Understanding Online Political Participation: Theory of Planned Behavior and Social Identity Model of Deindividuation Effect to Predict Online Petition Behavior

Kata kunci/


Introduction
It is unequivocally assumed that the Internet has changed the terrain of politics and political communication as this technology enables diverse political activities and participation to empower citizenship.The internet has become a new place of polity of space of flows (Castells 1999) which propagates horizontal communication among citizens in a network of solidarity that would promulgate social movement (p.

299).
Moreover, other scholars argue the internet also transposes hierarchical system of power into a decentralized structure that flattened government and citizen relationship (Coleman & Blumler, 2011, Zúñiga, Veenstra, Vraga, & Shah 2010).People can exercise power through networking, for example voicing critics to the government and almost unlikely this voice can be suppressed once it was distributed around the since it has no killer switch.
Does the Internet really liberate power in the hand of people?The answer can't be so transparent as it indeed provides chances to decentralize, equalize status and democratize decision making and therefore liberate and empower individuals (Spears & Lea 1994), yet, simultaneously yields a backlash of passivity, clicktivism or slacktivism (Morozov, 2009).
Extant literatures of internet studies identify two camps of thinking: the internet proponent and the opponent.The former argues technological capacities have changed the terrain of political processes into more open, level and lucid that diminish the old structure and set up a new social and an organizational pattern (Gonzales-Bailon, 2013, Dubois & Dutton, 2014).This means the Internet is perceived of being instrumental (Dutton, 2013;Webster, 2013) that it promotes variety of political participations and involvements that ranging from the delivery service to decision making, these included electronic (e)-petition (Earl & Schussman, 2008;Wright, 2012); political consultation (Coleman & Blumler, 2011); online social movement (Gonzales-Bailon, 2013); fifth estate role (Dubois & Dutton, 2014); blogs to induce public opinions (Zúñiga, Veenstra, Vraga, & Shah 2010); and digital campaign (Marcinkowski, & Metag, 2014).Meanwhile, the later raises issues of constant surveillance and coveillance (Rainie and Wellman, 2012); centralized control of powerful authority (van Laer & van Aelst, 2010) and shallow political participation of "slacktivism" through away clicks (Morozov, 2009) which degrade political sphere and citizenry.
The findings from empirical studies were not compelling and oftentimes contradictory, thus it is hard to measure the directionality of the internet effect on politics.Subsequently it begs a question whether this digital technology can benefit democracy, mainly to the people to exercise power over status quo.
Building on the question above, this study tries to investigate individuals political participations in a digitally mediated environment which a citizen can directly exercise his power to the government and state (Earl & Schussman, 2008, Wright, 2012,).An electronic petition (e-petition) was chosen as the quintessence of digital political participation which it is deemed as the materialization of an ideal digital democracy to facilitate direct links between citizens, politicians and government.
The e-petition allows citizens to instill political agenda that can influence a decision-making process through the Internet.It is produced, distributed and presented online and involves low-cost effort (Van Laer, & Van Aelst, 2010;Segerberg & Bennet, 2012;Della Porta & Mosca, 2005;Bennett & Segerberg, 2013;Segerberg & Bennett, 2011).Furthermore, similar to off-line political engagements it can also assist and accomplish its ultimate goal, that is, to serve a democratic purpose (Chadwick, 2008).
Several events have shown e-petition is an effective means to initiate collective actions, for example the British's case in 2011 when 8 million signatures were generated from over 5 million unparalleled email addresses to bombard the Prime Minister official website.Petitioners successfully rectified or cancelled out government policy initiatives that were opposed by the public (Margetts, Hale and Yasseri, 2014).Another case was in February 2007, 1.8 million signed an online petition against a road charging policy initiative that was endorsed by the UK Government (BBC, 2007) and subsequently this strong public opposition resulted in the postponement of the initiative.
The internet has changed the way public solidarity and cooperation was formed (Castells 1999) and previous theory such as Resources Mobilization theory (RMT) is inadequate in explaining digital collective action.The theory postulates top down mechanism to instigate a collective action through structural processes (McAdam, Tarrow & Tilly, 1996;Fligstein & McAdam, 2011) and organizations are important actors which engender resources to initiate and to direct individual involvement in the movement (McCarthy & Zald, 1977).In digital milieu a collective action is built around non-personal attachment, anonymity and decentralized approach (van Laer & van Aelst, 2010).
A movement can be a spontaneous action that the idea was circulated in the net and gains supports from individuals who share similar interest.It had lowered barriers to entry to activate collective action such as time, money, skills and efforts (van Laer & van Aelst, 2010) and this provides opportunities for citizens to engage closely with politics without concerns about the availability of resources and risks to support a political cause.Yet, it is hard to measure the effectivity of digital collective action since this action consists of weak ties, thus this raises questions about collective identity, social capital, norms and power distribution that were the main ingredients of off-line collective action (Tarrow & Tilly, 1996;Morris, 2000;Tarrow & Tollefson, 1994).
One of the answer is that Computer-Mediated Communication (CMC) facilitates a new social setting which diminishes the old boundaries (Postmes, Spears & Lea, 1998) when anonymity leads to normative conduct (Postmes, Spears, Sakhel and de Groot, 2001) and strong social attachment (Leas, Spears, and de Groot, 2002).It can be argued that an online network, which consists primarily of weak, bridging ties (Granovetter, 1973) is likely to benefit digital connection more so than strong, bonding ties (Putnam, 2001) because the Internet would enable anonymous individuals who share a similar interest in certain issues to connect and join with the civic action.
Pew Global research ( 2014) indicate that online collective actions have increased markedly in the past few years and almost exceeded traditional ones.This finding supports CMC approach that an individual bottom up initiative works better than structural influence.Therefore, individuals have more control over collective action than organizations.To address this issue, the social movement studies have extensively researched on an individual level to disclose social psychological mechanism that drive individuals' involvement in the collective action (Tarrow & Tilly, 1996;Morris, 2000) and these studies suggest human action is a rational and motivated behavior which utilize a systematic scheme to process information (Ajzen & Fishbein, 1980;Ajzen, 1988;Fishbein & Ajzen, 2010;Ajzen, 2012).
Therefore, this study tries to elucidate the extent social and psychological factors affect individuals involvement in a collective action.A reasoned action theory (Ajzen 1988, Fishbein & Ajzen, 2010;Ajzen & Fishbein, 1980) provides a framework to interrogate psychological elements to predict behavior, while collective identity as social determinant can also influence these action as assumed by CMC approach (Postmes & Spears, 1998;Postmes & Brunsting, 2002).

Literature Review Social Movement Repertoire
There are two big school of thoughts in the social movement literatures: (1) structural based theory from classical social theories of Marx and Durkheim (Tarrow & Tilly, 1996;Morris, 2000;Tarrow & Tollefson, 1994).A movement was perceived as a function of structure which an individual who occupies a role in the structure will function accordingly and subsequently this approach was developed into resource mobilization theory (McCarthy & Zald, 1977;McAdam, Tarrow & Tilly, 1996;Tarrow & Tollefson, 1994;Fligstein & McAdam, 2011); (2) actor-based theory emphasizes a social psychology paradigm to interrogate psychological and social factors of a movement (Morris, 2000;Tarrow & Tollefson, 1994;Klandermans 1984;McAdam) Resource mobilization theory (RMT) emphasizes organizational resources, political opportunity and collective identity (McCarthy & Zald, 1977;McAdam, Tarrow & Tilly, 1996;Tarrow & Tollefson, 1994;Fligstein & McAdam, 2011).Social movement can only be understood as a struc-tural process where organizations become prominent actors to instigate, coordinate and mobilize actors.Organizational resources become prominent factors for action where all various forms of capital (e.g.material and non-material) are possessed by an organization (McAdam, Tarrow & Tilly, 1996;Tarrow & Tollefson, 1994).With abundant resources, organizations can frame an issue to form a collective identity (Fligstein & McAdam, 2011) that will attract people to join with collective action, and organizations also can see and create a political opportunity for the action (McCarthy & Zald, 1977).The individual is a rational actor who is involved in social action in response to external stimuli and his/her behavior can be perceived as playing a role as a member of a group or organizations.RMT has been the dominant paradigm in social movement and many social movements around 1960-1980 have been studied using this theory (Tarrow & Tilly, 1996;Morris, 2000;Tarrow & Tollefson, 1994).
RMT failed to capture the free-riding phenomena in which individuals often ride the wave of a movement to reap benefits with minimal to no effort (Klandermans, 1984;Morris, 2000;Tarrow & Tollefson, 1994;Fligstein & McAdam, 2011).An American economist, Mancur Olson, proposed a cost-benefit model to examine individual behavior in social movement (McAdam, Tarrow & Tilly, 1996;Morris, 2000) and suggested that individuals participate because of a cost-benefit calculation, getting high benefit with low effort.This model provided an answer the free-riding problem, but failed to capture the complexity of individual behavior.The cost-benefit model emphasizes personal advantage while neglecting non-material or social motives of behavior.Later the expectancy model was developed to synthesize diverse motives of human behavior that may underpin involvement in a movement (Klandermans, 1984;Brunstig & Postmes, 2002).Klandermans (1984) identified three distinct motives: collective or goal motives, social motives and reward motives.Collective or goal motives refer to the attainability of the collective goal (e.g.number of participants who join the movement, expectancy regarding the effectiveness of movement, and evaluate the effective individual contribution during action); social motives indicate the perception of expectation of significant others and the personal importance reaction; and reward motives stress non-social cost and benefit of participation, such as money, skills and time (Klandermans, 1984;Brunstig & Postmes, 2002).The expectancy value has been criticized as overemphasizing on individual capability having full control over its action which ignores individual attribute differences and social stratification.This model has been tested to analyze diverse social movements, including: the labor movement in Netherland (Klandermans, 1984), the old people movement in Germany, the gay movement in the US (Simon, et al, 1998); and the Netherland environmental movement (Postmes & Brunstig, 2002;Brunstig & Postmes, 2002).The expectancy model is still a valuable framework to study social movement from the micro-level of individual behavior and this is relevant with modern condition of society where political decision-making is highly personal and people join movements as individual lifestyle choices (Giddens, 1990).Collective action and identity formation are a personal choice that are facilitated through the fluidity of social networks (Bennet & Serberger, 2013;della Porta & Mosca, 2005;Castells, 2013)

Social Psychology Perspective on Online Social Movement Psychological Components of Behavior
Social psychology has been used extensively in social movement studies to examine the individual social behavior and two models have emerged: the cost-benefit model and the expectancy-value model (Klandermans, 1984;Tarrow & Tilly, 1996;Morris, 2000;Tarrow & Tollefson, 1994).Individual autonomy and capability to conduct higher cognitive processing is central to a social psychology perspective (Ajzen, 1988;Ajzen, 1991;Fishbein & Ajzen 2010;Bursting & Postmes, 2002).The cost reward model was used for a limited time due to lack of explanatory power and therefore left expectancy value as a dominant framework for studying social movements (Tarrow & Tilly, 1996;Morris, 2000;Tarrow & Tollefson, 1994).
During the course the expectancy value model has morphed into theory of reasoned action (TRA) since both constructs are overlapping (Klandermans, 1984;Simon, et al 1998) and they were built upon theory of attitude (Simon, et al 1998;Ajzen, 2012).The original Klandermans' expectancy-value identified three motives as the main predictors of willingness to participate: collective, social and reward (Klandermans, 1984;Simon, et al 1998) and these constructs are analogues to subjective belief and subjective norms of TRA (Ajzen, 2012;Postmes & Brunig, 2002;Brunig & Postmes, 2002).
However, to predict intention to act using TRA is a flaw since political behavior is not fully volitional control behavior.Political action often exceeds personal motivational factors, for example, when the availability of requisite opportunities and resources may interfere the outcome (Ajzen, 1988).Therefore the extension of TRA, later called theory of planned behavior (TPB), will be suitable to overcome volitional control problem (Ajzen, 2012;Ajzen 1991).A study of the voting behavior of US citizens found that political behavior is unpredictable as evidenced by a low voter turnout during elections (Fishbein & Ajzen 2010).Many elements can inhibit a person's decision to vote or not vote, particularly background factors such as individual, social and information (Fishbein and Ajzen, 2010).
Understanding a particular behavior of inter-est could be challenging, but TPB provides a parsimonious model that encapsulates basic determinants of human social behavior (Ajzen, 1988;Ajzen 1991;Fishbein & Ajzen 2010).As a model that explains human social behavior, TPB has been used to examine diverse social behavior and meta-analysis proved TPB as an integrated model has good predictive power (i.e.Theory of planned behavior (TPB) is extensively referenced and used in social psychology research to predict and understand individual's motivational influences on intention and behavior (Madden, Ellen, & Ajzen, 1992).TPB acknowledges fluctuation in controlling behavior because individuals were unlikely having complete volitional control (Fishbein & Ajzen, 2011;Armitage & Conner, 2001).TPB is commonly mistaken as a theory that measure only rational and deliberate behavior because "planned" term in TPB.Planned refers to the consistency of formation of attitudes, perceived norms and perception of control and the intention they produce.This flow is not always perceived as reasoned, but often automatic from their belief about performing behavior and also TPB never assumed that belief is accurate, unbiased and rational (Fishbein & Ajzen, 2010).Cognitive foundations of belief can produce spontaneous and automatic fashion of behavior intention.
TPB can be a perfect substitute for Klandermans' expectancy-value model because it provides more variance to predict political behavior.Individuals don't have full control over political behavior, therefore, perceived behavioral control would affect the outcome: intention to participate and behavior.Research on political behavior using TPB has proved that this model is a better predictor of willingness or intention to participate than Klandermans' model (Postmes & Brunstig, 2002;Simon, et al, 1998;Armitage & Conner 2001;Fishbein & Ajzen, 2010).
Belief is an imperative factor in online petition behavior.It is assumed people consciously support a political cause and TPB is developed to explain human behavior that based on belief (Ajzen, 1988;Ajzen, 1991;Fishbein & Ajzen, 2010).According to TPB, behavioral belief, normative belief and control belief are antecedents for attitude toward behavior, subjective norms and perceived behavioral control consecutively.From a theoretical perspective, TPB is highly compatible with the online petition phenomenon.
The principle of compatibility of TPB (Ajzen, 1988;Fishbein & Ajzen, 2010;Ajzen, 1991) requires behavior under interest should be specific action than the general behavior pattern, the principle of compatibility advises that attitude and behavior of interest must correspond in terms of action, target, context and time.The more specific behavior of interest the more accurate TPB level of predictability (Fishbein & Ajzen, 2010;McEachan, Conner, Taylor, & Lawton, 2011;Madden, Ellen, & Ajzen, 1992).

Social Component of Behavior
An online petition is categorized as a social movement in which individuals voluntarily participate in group action through persuasive or confrontational means to achieve a collective end driven by a distinct ideology, value or political purpose (Postmes & Brunsting, 2002).However, different lines of study have integrated online petition as a form of social movement (van Laer & van Aelst, 2010;Brunstig & Postmes, 2002) defined as "networks of informal interaction between a plurality of individuals, groups and/or organizations, engaged in a political or cultural conflict on the basis of a shared collective identity" (Diani, 1992, p. 13).
This would make social movement is a repertoire of collective action or put differently, a social movement is the networks of collective action and in recent studies, these terms are often used interchangeably (Postmes & Brunig, 2002;Brunig & Postmes, 2002;Varnali & Gorgulu, 2015, Van Zomeren, Postmes, & Spears, 2008;Kelly & Berlinger, 1995).Both concepts have overlapping constructs and there are three essential ideas that underpin both: interaction among people in a group setting, shared values that form group cohesion and agreement to work together to achieve a common outcome.The author will use social movement since it is a general term used by political scientists.
A study of older peoples' movement in Germany and the gay rights movement in the USA found that collective motive, social motive, and reward motive are predicting of willingness to participate and added social identification was significantly increased the variance explained (Simon, et al, 1998).Other studies of the environmental movement in the Netherlands also obtained similar findings that social identification was a significant predictor of intention to participate in the movement (Postmes & Brunig, 2002;Brunig & Postmes, 2002).These studies confirm the social movement definition (Diani, 1992) that social identity is a crucial concept in understanding individual participation in collective action (Klandermans, 2004).
However, the individual is living in a web of social structures and one's identity may be stratified, that is, one can simultaneously identify as man, Asians, middle age and journalist, which one is the key identity?The research findings indicate that in the context of social movements, individual attachment to collective movement is the most visible social identity (Postmes & Brunig, 2002;Brunig & Postmes, 2002;Simon, et al, 1998;Kelly & Berlinger, 1995).For the purpose of this study, two distinct constructs have been developed to differentiate traditional and online social identity: one measuring attachment to a social group and the other gauging collective movement.
SIDE perspective is congruent with the social identity perspective of a social movement that individuals will cognize with lower levels of identification in this case is the collective group identification by releasing higher levels of social category.This means the SIDE model is compatible with the study of collective action in an online setting irrespective of the present state of the individual (i.e.anonymous and isolated) and proximity of the other group members (Postmes & Brunig, 2002).
SIDE model can provide the ideal circumstance for online political participation that was less constrained.People can express or exercise their action with low effort and less time consuming.Put in the real context some people will feel comfortable in anonymous condition, particularly when donating through CMC or want to join with a cause through online petition rather than attending a public demonstration.However, one critique should be addressed that a collective action is a form of "slacktivism" when participation is shallow action.
There are two objectives that underpin this study: first combining two perspectives: social psychology and resource mobilization, and to examine online petition activity that has never studied before.A synthesis model is built that based on expectancy-value and social identity approach.Unlike the previous expectancy model, TPB is used as a basic theoretical model because this theory can provide better prediction of human behavior (McEachan, Conner, Taylor, & Lawton, 2011;Armitage & Conner 2001;Albarracin, Johnson, Fishbein, & Muellerleile, 2001).
Additional concepts of social identity and collective identification were added into the model to increase the model's explanatory power.This model also aligns with previous research to predict intention through three factors: affective, individual and collective belief and cognitive calculation.Affective will be measured by collective identification, individual and collective belief or attitude toward behavior and subjective norms and cognitive calculation is perceived behavioral control
A few studies have been done to elaborate structure and actor paradigms in the social movement fi eld (Postmes & Spears, 2008;Brunsting & Postmes, 2002;Bennett & Segerberg, 2013), and this study tries to corroborate previous fi ndings and to also elucidate whether weak ties play a signifi cant role than strong ties.

Theoretical Model
The model (fi gure 1) is a combination of TPB and SIDE model which several hypotheses (table 1) are developed to provide explanation of online petition.Hypotheses 1 to 4 were drawn from original study of TPB (Ajzen, 1988;Fishbein & Ajzen, 2010;Ajzen, 2012) and these have been employed to several studies (e.g.Simon, et al, 1998;Lee & Robins, 1998;Armitage & Conner, 2001;Madden, Ellen, & Ajzen, 1992;McEachan, Conner, Taylor, & Lawton, 2011), therefore: H1: positive attitude towards online petition associates with higher intention to an online petition H2: the strength of subjective norms increases intention to online petition H3: perceived behavior control gives rise to intention to an online petition H4: perceived behavior control infl uences to signing an online petition H5: behavior intention and perceived behavior control positively associates with signing an online petition.
Subsequently SIDE model was incorporated to provide more variance into the model and social identity is identifi ed to have a direct connection to intention to participate (Postmes & Brunig, 2002;Brunig & Postmes, 2002): H6: higher attachment social identity as Americans increases intention to an online petition H7: higher attachment to a movement identity positively associate with intention to an online petition Furthermore, the author wants to examine which social identifi cation would signifi cantly infl uence norms, is it as Americans or attachment to a movement?If an individual's norms heavily rely to the movement, this would support CMC

Method
To test the hypotheses, the study is designed and conducted through an online survey on Amazon Mechanical Turk.An invitation to participate in the study was posted on MTurk for 7 days in week four on November, 2015.MTurk worker qualifi cations for the current study included a "human intelligence task" (HIT) acceptance rate of 90% or greater.The subject pool was further limited to those who were located in the United States and at least signed one online petition in the past six months that related to political and environmental issues.220 MTurkers completed the survey and they received 0.33 cents as an incentive, yet only 187 questionnaires were passed the attentive check questions and used for the final analysis.

Respondents
Demographic data demonstrate the gender proportion is equally distributed 49.7% male and 50.3% female yet, skewed toward Caucasian (80%).Approximately over half of the respondents have a degree from a college or equal to college with income over $40,000 a year.These fi ndings likely support digital inequality thesis, which savvy Internet users are coming from middle-upper social economic class.Almost half of respondents have strong connections with democrat party and another one-third claimed as an independent.On average, three petitions were signed in the past six months and mostly related to political issues.However, the fi ndings can't be generated outside this research context since participants are MTurkers who voluntarily involved can be perceived as a non-probability con-venient sampling method.
A self-report method for TPB has been heavily criticized for promoting bias of self-presentation and social desirability, but meta-analysis indicated only a small effect (5%) has been identifi ed these factors affect TPB components (Armitage & Conner, 2001).Align with this, another problem was addressed to question behavior measurement.Some studies used a longitudinal design to measure behavior, but for cross sectional research often TPB measures past behavior as proxy of future behavior (Ajzen, 2012;Ajzen, 2011).

Validity and Reliability
Adopting constructs from the theory of planned behavior and SIDE, all indicators yield good convergent and discriminant validities, and reliability (table .2) Reliability tests were conducted to evaluate items consistence across constructs and all variables are yielding a high cronbach's alpha ranging from 0.75 to 0.90 (table.2).Subsequently, no items were dropped from this study.
Factor analysis also yields a similar result as all items were loaded into their constructs and no items were dropped from the analysis.This indicates the study has strong internal validity, mainly for confi rmatory factor analysis, which the theoretical model was supported by the statistical results.Moreover, further analysis using SEM can be conducted as good latent variables would adequately estimate parameter of the entire theoretical model (Dijkstra & Henseler, 2015).

Behavioral Intention
Initially, OLS regression was conducted to test behavioral intention and regression analysis corroborates TPB model attitude towards behavior, subjective norms and perceived control were simultaneously explained individual behavioral intention.Perceive behavioral control is the strongest predictor of behavioral intention and its value almost twice of attitude toward be- havior.It is likely individuals confidently believe that they are capable signing an online petition in the coming months and there are no inhibiting factors that can deter them from doing it.High self-efficacy can also directly promote real action when an individual believes he/she can deliver action in a specific circumstance to accomplish a goal (Bandura, 1986).
Moreover, the result also indicates three components of attitude, norms and control generate 45.2% proportion of variance in predicting behavior intention and this finding corroborates previous TPB research that the model can yield moderate to high score of variance explained (Albarracin, et al 2001;Armitage & Conner, 2001).Signing an online petition can be assumed as a volitional behavior which an individual carefully planned action through the consistency of formation of attitudes, perceived norms and perception of control and the intention they produce.Someone who has a propensity towards an online petition, under influence of social norms to support a cause and having a high level of efficacy will engender intention that leads into real action.Hypotheses one to four were confirmed that TPB model is sufficiently predicted an individual behavior signing an online petition.
Further model 2 (Table .2) attests limitation of computer-mediated environment where verbal and non-verbal cues are barely detected (β = -.0.82, p > .05)and this means individuals detracted from social groups when forming a collective action.Personal attributes were intangible that makes situational norms (i.e.identification to action) and group identity become salient (Postmes & Spears, 1998;Lea & Spears, 1991;Lea, Spears & de Groot, 2001).The result suggests attitude and group identification are statistically insignif-icant and failed to generate variance explained to improve model prediction (Table 2, model 2).The finding only supports hypothesis: higher attachment to an online petition identification increases intention to an online petition and identification with the broader social group didn't give much sense to an individual who support a petition.
Anonymity leads individuals to ground decision using relevant and available resources.In the context of digital collective action, individuals who supported and signed a petition are less likely to have a personal attachment from previous encounters, thus a group identity is unattainable and people will heavily rely to an online petition itself, identification with the cause.Since individuals can't relate to group identity, it deflates their subjective belief and simultaneously create uncertainty for final goal achievability.This process is similar to computer-mediated communication (Walther, 1996) assumption that communication through the computer is impersonal therefore the richness of text will drive further relationship for people whom they connected through the Internet.The cues-filtered-out condition, lack of socioemotional expression and text-based message become instrumental and functional which message elements (i.e.language content and style) are conduits of interpersonal cues (Walther, van de Heide, Ramirez Jr, Burgoon and Pena, 2014).
Even though text-based message can convey nonverbal cues for emotional attachment, the process is taking longer than the face-to-face condition.The lack of group identity and interpersonal encounter, an individual who support the petition decodes a message thoroughly to instigate social identity to the collective action.
Full model examines demographic data (Table 1, model 3) as control variables that might influence

Example of Indicator
I intend signing an online petition to support political action in the forthcoming six months (extremely unlikely/likely, 7 scales) For me signing an online petition to support political action in the forthcoming six months (harmful/beneficial, 7 scales).

Behavior of Interest
Furthermore, OLS tested two main determinants of behavior in action and the fi ndings were surprising as none of them are statistically signifi cant.However, when each variable was tested independently, the fi nding shows, fi rst an individual action signing an online petition is a product of behavior intention (β = 0.074, p < 0.001) and this confi rms the expectancy model (Ajzen, 2010;Ajzen, 1988) while an individual who has a strong disposition toward a behavior will almost certainly do the behavior intention.
This research also found no association between behavioral control and individuals' action (β = 0.022, p > 0.05) and it indicates in computer-mediated setting, a high degree of control of internal and external factors hinders an individual to act spontaneously as intention mediates perceived control and actual behavior.The Internet enables a less constrained environment where one can safely conduct collective action with less risk or cost (Van Laer, & Van Aelst, 2010), but this only happens if an individual acquires positive intention toward behavior.
It can be concluded based solely on the regression analysis, TPB model works well to explain signing an online petition.Individual behavior is perceived rational and volitional, and intention is a signifi cant predictor of behavior of target signing online petition.Hypotheses one to four are supported.

Online Petition between Rational Action and Social Infl uence
Structural Equation Modelling (SEM) was utilized to analyze the full model of signing online petition (Model fi t SRMR = 0.075, NFI 0.728) and it supports hypotheses 5 (r = 0.188, p < 0.05) and 6 (r = 0.169, p< 0.05).But, SEM slightly changes TPB model as attitude toward online petition is insignifi cant, and its relationship to intention is fully mediated by subjective norms.This warrants that anonymity has created uncertainty for individuals and a positive attitude does not drive intention to participate in online petitions, rather they are trying to confi rm with signifi cant other's preferences, the subjective norms.Social ties, mainly important persons, have a greater infl uence in CMC setting, on which individual lean when social cues are unavailable.
This fi nding may relate to the fi lter bubble (Pariser, 2012) and social media polarization phenomena (Lee, Choi, Kim & Kim, 2014), which describe a phenomenon in which social networks have strong infl uence on an individual's action on the Internet.Individual behavior can be predicted by him/her social networks and often creates inclusive ties.In this regards, CMC may yield a Table 3. oLS Regression of TPB drawback and negative action, yet in the TPB model this is not the case as another construct, the perceived behavior control, can counterbalance negativity, when a person has fully control over behavior, not merely rely on signifi cant others.
Despite mediated through subjective norms, the study also identifi es a direct correlation between attitude and actual behavior (r = 0.236, p < 0.05).This path supports previous fi ndings of attitude-behavior research that attitude signifi cantly and substantially predicts behavior (Kraus, 1995).Following TPB model, attitude consists of behavioral beliefs which an individual makes an evaluation of the expected outcomes and this belief is the subjective probability that is readily available and can be accessed at a given moment (Ajzen, 1988).
This evidences SIDE has played a greater role in CMC setting when persons immerse in a collective action and anonymity hint to a loss of selfhood (Reicher, Spears & Postmes, 1995).Persons less inclined toward the outcome of the action, since they heavily attach to the petition and de-individuation gain positive effects without losing control over behavior, maximize the opportunity of individuals to exploit their collective iden-tity (Ajzen & Fisbein 1980;Ajzen, 1988).It can be assumed that a person yields simple calculation of probability whether his/her action would have a future impact.
Someone who signs an online petition may engender positive beliefs about the action and its future outcomes.This is a shortcut mechanism whenever a decision should be made immediately and no subjective norms or state control are available.In this context, this path attitude-behavior can be perceived as "one click away activism" that has been criticized heavily as passive or clicktivism (Morozov, 2009).
Another important aspect to be highlighted is subjective norms to social identity relationship as weak ties have broadened the scope of referent groups.The TPB constructs only defi ne norms as perceived of desirable behavior that referred to important others' expectation and behavior (Fishbein & Ajzen, 2011) and this study found individual membership as part of the greater social group affects perception about norms.In the CMC context, information becomes an important commodity as individuals' reliance on text-based petition would be prominent and weak ties can serve the informational purposes (Granovetter 1973).Several studies show weak ties is benefi t-  & Westhoff, 2006;Weening & Midden 1991) and this to be the case in the context of online petitions.People are looking for conformity not only through their importance person, but also look up to the social group.Internet, mainly news websites and social media offer the platform where information about social group are free and can be accessed anytime.In the future, a distinction between strong and weak ties may be irrelevant as the Internet can provide dense connectivity that will diminish close and distance relationship.
Meanwhile, identification of the cause of a movement directly affects intention to sign an online petition and this support previous studies (Simon, et al, 1998;Lee & Robins, 1998) which social and emotional attachment stimulates an individual's readiness to execute a given behavior (Ajzen & Fishbein, 1988).A person's attachment to a movement would instigate positive intention signing online petition and then leads to actual behavior.

Conclusion
Traditionally petition is a top-down collective action where an organization instigates the action and information diffuses through interpersonal communication.Society has changed and new technology has become prominent as our social morphology (Castells, 2000;Castells, 2013) and has reformatted social connection and simultaneously political participation.The Internet diminishes time and space constraints for political activity while simultaneously reducing social identity as people can anonymously join in a collective action.Due to the low cost of involvement and anonymity condition, the Internet opponents claimed digital collective action is clicktivism or slacktivism activity (Morozov, 2009).
This study identified that online petition to a certain degree is a volitional behavior when social norms and behavioral control affects an individual's intention which leads to signing an online petition.Positive attitude toward signing online petition creates two routes: first, it directly affects behavior, and secondly mediated by subjective norms it can arouse favorable intention toward behavior.Align with this, the identification of an issue has compensated the lack of group identity which individual derived socioemotional attachment through the text-based message.
Furthermore, this study elucidates two types of social connection: identification of social group and identification with the movement.The former correlate with a person's subjective norm, while the latter associated with the intention to a given behavior.All in all, individual political participation in CMC setting is a rational action with a degree of control that was influenced by social pressure: attached to a social group and tied to the movement.The findings have challenged previous studies that digital political participation is neither rational action nor slacktivism.Social pressure still plays an important role in supporting or inhibiting individual action.
The author acknowledges that this study did not disclose a mechanism to what extent one petition can generate thousands of signatures while others were failing in the first place, nor examine message elements as a medium of persuasion that can explain the effectivity of petition to arouse socioemotional attachment to compensate face-to-face communication and non-verbal cues.These two suggestions can be pursued in future research.

Figure 1 .
Figure 1.Theoretical Model for online Petition Behavior Most people who are important to me think that signing an online petition to support political action in the forthcoming six months (I should not/should, 7 scales) For me signing an online petition to support political action in the forthcoming six months is (impossible/possible, 7 scales) In many respects, I am like most other Americans (not true/true, 7 scales) I feel strong with other political activists (not true/true, 7 scales) intention to sign an online petition and the results show no demographic variables are signifi cant.It seems participants are Internet literate that they are a homogenous group who acquire digital dividend with their Internet use.

Figure 2 .
Figure 2. Structural equation Model of Factors that Contributes to online Petition Behavior Model fi t indicators: SRMR 0.075, NFI 0.728

Table 1 .
Hypotheses of the Model

Table 2 .
Validity and Reliability