mikropoulos, 2001;Schilbach et al., 2006;Wiederhold et al., 2000), and many others (Schultze, 2010) have been used also with clinical patients (Albani et al., 2012;Raspelli et al., 2012;Repetto et al., in press).In this chapter, we were interested in examining subjects' experience during the use of SNSs; thus, our objective was to identify presence during such experience and the ways in which it leads to an optimum.A measure of presence in this study needs to consider the temporal continuity of the experience, additionally, for an objective and accurate assessment, we had to measure the key factors that previous research took into account for presence and optimal experience.Thus, we focused on three main aspects of time-space continuum of subjects' states: physiological arousal, emotional valence, and sustained attention.
The third dimension used to assess the SNSs experience was the attention.According to Draper and Colleagues (Draper, Kaber, & Usher, 1998), presence occurs when more attentional resources are allocated to the computer-mediated environment: "The more attentional resources that a user devotes to stimuli presented by the displays, the greater the identification with the computer-mediated environment and the stronger the sense of presence."If we are able to demonstrate that users are engaged during an experience and that during the same experience their level of attention (Schupp et al., 2007) to the source of the experience is high, then we can affirm that their sense of presence is high, which increases the optimal experience of subjects through higher involvement (Riley, Kaber, & Draper, 2004;Giuseppe Riva, Davide, & Ijsselsteijn, 2003).

Psychophysiology and Affective States
According to the classic valence-arousal model (Lang, 1995b;Russell, 1979) for experimentally identifying affective states in subjects, we can consider the two dimensions of "activation", namely, physiological arousal and emotional valence.
Physiological arousal can be measured by using electroencephalogram (EEG), galvanic skin response (GSR), cardiovascular activity (ECG or BVP), and respiration signal (RSP); emotional valence can be measured by using EEG, self-reports, facial expression identification, eye-blink startle, and facial EMG corrugator and/or zygomatic.According to Blumenthal and colleagues (Blumenthal et al., 2005), facial EMG-CS (corrugator) can be considered the best measure of emotion valence.
Frontal EEG activation asymmetry has been generally used, giving evidence that greater left frontal activity seems to be more related with positive emotional valence, whereas greater right frontal activity seems to be more involved in negative emotional valence (Debener, Beauducel, Fiehler, Rabe, & Brocke, 2001).Alpha index has been suggested to be the most adept for studying frontal EEG activation asymmetry (Debener et al., 2001).Alpha Asymmetry index can be calculated in a number of different ways in order to take into account the prevalence of one hemisphere over the other one, and correct the index accordingly.In calculating this index it is crucial to consider that higher cortical activation is revealed by lower Alpha waves, and thus this need to be considered in the computation and formula derivation.
As recently demonstrated by Mojzisich and colleagues (Mojzisch et al., 2006), selfinvolvement during social interactions is specifically related to attention allocation.For this reason we are also interested in investigating sustained attention as another dimension of analysis.To this ends, slow Alpha EEG bands (following Slow Alpha) have been demonstrated to be a valid measure of sustained attention (Klimesch, 1999;Klimesch, Doppelmayr, Russegger, Pachinger, & Schwaiger, 1998).

Hypotheses and Research Questions
Eight hypotheses can be formulated, four regarding navigating SNSs using a PC and four considering navigation using a mobile device (Table 1).Each hypothesis is based on one dimension of Arousal-Valence-Attention factors.We added the fourth dimension of anxiety to act as a control dimension.For this purpose, we used two basic affective states, namely "Relax" and "Stress," that can be easily elicited in healthy subjects, representing the ground truth in the tridimensional space (Cipresso, Gaggioli, et al., 2012;Magagnin et al., 2010;Mauri et al., 2010a;Villani et al., 2011).This operation is conducted to compare along each axis the SNSs condition with a standard affective states generated in the subject.The significant quadratic trend in the within-subjects contrasts can be used to further test each hypothesis.The Bayes factor can also be considered as alternative to the conventional t-test to express preference for either the null hypothesis or the alternative.
For Physiological Arousal, we hypothesized that subjects are more physiologically activated when navigating SNSs, which make them have an arousal quite similar to the one in the stress condition and different enough from the one during the relax condition (Hypotheses 1 and 5).
Positive emotions during the SNSs navigation were also proposed, in that the subjects were proposed to have a higher emotional valence during SNSs navigation, as high as in a Relax condition.On the other hand, low emotional valence due to the elicited negative emotions is expected during the Stress condition (Hypotheses 2 and 6).
For attention, it was easy to hypothesize that a higher level of attention characterizes both of the SNSs conditions (if an interest is provoked) and the stress condition, where the attentional resource is needed to focus on the complex task.On the other hand, the relax condition is expected to have a low level of attention, which is different from the SNSs condition (Hypotheses 3 and 7).
Anxiety level is an important dimension to be considered, since it allow us to correctly discern stressful experiences from optimal experiences.In this context, this dimension is crucial for avoiding erroneous conclusions about the effect of SNSs navigation.The hypothesis, therefore, is that the level of anxiety during SNSs navigation is quite different from the level of anxiety during the stress condition.In this sense, the level of anxiety during the SNSs navigation is supposed to be quite similar to the level during the relax condition (Hypotheses 4 and 8).
A synthesis of the hypotheses is reported in Table 1, where were coded the SNSs navigation for both sessions: during the PC navigation (shortly PC session), and during the navigation using a specific app in a mobile smart phone (shortly Mobile session).
The goal of studies which are interested in SNSs experience should be to investigate the effects of navigating SNSs on subjects' experience by measuring psychophysiological correlates.In particular the following research questions (RQ) arise: RQ1: Can SNSs navigation leads to an engagement state, aside from the specific platform used and the related levels of pervasiveness?RQ2: Can SNSs navigation through the mobile be more effective, leading to a higher pervasiveness able to induce an optimum state?
To answer these questions, we can use standard psychophysiological measures.We need to consider that the more the subjects will be engaged and attentive, the more likely they will be able to achieve an optimal state characterized by positive valence, high arousal, and high attention.Thus, we are interested in objectively identifying specific pattern of users' affective states in the Arousal-Valence-Attention space while experiencing SNSs navigation through a PC or a mobile platform.

Conclusion
We presented a multidimensional model to understand the experience of SNSs navigation by means of psychophysiological correlates.The dimensions considered were physiological arousal, emotional valence, and life satisfaction.Psychophysiological measures can be extremely useful for understanding users' behavior without interfering with their experience, and for this reason are considered to be very effective psychometric instruments.The idea of creating a map of the affective states during user navigation can be considered either in terms of simple observation or as an intervention device for any bad state arising during SNSs use.A real-time intervention to inform users of their affective states during SNSs navigation requires so-called affective computing.According to Rosalind Picard (1997), who coined the term, "affective computing is computing that relates to, arises from, or deliberately influences emotion or other affective phenomena."Of course affective computing is a fascinating field; however, it does have intrinsic limitations due to a lack of classification.It is in fact complex to recognize an affective state after months of signal processing and data analysis, let alone in real time.However, the big advantage of affective computing is that it has fostered the development of plenty of classification methods through vivid and productive international discussion at extremely high scientific levels.The continuous development of new artificial intelligence techniques further enriches this scenario; however, these methods have not often been applied to understanding SNSs navigation.
The spread of SNSs use among the population has been significant in the last ten years, and we can now affirm that most people in a way or another use social connections with direct consequences on their lives.To quantify these consequences is important not only at the social level but also at the individual level, since better information about our own states is crucial for personal decision-making.In particular, if I don't feel stressed during an SNSs navigation, but indeed I am, it would be better if an automatic system would be able to suggest I change my activity and try to relax for a while.Science fiction?Not at all.Automatic detection of physiological states and understanding of affective states is becoming a reality, and thousands of researchers are working to make this scenario possible.

Table 15 .1. Test
Hypotheses (~ is for similar, > is for greater than, < is for lower than)