From access deprivation to skill acquisition: Cluster analysis of user behavior in face of a 12-hour legal blockage of WhatsApp in Brazil
First Monday

'From access deprivation to skill acquisition: Cluster analysis of user behavior in face of a 12-hour legal blockage of WhatsApp in Brazil by Marcelo Santos, Magdalena Saldana, and Andres Rosenberg



Abstract
This study takes advantage of a forceful legal 12-hour deprivation of access to WhatsApp messaging service nationwide in Brazil on 18 December 2015. Right after the blockage, we ran a survey to capture the reaction of Brazilians that were cut off from the App, aiming to understand which factors point to the capacity to successfully circumvent the blockage. Anxiety, digital skills and gender were found to be related to success, while isolation and age were not. Furthermore, a cluster analysis of 306 respondents that attempted to bypass the blockage identified four groups that summarize the reaction patterns in face of the blockage: the deprived, the challengers, the addicted and the elite. We discuss the possible implications of the findings for the field.

Contents

Introduction
ICT deprivation
Emotional arousal
Digital inequalities
Explaining the blockage
Method
Variables
Results
Discussion
Conclusion

 


 

Introduction

Digital instant messaging (IM) platforms have opened new near-ubiquitous mediated forms of communication not only to keep connected to significant others, but also to receive news (Opinião Pública, 2018) and foster political campaigns (Nemer, 2018). Amongst such IM tools, WhatsApp’s relevance is undeniable. While many countries in the northern hemisphere rely on technologies such as SMS or other apps for interpersonal communication, Latin-American countries like Brazil and Mexico are among the top five in worlds’ WhatsApp penetration, with 56 percent of both their population using this app (Statista, 2019). In Brazil’s case, the Internet penetration was 65 percent in 2017 (Gomes, 2018), so an estimate 86 percent of the population with access to Internet was actively using WhatsApp in the beginning of 2017. Estimated in 120 million users around the same time, Brazilians represented (and still represent) around 10 percent of WhatsApp’s total population (Mazzeto, 2017). Brazilians are perhaps as important to WhatsApp as WhatsApp is to Brazilians.

Nevertheless, a forceful legal deprivation of access to WhatsApp IM service took place in Brazil on 18 December 2015, shutting down access nationwide for around 12 hours, leading many of the users to pursue tools to circumvent the blockage.

This study takes advantage of such a unique situation to understand some aspects of skills acquisition in the context of technology deprivation: on the days that followed the blockage we ran a survey to capture the reaction of a part of the 100 million Brazilians that were cut off from the app. Though the blockage was due to WhatsApp’s failure to comply with a solicitude of disclosure of information to feed a drug traffic ongoing investigation [1], our data allows to shed light on the social and technical factors related to the success or failure as users attempted to bypass the blockage.

The most common method to circumvent the blockage was the installation of a Virtual Private Network software that provides a “fake” IP address, emulating the access from a different country to evade the blockage. We claim that such procedure is a relatively high-skilled tech bypass, for a country unaccustomed to government censorship practices on the digital realm.

Findings from this study suggest commonalities that help us understand the successful transition from access deprivation to skill acquisition — as well as the unsuccessful efforts to do so.

 

++++++++++

ICT deprivation

Complete or partial Internet shutdowns are becoming more common in the last few years (Access Now, 2018). Besides Brazil, 81 other temporary digital media-related [2] disruptions were detected in 2015 (West, 2016). While some governments aim to better control the information citizens consume (Esarey and Xiao, 2011), others justify it by making a case for public safety, combating disinformation and national security (Access Now, 2018).

There are (at least) two distinguishable types of ICT deprivation: (1) Shutdowns, when an entire application or network is blocked; and (2) Content censorship, when certain themes, visual representations or keywords are subject to control. While the present research identifies with the former, the latter may differ as to an ex-ante or ex-post implementation. Thailand’s law lèse majesté not only allows authorities to block content on the ISPs (ex-ante), but also punishes users (ex-post) from defaming monarchy, such as the late king, the royal dog and even “liking” such content on social media (Gebhart and Kohno, 2017). The Chinese “Great Firewall” (GFW) is perhaps the most infamous and elaborate example of content censorship ex-ante, including “network filtering, search filtering, chat censorship and blog censorship” [3]. In order to be more efficient in such censorship, “any user-generated content websites without a proper government gatekeeper are generally censored” [4], and nationally owned tools are offered in its replacement not only in China, such as Weibo, but also in Iran as the country has been trying to ban Telegram for quite a while now (Kargar and McManamen, 2018).

The social and behavioral consequences of any variation of this nature of disruptions are very difficult to predict. On one side, Comunello and Anzera (2012) claim that censorship has detrimental effects on civic engagement, since users don’t have access to all content, thus narrowing their views. But according to Hassanpour (2014), when the Egyptian government promoted an Internet shutdown in January 2011, during the so-called ‘Arab Spring’, the blockage was counterproductive for the regime because “in the absence of information about the crisis, others took to the streets, eager to find out what was going on” [5]. Additionally, Russia’s ban of Telegram has led to the malfunction of banking and retail services (Savov, 2018) and triggered social unrest, fed by its very founder Pavel Durov, an exiled Russian: “The history of our ancestors has taught us to fight until the end” (Durov, 2018, translated with Google).

Though most of the examples depicted are more extreme than the case presently studied, the deprivation of digital communication services should spark users’ emotions, which might act as catalysts for behavior, as we will make the case in the next section.

 

++++++++++

Emotional arousal

The only approach to emotions we found in the literature on censorship circumvention and ICT deprivation is an analysis of the relationship of users’ attitude towards censorship and circumvention tools in the Chinese context (Mou, et al., 2016); the results were not statistically significant. We chose, instead, to study users’ emotional arousal after the blockage in Brazil, regardless of their attitude towards censorship. This approach suggests that emotions should also be studied as to its physiological arousal which respond to patterns of cognitive appraisals (Smith and Ellsworth, 1985), rather than the individual impact of specific emotions or the combined valence. It points to the intensity of the emotional response: while high levels of emotional arousal are associated with activation, or action-oriented behavior, low levels are associated with deactivation or relaxation (Feldman Barrett and Russell, 1998; Berger and Milkman, 2012).

Based on reactance theory, Mou, et al., (2016) claim that “when people have an unpleasant psychological state because of their loss of free information, they will form a series of psychological responses and make an effort to regain the censored information” [6]. It seems plausible that high levels of emotional arousal could operate as triggers, to some measure or within a set of conditions, to a somewhat radical behavior to recover access not to some specific censored information, but to the blocked IM app. Drawing upon this literature, this study suggests that high emotional arousal will increase the likelihood of a user bypassing the WhatsApp blockage, leading to Hypothesis 1:

H1: Emotional arousal will positively affect users’ ability to overcome the WhatsApp blockage in Brazil.

 

++++++++++

Digital inequalities

While there are many forms digital inequalities may take, some demographic variables are usually related to advantages on ICT appropriation and usage. For this study, we consider education level, gender and age. Additionally, we acknowledge digital skills as one of the explaining factors, considering the adoption of a VPN as a digital skill to be acquired by users.

Skills

The literature identifies key demographic characteristics that explain digital divides within different contexts: age, geographic location, gender, socioeconomic status, race and ethnicity (Gilbert, 2010; van Dijk, 2005). But skills also matter, as per the second digital divide which addresses gaps in skills and usage (van Dijk and van Deursen, 2014). Studies have also found that younger people, white people, and more educated people have better digital skills, not just to consume technology but also to accomplish different tasks, and to do so efficiently (Hargittai, 2002).

Such divide is socially relevant in many dimensions, as better digital skills provide advantages in economic, social, political, cultural, educational and institutional aspects and even in citizen participation in society (van Dijk, 2005). Therefore, our second hypothesis expects digital skills to make a difference when explaining users’ reaction towards the WhatsApp blockage:

H2: Digitally skilled users will be more likely than less skilled users to overcome the WhatsApp blockage in Brazil.

Education

Brazil’s digital inequalities are highly affected by education: as a general rule, the higher the level of education, the higher the levels of Internet penetration in the population (Instituto Brasileiro de Geografia e Estatistica, 2018) in great part due to the correlation between education and income. In fact, 95 percent of the individuals with higher education are Internet users, while only five percent of those with no education are able to use the Internet [7]. Among houses with Internet access, 64 percent of Brazilians had landline broadband for Internet connection, while 25 percent connected through mobile plans, a proportion that remained stable from 2016 to 2017 [8].

Additionally, people who are unemployed and with weaker social networks are granted less support for learning from proximate others (DiMaggio and Hargittai, 2001) than those with more developed networks and in possession of a job (van Dijk, 2005) or with college degree (Mossberger, et al., 2008). In this context, education can be considered as a proxy for social capital — less educated users will be less likely to access more knowledgeable networks that might help them with the blockage, or might be less likely to teach themselves how to do so. Consequently, our third hypothesis suggests:

H3: Education level will positively affect users’ ability to overcome the WhatsApp blockage in Brazil.

Gender

While studies have systematically shown that women are as capable as men at learning and improving their digital literacy (Liff, et al., 2004), Cooper and Weaver (2003) claim that differences on how men and women have been educated regarding technology and overall digital literacy lead to such gender gap. As explained by Correa (2016), gendered social norms have socialized the idea that technology is a male domain, which might explain differences in digital skills even among highly connected groups.

On the other hand, in a study of females from twelve Latin American and eighteen African countries in a three-year span, Hilbert (2011) came to the conclusion that the gender divide is an extension of unfavorable conditions of employment, education and income. In other words, it would be a collateral effect of structural gender inequalities. Either way, our fourth hypothesis expects to find differences regarding the blockage attributable to gender:

H4: Males will be more likely than females to overcome the WhatsApp blockage in Brazil.

Age

Matassi, et al. (2019) found that life stages may entail different aspects of domestication of WhatsApp in neighboring country Argentina. They vary from an ‘always on’ mode for teens to a generational bridge tool used less frequently as users belong to older groups.

Hargittai (2002) claims that “young people (late teens and twenties) have a much easier time getting around online than their older counterparts (whether people in their 30s or 70s)” [9]. Younger generations also act as ‘digital media brokers’ (Correa, et al., 2015) and as inducers of digital skills in isolated areas (Correa and Pavez, 2016).

Drawing upon these findings, we propose our fifth hypothesis:

H5: Age will negatively affect users’ ability to overcome the WhatsApp blockage in Brazil.

 

++++++++++

Explaining the blockage

The literature presented above reassures the relevance of identifying different levers that facilitate digital skills acquisition, either as a personal advancement or as a perceived necessity in times of politically motivated censorship, public security, or legal conflict, as the present case. Such moments of exception lead users to “incident-driven strategies” (Gebhart and Kohno, 2017) to circumvent the shutdown. Gender, age and education have proven to be related to digital skills acquisition, while emotional arousal is a relevant factor to predict people’s reactions to unexpected events. Together, these variables are important to look at when understanding users’ behaviors toward a legal technology blockage. Thus, this study aims to understand which factors are more influential to explain whether users were or were not able to bypass the WhatsApp blockage. Thus, we pose the following research question:

RQ1: How do WhatsApp users classify depending on their digital skills, arousal and ability to bypass the blockage in Brazil?

 

++++++++++

Method

This study relies on data from an online survey conducted right after the WhatsApp legal blockage in Brazil, in 18 December 2015. The questionnaire was written in Portuguese and administered using Qualtrics. The link to the survey was advertised to Brazilian users through a Facebook page created for Social Media Studies, and it went live from 22 December 2015 to 13 January 2016 (see survey summary in the Appendix).

In this timeframe, more than 600 online users completed the survey. This study explores a subset of the answers, since we look at those who attempted to bypass the blockage (N=306), observing those who failed and those who succeeded in bypassing it. Due to the fact that the sampling procedure did not allow for a random selection of participants, respondents to the survey are not necessarily representative of Brazil’s online population. Our sample is younger and more educated, and women are slightly overrepresented.

 

++++++++++

Variables

Blockage bypass. Among those who did attempt to bypass the blockage, we asked respondents whether they were able to successfully overcome the blockage (Yes = 71 percent), mostly by downloading a special VPN app that simulates the access from a different country.

Emotional arousal. Respondents were asked whether they felt anxiety (Yes = 33 percent), anger (Yes = 40 percent), isolation (Yes = 13 percent), despair (Yes = 29 percent), and loneliness (Yes = 15 percent) when they knew about the blockage. The literature about this topic claims that physiological arousal of emotions responds to patterns of cognitive appraisals (Smith and Ellsworth, 1985). So, while high levels of emotional arousal are associated with activation, or action-oriented behavior, low levels are associated with deactivation or relaxation (Feldman Barrett and Russell, 1998; Berger and Milkman, 2012). Following this literature, we grouped the above items into two variables: isolation (including isolation, despair and loneliness; Cronbach’s α = .61) and anxiety (including anxiety and anger; inter-item correlation r = .24).

Digital skills. Respondents were asked how familiar they were with a series of concepts related to digital technologies developed and rigorously validated by Hargittai and Hsieh (2012), on a five-point Likert scale where 1 = not familiar at all, to 5 = very familiar: “Advanced search,” “Spyware,” “Wikipedia,” “Favorites,” “JPG,” “Blog,” “Malware,” “Bookmark,” “Tag,” “Firewall” and “Podcast.” These items were combined into an index of digital skills (11 items, Cronbach’s α = .94, M = 2.9, SD = 1.2, range = 1 to 5).

Frequency of WhatsApp use. Respondents were asked how often they used WhatsApp on a five-point Likert scale where 1 = seldom, to 5 = all the time (M = 4.7, SD = .63, range = 1 to 5).

WhatsApp skills. This variable measured different activities users can perform on WhatsApp. Respondents answered whether they were able to send text messages (Yes = 98 percent), send stock photos and images (Yes = 93 percent), use icons/emojis (Yes = 90 percent), send audio messages (Yes = 89 percent), send videos (Yes = 86 percent), take direct photos through WhatsApp (Yes = 78 percent), forward other content received on the phone (Yes = 78 percent), create groups (Yes = 78 percent), mute groups (Yes = 62 percent), empty chats (Yes = 63 percent), change sounds (Yes = 61 percent), archive chats (Yes = 73 percent) and share location (Yes = 55 percent). These items were combined into an additive index of WhatsApp skills (11 items, Cronbach’s α = .92).

WhatsApp gratifications. Respondents were asked how often they used WhatsApp for conversation, for entertainment, for problem solving, to work, to debate politics, to organize events and to keep in touch with family/friends, on a five-point Likert scale where 1 = never, to 5 = all the time. We performed an exploratory factor analysis with the items described above. The items loaded in two factors we labeled as WhatsApp for work (for problem solving, to work, to debate politics, to organize events) and WhatsApp for joy (for conversation, for entertainment, to keep in touch with family/friends). Items in each factor were combined to build two indexes of WhatsApp gratifications — work (4 items; Cronbach’s α = .63, M = 2.6, SD = .82, range = 1 to 5), and joy (3 items; Cronbach’s α = .41, M = 4.0, SD = .73, range = 2 to 5).

Demographics. Respondents were asked their gender (Male = 41 percent), their age (M = 21.3, SD = 7.7) and their highest level of education, from a list that ranged from “elementary school” to “college degree” (median = 4, high school).

Analyses

To test H1, H2, H3, and H5, independent sample t-tests were performed to compare the means of those who did overcome the WhatsApp blockage in Brazil with the means of those who did not. To test H4, a chi-square test was run to measure whether bypassing the blockage was associated with gender. To answer RQ1, a two-step cluster analysis was performed using blockage bypass, digital skills and emotional arousal as the variable criteria to classify cases. Clusters were compared to each other in terms of gender, age, education, frequency of WhatsApp use, WhatsApp skills and WhatsApp gratifications (work and joy).

 

++++++++++

Results

H1 suggested that emotional arousal would positively affect overcoming the WhatsApp blockage. In other words, users who felt high levels of a) anxiety and b) isolation would be more likely to bypass the blockage. Results from the t-tests show that users who were able to use WhatsApp during the blockage presented significantly higher levels of anxiety (M = .78, SD = .75) than those who did not overcome the blockage (M = .61, SD = .65) (t = 1.89, p<.05). As such, H1a is supported. However, we did not see this pattern for isolation, as differences between those who did and did not bypass the blockage were not significant. Therefore, H1b is not supported.

H2 suggested that digitally skilled users would be more likely than less skilled users to overcome the WhatsApp blockage. Results from t-tests show that users overcoming the blockage were significantly more skilled (M = 3.0, SD = 1.2) than those who could not install a VPN software (M = 2.7, SD = 1.2) to use WhatsApp (t = 2.05, p<.05). Consequently, H2 is supported.

H3 proposed that education would positively affect users’ ability of overcoming the blockage in Brazil. Results from t-tests show this was not the case, as users who did and did not bypass the blockage had similar levels of education (M = 4 and M = 3.7, respectively). Thus, H3 is not supported.

H4 proposed that males would be more likely than females to overcome the WhatsApp blockage. Results from the chi-square test indicate a significant relationship between gender and overcoming the blockage (χ2(1) = 8.26, p<.01), showing that 81 percent of the males in the sample were able to bypass the blockage, while 65 percent of the females were able to do the same. As such, H4 is supported.

H5 suggested that age would negatively affect users’ ability of overcoming the blockage. Results from t-tests indicate that age did not make a difference, as the average age of those who overcame the blockage was very similar to the age of those who did not (M = 21.1 years old versus M = 21.7 years old, respectively). Consequently, H5 is not supported.

RQ1 asked how WhatsApp users classify depending on their digital skills, arousal and ability to bypass the blockage. Cluster analyses revealed respondents cluster into four groups based on the criteria variables: the deprived, the addicted, the challengers and the elite, as illustrated in Table 1. We compared these four groups regarding their levels of WhatsApp use (in terms of gratifications and frequency), WhatsApp skills and demographics (gender, age, and education).

 

Table 1: Two-step cluster analysis summary.
Variable criteriaClusters
 The deprivedThe addictedThe challengersThe elite
Bypassing WhatsApp blockage
(cells report categories)
NoYesYesYes
Digital skills
(cells report means on a scale from 1–5)
2.702.872.234.36
Arousal — Isolation
(cells report means on a scale from 0–3)
0.641.530.250.17
Arousal — Anxiety
(cells report means on a scale from 0–2)
0.611.820.520.35
Cluster size29%17%33%21%

 

The first cluster accounts for 29 percent of respondents. They were not able to bypass the blockage, showed high levels of anxiety, and their digital skills were lower than the average user (see Figure 1). Probably because of their lack of skills, they could not install and/or operate a VPN software, despite their attempt to do so. We labeled this group as the deprived.

There were significantly more women in this cluster (73 percent) than in other clusters. No differences in terms of age, WhatsApp gratifications (work or joy), WhatsApp skills, and frequency of WhatsApp use were found as compared to other clusters.

 

The deprived cluster
 
Figure 1: The deprived cluster.
Note: Larger version of Figure 1 available here.

 

The second cluster accounts for 17 percent of respondents. This group comprises respondents who were able to bypass the blockage, had average skills and high levels of both anxiety and isolation (see Figure 2). Even though they werent the savviest, they probably surpassed the blockage because of their high levels of emotional arousal when they found out about the blockage. Despite their average digital skills, this cluster was significantly more skilled than other clusters in terms of WhatsApp skills and had the highest frequency of WhatsApp use (4.9 on a scale from 1 to 5). We labeled this group as the addicted due to their intensive use and knowledge of the app.

This cluster had more males (55 percent) than females and was the youngest of the four clusters (19 years old on average). No differences were found in terms of WhatsApp gratifications or education.

 

The addicted cluster
 
Figure 2: The addicted cluster.
Note: Larger version of Figure 2 available here.

 

The third cluster accounts for 33 percent of respondents and is the largest group. Users in this cluster are very similar to users in the deprived cluster — less skilled than the average user and with high levels of anxiety (see Figure 3). Yet, users in this group did overcome the blockage. We labeled them as the challengers. Just like the deprived, the challengers also had more females than males (64 percent) but no other differences were found between challengers and deprived.

 

The challengers cluster
 
Figure 3: The challengers cluster.
Note: Larger version of Figure 3 available here.

 

Finally, the fourth cluster accounts for 21 percent of respondents. These users were able to overcome the blockage, were highly skilled and presented low levels of both anxiety and isolation. Probably because they knew their digital skills would allow them to bypass the blockage, their emotional arousal was far below the mean (see Figure 4). We labeled this group as the elite. This is the cluster with more males (56 percent, similar to the addicted), had older users (23 years old on average) and had the highest level of education.

 

The elite cluster
 
Figure 4: The elite cluster.
Note: Larger version of Figure 4 available here.

 

 

++++++++++

Discussion

This study aimed to understand reasons behind the success of failure of Brazilian WhatsApp users who tried to bypass a forceful legal 12-hour deprivation of access to WhatsApp in Brazil in December 2015. Our findings indicate the blockage generated important levels of anger, isolation and anxiety in the population, which turned out in users attempting to overcome the blockage by installing VPN software to use WhatsApp despite the prohibition.

Yet, not all emotions work in the same way, and emotions might not be enough to overcome the problem. Findings indicate that anxiety was a more powerful catalyst than isolation, but even those with higher levels of anxiety were not able to bypass the blockage if their digital skills were lower than the average user, as it was the case for the deprived.

The classic factors of digital divide

Results regarding digital skills were consistent with previous literature, since they are a good predictor of successfully bypassing the blockage. Our findings indicate, though, that low or average levels of digital skills are not necessarily a predictor of failing to bypass; they may be prompted by high levels of emotional arousal, in the case of anxiety and anger, as best illustrated by the addicted.

In terms of gender, the results of this study are consistent to previous research about digital divide and acquisition of digital skills. First, we observed that males were significantly more likely than women to overcome the blockage. This pattern was evident in the cluster analysis as well — the only cluster not able to successfully bypass the blockage (the deprived) was mainly composed by women, while the cluster dominated by men (the elite) bypassed it supposedly with a great degree of ease.

Such findings add up even when considering digitization prior to Internet, mobile Internet and smartphones. Ever since the diffusion of computers in society, differences in gender have been detected in distinct areas such as sex-role stereotyping, self-efficacy and positive affect (Whitley, 1997), always favorable to men. As suggested by Hilbert (2011), the gender divide could be a spurious effect derived from the other structural divides to which women are subject.

Other demographics showed no effect on overcoming the blockage — both education and age did not make a difference. Yet, these variables might be affecting the dependent variable through digital acquisition or might be the outcome of a sample slightly skewed in terms of age. Though we had variability (15–68), our sample was skewed toward younger users and that might be the cause of age not being a significant factor in the equation. Education, on the other hand, might act through a different route. The elite cluster overcame the blockage with the lowest levels of emotion arousal and highest levels of digital skills and presented the highest levels of education. Then, education could be crucial to acquire skills that allowed users to bypass the blockage with almost no anxiety/isolation feelings.

The clusters: Who did what and how?

Despite low levels of emotional arousal, the elite not only did attempt to bypass the blockage, but they also succeeded. We believe this happened because of a perception of self-efficacy: they knew they would be able to bypass the blockage and therefore were not significantly aroused. As such, bypassing the blockage was a low-cost behavior to them.

The addicted were users with average digital skills but are extremely aroused in terms of both isolation and anxiety. These factors, combined with the intense use, frequency and skills related to the app, probably explain their emotional reaction. Such users might have a special relationship with the app, either by option or as an economic limitation, as mobile Internet access plans in Brazil frequently offer zero-rating [10] for WhatsApp data traffic.

Those with lower digital skills were grouped in two clusters: the deprived and the challengers. They were both very similar, except for subtle differences. The challengers cluster had fewer women (65 percent versus 73 percent), while the deprived were a little better qualified regarding digital skills.

 

++++++++++

Conclusion

Impromptu censorship of an IM App in a democratic country surprises users. This is a distinguishing trait of the present study in comparison with censorship circumvention literature. Such context leads to incident-driven strategies (Gebhart and Kohno, 2017) such as search for tools and learning new skills to bypass the blockage. But skill acquisition in digital environments is a complex phenomenon involving many variables. In the case of this study, we observed that digital skills and emotional arousal were crucial to trigger behaviors toward the WhatsApp blockage in Brazil. The combination of such variables resulted in four clusters of users: the deprived, who failed to overcome the blockage; and the addicted, the challengers and the elite, who successfully bypassed it.

As such, our study is not without limitations. First, we worked with a convenience sample to prioritize immediacy over representation. Second, we measured emotional arousal using self-reported answers, though several studies have documented respondents’ self-reports may be inaccurate because of factors such as perception bias or memory loss (Holbrook, 2008). Third, we did not consider macro structural variables that could also explain the phenomenon observed here — working conditions and social status, income, household position, location (rural-urban) — as well as accurate measures of social capital. We invite other scholars to include these factors when working on similar topics. Future research could also consider with greater detail the structural conditions of access (such as ISP contracts to access Internet) and the positional relations (van Dijk, 2005) of the subjects who use digital communication technologies. Additionally, it would be valuable to reinforce the findings with a qualitative design to explore narrated experiences of those deprived from ICT, filling in some of the gaps left by survey data.

This research adds important insights to the limited literature on technology deprivation and the digital inequalities field. The present paper highlights relevant patterns of user characteristics, emotions and behaviors in face of an extremely unexpected event at that moment: the forceful deprivation of more than 100 million WhatsApp users for half a day by the Brazil legal system. In a nutshell, our study shows that digital skills matter, but they work in conjunction with people’s emotions and needs. End of article

 

About the authors

Marcelo Santos (Ph.D., PUC Chile, 2018) is a researcher at CIDOC and an assistant professor in the School of Communications at Universidad Finis Terrae, Chile. His research is in the crossroads of ICT with democracy, including digital inequalities, ICT social appropriation, datafication of the society among others.
Direct comments to: msantos [at] uft [dot] cl

Magdalena Saldaña (Ph.D., University of Texas at Austin, 2017) is an assistant professor in the School of Communications at Pontificia Universidad Católica de Chile, and a researcher at the Millennium Institute for Foundational Research on Data (IMFD). Her research interests include political communication, public opinion, digital and social media and Latin American studies.
E-mail: magdalena [dot] saldana [at] uc [dot] cl

Andrés Rosenberg (M.A., Autonomous University of Barcelona) is a Chilean-Spanish journalist and Ph.D. candidate at Pontificia Universidad Católica de Chile. He also holds a Master’s degree in communication and education from the Autonomous University of Barcelona. His research interests include uncivil speech in online communication, ICTs and political communication.
E-mail: aarosenb [at] uc [dot] cl

 

Acknowledgments

The authors are grateful for the contributions of Angela Brandão, Sebastián Valenzuela and Rayén Condeza, who helped build the survey, and both the special issue editors and the reviewers for their valuable contributions to this article. Additionally, the present work benefited from internal funding CAI 2018, from University Finis Terrae.

 

Notes

1. http://bloqueios.info/en/casos/block-for-non-compliance-with-judicial-requests-for-user-data-2/, accessed 20 June 2019.

2. As West clearly puts it, he identified “temporary shutdowns of the entire Internet (nationally or locally), temporary shutdowns of the mobile Internet (nationally or locally), and temporary blocking of specific applications and/or services (nationally or locally)” (West, 2016, p. 4).

3. Yang and Liu, 2014, p. 249.

4. Mou, et al., 2016, p. 841.

5. Howard, et al., 2011, p. 217.

6. Mou, et al., 2016, p. 838.

7. Comitê Gestor da Internet no Brasil, 2018, p. 120.

8. Comitê Gestor da Internet no Brasil, 2018, p. 117.

9. Hargittai, 2002, p. 8.

10. By December 2017, there where 58 different plans to access Internet that offered zero-rating for some apps, most of them including WhatsApp, due to its popularity (Gragnani, 2018). This means that data exchanged in the selected apps do not account for the data plan consumption limits.

 

References

Access Now, 2018. “Internet shutdowns in context” (13 December), at https://www.accessnow.org/keepiton/, accessed 10 June 2019.

Jonah Berger and Katherine L. Milkman, 2012. “What makes online content viral?” Journal of Marketing Research, volume 49, number 2, pp. 192–205.
doi: https://doi.org/10.1509/jmr.10.0353, accessed 11 December 2019.

Comitê Gestor da Internet no Brasil, 2018. 2017 ICT households: Survey on the use of information and communication technologies in Brazilian households. São Paulo: Núcleo de Informação e Coordenação do Ponto BR.

Francesca Comunello and Giuseppe Anzera, 2012. “Will the revolution be tweeted? A conceptual framework for understanding the social media and the Arab Spring,” Islam and ChristianMuslim Relations, volume 23, number 4, pp. 453–470.
doi: https://doi.org/10.1080/09596410.2012.712435, accessed 11 December 2019.

Joel Cooper and Kimberlee D. Weaver, 2003. Gender and computers: Understanding the digital divide. Mahwah, N.J.: Lawrence Erlbaum Associates.

Teresa Correa, 2016. “Digital skills and social media use: How Internet skills are related to different types of Facebook use among ‘digital natives’,” Information, Communication & Society, volume 19, number 8, pp. 1,095–1,107.
doi: https://doi.org/10.1080/1369118X.2015.1084023, accessed 11 December 2019.

Teresa Correa and Isabel Pavez, 2016. “Digital inclusion in rural areas: A qualitative exploration of challenges faced by people from isolated communities,” Journal of Computer-Mediated Communication, volume 21, number 3, pp. 247–263.
doi: https://doi.org/10.1111/jcc4.12154, accessed 11 December 2019.

Teresa Correa, Joseph D. Straubhaar, Wenhong Chen and Jeremiah Spence, 2015. “Brokering new technologies: The role of children in their parents’ usage of the Internet,” New Media & Society, volume 17, number 4, pp. 483–500.
doi: https://doi.org/10.1177/1461444813506975, accessed 11 December 2019.

Paul DiMaggio and Eszter Hargittai, 2001. “From the ‘digital divide’ to ‘digital inequality’: Studying Internet use as penetration increases,” Princeton University, Center for Arts and Cultural Policy Studies, Working Paper Series, number 15, and at https://culturalpolicy.princeton.edu/sites/culturalpolicy/files/wp15_dimaggio_hargittai.pdf, accessed 11 December 2019.

Pavel Durov, 2018. “VK wall post” (29 April), at https://vk.com/tnews?w=wall1_2442097, accessed 19 November 2019.

Ashley Esarey and Qiang Xiao, 2011. “Digital communication and political change in China,” International Journal of Communication, volume 5, pp. 298–319, and at https://ijoc.org/index.php/ijoc/article/view/688/525, accessed 11 December 2019.

Lisa Feldman Barrett and James A. Russell, 1998. “Independence and bipolarity in the structure of current affect,” Journal of Personality and Social Psychology, volume 74, number 4, pp. 967–984.
doi: https://doi.org/10.1037/0022-3514.74.4.967, accessed 11 December 2019.

Genevieve Gebhart and Tadayoshi Kohno, 2017. “Internet censorship in Thailand: User practices and potential threats,” 2017 IEEE European Symposium on Security and Privacy (EuroS&P), pp. 417–432.
doi: https://doi.org/10.1109/EuroSP.2017.50, accessed 11 December 2019.

Melissa Gilbert, 2010. “Theorizing digital and urban inequalities: Critical geographies of ‘race’, gender and technological capital,” Information, Communication & Society, volume 13, number 7, pp. 1,000–1,018.
doi: https://doi.org/10.1080/1369118X.2010.499954, accessed 11 December 2019.

Helton S. Gomes, 2018. “Brasil tem 116 milhões de pessoas conectadas à internet, diz IBGE,” G1 (21 February), at https://g1.globo.com/economia/tecnologia/noticia/brasil-tem-116-milhoes-de-pessoas-conectadas-a-internet-diz-ibge.ghtml, accessed 6 June 2019.

Juliana Gragnani, 2018. “Como planos de celular com Facebook e WhatsApp ilimitados podem potencializar propagação de notcias falsas,” BBC News Brasil (16 April), at https://www.bbc.com/portuguese/brasil-43715049, accessed 6 June 2019.

Eszter Hargittai, 2002. “Second-level digital divide: Differences in people’s online skills,” First Monday, volume 7, number 4, at https://firstmonday.org/article/view/942/864, accessed 11 December 2019.
doi: https://doi.org/10.5210/fm.v7i4.942, accessed 11 December 2019.

Ezster Hargittai and Yuli Patrick Hsieh, 2012. “Succinct survey measures of Web-use skills,” Social Science Computer Review, volume 30, number 1, pp. 95–107.
doi: https://doi.org/10.1177/0894439310397146, accessed 11 December 2019.

Navid Hassanpour, 2014. “Media disruption and revolutionary unrest: Evidence from Mubarak’s quasi-experiment,“ Political Communication, volume 31, number 1, pp. 1–24.
doi: https://doi.org/10.1080/10584609.2012.737439, accessed 11 December 2019.

Martin Hilbert, 2011. “Digital gender divide or technologically empowered women in developing countries? A typical case of lies, damned lies, and statistics,” Women’s Studies International Forum, volume 34, number 6, pp. 479–489.
doi: https://doi.org/10.1016/j.wsif.2011.07.001, accessed 11 December 2019.

Allyson Holbrook, 2008. “Self-reported measure,” In: Paul J. Lavrakas (editor). Encyclopedia of survey research methods. Thousand Oaks, Calif.: Sage.
doi: http://dx.doi.org/10.4135/9781412963947.n523, accessed 11 December 2019.

Philip N. Howard, Sheetal D. Agarwal and Muzammil M. Hussain, 2011. “The dictators’ digital dilemma: When do states disconnect their digital networks?” Issues in Technology Innovation, number 13, at https://www.brookings.edu/wp-content/uploads/2016/06/10_dictators_digital_network.pdf, accessed 11 December 2019.

Instituto Brasileiro de Geografia e Estatistica (IBGE), 2018. “Acesso à Internet e à televisão e posse de telefone móvel celular para uso pessoal 2016,” at https://biblioteca.ibge.gov.br/visualizacao/livros/liv101543.pdf, accessed 11 December 2019.

Simin Kargar and Keith McManamen, 2018. “Censorship and collateral damage: Analyzing the Telegram ban in Iran,” >Berkman Klein Center, Research Publication, number 2018–4, at https://cyber.harvard.edu/publication/2018/censorship-and-collateral-damage, accessed 11 December 2019.

Mora Matassi, Pablo Boczkowski and Eugenia Mitchelstein, 2019. “Domesticating WhatsApp: Family, friends, work, and study in everyday communication,” New Media & Society, volume 21, number 10, pp. 2,183–2,200.
doi: https://doi.org/10.1177/1461444819841890, accessed 11 December 2019.

Luiz Mazzeto, 2017. “Com 120 milhões de usuários, WhatsApp vive momento crucial no Brasil,” IDG Now (31 May), at http://www.edicaoms.com.br/tecnologia/com-120-milhoes-de-usuarios-whatsapp-vive-momento-crucial-no-brasil, accessed 24 August 2017.

Karen Mossberger, David Kaplan and Michele A. Gilbert, 2008. “Going online without easy access: A tale of three cities,” Journal of Urban Affairs, volume 30, number 5, pp. 469–488.
doi: https://doi.org/10.1111/j.1467-9906.2008.00414.x, accessed 11 December 2019.

Yi Mou, Kevin Wu and David Atkin, 2016. “Understanding the use of circumvention tools to bypass online censorship,” New Media & Society, volume 18, number 5, pp. 837–856.
doi: https://doi.org/10.1177/1461444814548994, accessed 11 December 2019.

David Nemer, 2018. “The three types of WhatsApp users getting Brazil’s Jair Bolsonaro elected,” Guardian (25 October), at https://www.theguardian.com/world/2018/oct/25/brazil-president-jair-bolsonaro-whatsapp-fake-news, accessed 11 December 2019.

Opinião Pública, 2018. “24% dos eleitores usam Whatsapp para compartilhar contedo eleitoral,” Datafolha (27 October), at http://datafolha.folha.uol.com.br/opiniaopublica/2018/10/1983765-24-dos-eleitores-usam-whatsapp-para-compartilhar-conteudo-eleitoral.shtml accessed 20 June 2019.

Vlad Savov, 2018. “Russia’s Telegram ban is a big, convoluted mess,” The Verge (17 April), at https://www.theverge.com/2018/4/17/17246150/telegram-russia-ban, accessed 20 November 2019.

Craig A. Smith and Phoebe C. Ellsworth, 1985. “Patterns of cognitive appraisal in emotion,” Journal of Personality and Social Psychology, volume 48, number 4, pp. 813–838.
doi: https://doi.org/10.1037/0022-3514.48.4.813, accessed 11 December 2019.

Statista, 2019. “Share of population in selected countries who are active WhatsApp users as of 3rd quarter 2017” (22 July), at https://www.statista.com/statistics/291540/mobile-internet-user-whatsapp/, accessed June 6 2019.

Jan A.G.M. van Dijk, 2005. The deepening divide: Inequality in the information society. Thousand Oaks, Calif.: Sage.

Jan A.G.M. van Dijk and Alexander J. van Deursen, 2014. Digital skills: Unlocking the information society. New York: Palgrave Macmillan.
doi: https://doi.org/10.1057/9781137437037, accessed 11 December 2019.

Darren M. West, 2016. “Internet shutdowns cost countries $2.4 billion last year,” Center for Technological Innovation at Brookings, at https://www.brookings.edu/wp-content/uploads/2016/10/intenet-shutdowns-v-3.pdf, accessed 11 December 2019.

Bernard E. Whitley, Jr., 1997. “Gender differences in computer-related attitudes and behavior: A meta-analysis,” Computers in Human Behavior, volume 13, number 1, pp. 1–22.
doi: https://doi.org/10.1016/S0747-5632(96)00026-X, accessed 11 December 2019.

Qinghua Yang and Yu Liu, 2014. “What’s on the other side of the great firewall? Chinese Web users’ motivations for bypassing the Internet censorship,” Computers in Human Behavior, volume 37, pp. 249–257.
doi: https://doi.org/10.1016/j.chb.2014.04.054, accessed 11 December 2019.

 

Appendix

Survey summary — questions used in this study:

  1. Were you able to keep using WhatsApp during the blockage? Yes/No
  2. When the blockage occurred, did you feel: anxiety (Yes/No), anger (Yes/No), isolation (Yes/No), despair (Yes/No), loneliness (Yes/No).
  3. On a scale where 1 = not familiar at all, to 5 = very familiar, how familiar do you feel with the following concepts:
     12345
    Advanced search     
    Spyware     
    Wikipedia     
    Favorites     
    JPG     
    Blog     
    Malware     
    Bookmark     
    Tag     
    Firewall     
    Podcast     
  4. On a scale where 1 = seldom to 5 = all the time, how often do you use WhatsApp?
  5. When using WhatsApp, can you:
    • Send text messages (Yes/No)
    • Use icons/emojis (Yes/No)
    • Send stock photos and images (Yes/No)
    • Take photos directly through WhatsApp (Yes/No)
    • Send videos (Yes/No)
    • Send audio messages (Yes/No)
    • Forward other content received on the phone (Yes/No)
    • Create groups (Yes/No)
    • Share location (Yes/No)
    • Archive chats (Yes/No)
    • Empty chats (Yes/No)
    • Change sounds (Yes/No)
    • Mute groups (Yes/No)
  6. On a scale where 1 = never to 5 = all the time, how often do you use WhatsApp for the following activities:
     12345
    For conversation     
    To work     
    To debate politics     
    to be informed     
    For entertainment     
    To organize events     
    To keep in touch with family/friends     
    For problem solving     
  7. What’s your gender? Female/Male
  8. What year were you born?
  9. What’s you highest level of education?
    • Elementary school (incomplete)
    • Elementary school
    • High school (incomplete)
    • High school
    • College (incomplete)
    • College degree

 


Editorial history

Received 26 November 2019; accepted 8 December 2019.


Creative Commons License
This paper is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

From access deprivation to skill acquisition: Cluster analysis of user behavior in face of a 12-hour legal blockage of WhatsApp in Brazil
by Marcelo Santos, Magdalena Saldaña, and Andrés Rosenberg.
First Monday, Volume 25, Number 1 - 6 January 2020
https://firstmonday.org/ojs/index.php/fm/article/view/10401/8316
doi: http://dx.doi.org/10.5210/fm.v25i1.10401





A Great Cities Initiative of the University of Illinois at Chicago University Library.

© First Monday, 1995-2020. ISSN 1396-0466.