Note: This post is a response to Book Review: “Digitally Enabled Social Change: Activism in the Internet Age”

by Jennifer Earl (Professor of Sociology, University of California, Santa Barbara)

I appreciate the chance that Mary has given us to reply to her critical review of our book, Digitally Enabled Social Change (2011, MIT Press). Given the tone of Mary’s review, I think it is helpful to first step back and notice that there are many things on which Mary and I agree. Indeed, Mary ends her review with a laundry list of things she liked about our book, some of which are quite important themes. For example, she agrees with our arguments about the changing infrastructure of movements—which may seem like a simple point to her but is one that in some ways upends four decades of social movement scholarship. She also agrees with our argument about the likelihood of episodic activism, which again may seem minor to her but would represent a fundamental break in our academic understanding of social movements across two centuries. But, at an even bigger level, and perhaps most importantly, Mary and I both think digital activism is important and that people (activists and scholars alike) should pay more attention to it.

Where Mary and I diverge is in how you forward an agenda about raising the profile of digital activism. That divergence in large part owes to our expected audiences—Katrina and my audience is academic; we are trying to make a case to social movements researchers, who as a group have been exceedingly skeptical of digital activism. It has been an uphill battle to get social movement scholars to consider the possibility that digital activism has different dynamics and that studying those dynamics is important. Mary’s is a technology-rich audience where utopian visions of technology are as common as skepticism. Our primary audience uses email; hers tweets. Our primary audience is obsessed with the quality of research methods, theories of causality, and academic rules of evidence. Hers is obsessed with cutting edge technologies. So, it is understandable that despite common orientations to digital activism overall, we end up with very different means of forwarding that agenda. With this as background, let’s turn to Mary’s main concerns:

Why study online petitions, boycotts, and email and letter-writing campaigns?

Mary takes issue with our empirical focus on these tactics for a variety of reasons (indeed, if you read carefully, this is her biggest beef with our work), and certainly, if you don’t want to read about these kinds of online tools, you might take issue with us too. But, instead of critiquing our book based on the book you wish we had written, let’s discuss the merits of the book that was written. We focus on these tactics for several reasons.

First, as Mary points out, these are online incarnations of offline tactics. Although Mary takes this on its face as a negative, we think it actually gives us a lot of helpful research leverage. From a research methods perspective, it allows us to very precisely isolate the impact of action taking place online versus offline because we know how these specific tactics have worked offline in the past and can use that as a baseline. Also, by limiting the only source of variation to whether the action is taking place online or off, we immediately eliminate a ton of other causal explanations for what we find. For instance, if we had chosen other online protest forms, like the Google bomb that Mary mentions, critical social movement scholars would have been able to assume that the differences between on and offline activism we find owe to the exoticism of the tactic, not to its online elements.

Second, Mary argues that the tactics we study are least likely to showcase novel action. We agree—this is a chief reason we chose them! For an academic audience, choosing the hardest target and then still proving your point is a huge bonus, not a criticism: that we find important differences between online and offline tactics in places where you might expect those differences to not exist or to be minimal is our argument. Indeed, choosing a venue where you are mostly likely to be wrong and then testing your theory is a hallmark of good social science—despite what many people think, social scientists should try to avoid “cooking the books” in their favor through their selection of cases.

Third, we chose these tactics because they seem to be everywhere online. Mary asserts (without data, much as she accuses us of doing) that these tactics are not the most common online forms of action. Perhaps these are not the most common tactics in the complex world of the Obama campaign or in a training session for experienced digital activists, but in the everyday world where my aunt and her friends are looking to participate online, online petitions, boycotts, letter-writing and emailing campaigns are where it is at. And, while this book doesn’t present data on the frequency of these tactics versus other kinds of tactics (you got us there, Mary), I have recently finished a 5-year study of 20 social movement arenas and can tell you that those data conclusively show that the tactics Katrina and I are studying in this book are the most frequent online tactics. You can check out some early results from that study in the December 2010 issue of Mobilization. Later papers from this study will confirm what Mary thinks must be wrong: even in 2010, online petitions, boycotts, and email and letter-writing campaigns are very popular online.

Fourth, Mary objects to the lack of focus on social media and related social media tactics, arguing that the tactics we study are stale in digital terms. But, as I just mentioned, later work of mine shows that these are not stale—they are quite popular even today. And, my more recent data collection shows that the social media tactics Mary thinks are so prevalent still make up a very small share of the overall online protest universe. While social media maybe the shiniest thing out today, it’s not the only or the most empirically frequent form of online protest. Moreover, while we are looking at data from 2004, we don’t believe that the theoretical principals we are trying to illustrate with that data are much different in 2011 from 2004. Indeed, my current work is testing precisely that hypothesis. I also think there is another audience issue at work here: academics understand that writing a well-researched book and getting it through the academic publishing process takes a few years. Mary’s audience is now, new, next. But just because something took place yesterday, doesn’t mean it’s not instructive about today and tomorrow.

Finally, Mary suggests our tactics are not representative of the online universe. Here I could not strenuously disagree more and hope that readers will judge this for themselves. The methods that Katrina and I use are unique in that they actually give us a better chance of charting a representative population of online petitions, boycotts, and email and letter-writing campaigns. Check out the methods and decide for yourself. As for whether they are representative of the most common forms online, see my response above—new work shows, yes, they are. While Mary’s anecdotes about the popularity of other forms are interesting, they aren’t accurate in painting a larger view. Some of Mary’s examples certainly have gotten a lot of news coverage—but they are outliers in both their public notoriety and their novelty.

Are you claiming that some kinds of online protest are “better” than others?

No. I think this is a place where Mary misreads us. For whatever part of that Katrina and I are responsible for as authors, let me set the record straight. In a nutshell, we argue that theoretical processes that have been developed over the past 50 years to explain activism are only good guides for the theoretical processes driving some kinds of online activism today, and those theoretical expectations don’t lead us in the right direction for other kinds of online activism. Try this analogy: 50 years of research has shown that engines are based on combustion. We are saying that for some kinds of online protest, the engine is all electric (proverbially speaking). We are not claiming that an electric engine is better; we are saying it operates differently from a classic combustion engine. More precisely, we are saying that when people take advantage of and leverage two unique features of Internet-enabled technologies—low costs and coordinated action without co-presence—the theoretical explanations for participation and organizing change. When they don’t leverage these unique features, the engine stays the same. We are not claiming one is better, just that they are different. For Mary, those theoretical differences might not matter much. For social movement scholars, they are critically important.

Why didn’t you study which online protests were effective?

From her review, it appears Mary wanted a book that empirically examined which kinds of online protest are more or less effective. We don’t advertise such a book, nor did we write one. I think for a variety of methodological reasons, it will be a long time before someone writes the book that Mary wants in this regard, or at least writes one that, from a research methods point of view, I would also want to read. It turns out that the study of offline protest faces the same problem: it turns out to be very difficult to prove, from a social scientific perspective, which offline actions or movements are effective. I gave a paper on this very topic in Berlin this summer and would be happy to share it with people who email me.

Other Quibbles

Mary had other quibbles with the book. She thought we should have mathematically tested whether organizers’ time exactly conformed to a power law. We didn’t see that as necessary because even something that looks close to a power law—which we do show—is a very radical departure from what would be anticipated by social movement scholars. If it is off by a hair, it doesn’t really matter to the arc of our argument because it’s still in the ballpark. She wishes we didn’t use the terminology of e-tactics, preferring digital protest. I am hoping much of both of our audiences can get past such semantic differences in style.

So, where does this leave us?

As I said in the opening, Mary and I actually have pretty similar agendas: we both think people should pay more attention to digital activism. In my case, I want social movement scholars to dig more deeply into our theoretical approaches so we can figure out when and how protesting online differs—at a theoretical process level—from protesting offline. I also want Internet scholars to have to seriously engage with the literatures that have been developed around relevant offline areas of social life instead of engaging in drive-by theorizing that doesn’t connect with different areas’ rich research traditions. I think Katrina and my book gets scholars off to a good start on both of these endeavors. I hope readers will judge for themselves and I am confident that most will enjoy the book much more than Mary did.

Note: This post is followed by a response from the authors.

Broad-based empirical studies are sorely needed to understand the effect of digital technology on contentious politics, a field where battling anecdotes predominate. Jennifer Earl and Katrina Kimport’s new book, Digitally Enabled Social Change: Activism in the Internet Age attempts to fill this need. However, the book has several serious weaknesses in its methodology, theoretic goals, and conclusions.

The authors clearly wish to find disruptive new phenomena, evidence that digital activism challenges traditional social movement theory. Yet they choose to study digital tactics that are merely digital incarnations of analog ones and focus on outdated tactics and applications while ignoring tactics which are more relevant to current practices of digital activism. This desire to come to a certain theoretical conclusion leads the authors to make intellectually appealing yet practically unsound conclusions about how to evaluate digital tactics.

Outdated Tactics

In 2004, Jennifer Earl of the University of California at Santa Barbara and Katrina Kimport of the University of California at San Francisco created a data set composed of four types of digital tactics: petitions, letter-writing campaigns, email campaigns, and boycotts. The tactics, identified through a robust and innovative use of Google search, were split between “warehouse” sites like PetitionOnline.com (362 tactics), and non-warehouse sites, like personal blogs (748 tactics).

Already some problems are apparent. First, no social media platforms are included in their data set. Based even on a casual review of news headlines, social media seems extremely salient in “activism in the Internet age,” the avowed topic of the book. Social media is not discussed at length until the very last page of the conclusion, and is discussed in vague and platitudinous terms: “social networking sites like Facebook will encourage new uses and dynamics of online protest,” “networks could even be created around specific actions,” etc. (p 204). The disconnect is perhaps most visible in the index, which includes one entry and four sub-entries on “fax campaigns,” but not a single entry of any kind for “social networking” or “social media”.

Earl and Kimport do briefly acknowledge this shortcoming in the beginning of their book when they note that their “data are drawn from the period just preceding the rise of many dominant social networking Web sites” (p 27). However, this statement is problematic. First of all, it is not really true. The first wave social networks like Friendster (2002) and MySpace (2003) existed when they collected their data. Other social media platforms more important for activism, like Facebook (2004), YouTube (2005), and Twitter (2006), were founded soon after, and long before the publication of their book.

To draw conclusions about activism in the Internet age without reference to social media seems almost as negligent as drawing conclusions about weight gain without reference to carbohydrates. During the seven years between data collection and publication could they not have collected new data to reflect current practice? And, if such revision was impractical, could they not have couched their findings with a strong disclaimer? Publishing outdated material is a particular danger in this rapidly changing field, but downplaying or ignoring valid concerns about old data is more problematic.

What about the tactics they did choose: petitions, letter-writing campaigns, email campaigns, and boycotts? Earl and Kimport state that “most examples of relatively inexpensive activism online take the form of the e-tactics we examine in this book” (p 73), yet they offer no evidence for this claim. Did they do an earlier survey of a broad range of digital tactics and select the most popular to focus on? There is no evidence of this. Though anecdotal evidence supports the continuing popularity of email in digital campaigns, it’s hard to imagine that digital boycotts are as pervasive. After failures of high-profile e-petition sites like Number 10, the tactical value of e-petitions has also been questioned.1 In fact, Earl and Kimport acknowledge that “seven of the fifteen [warehouse sites] we studied are gone from the Web” (p 195), a fact that does not bode well for the current validity of their seven-year-old tactical data.

This focus on the pre-social media, Web 1.0 era of the Internet is also evident in their terminology: e-tactics, e-movements, e-mobilization. These terms are a throwback to the late nineties when the terms grew out of the word e-mail, short for “electronic mail.” Back then it was assumed that the principle difference between digital and paper media was that the former required electricity to be created and disseminated. We now have a more sophisticated understanding for how digital media is different from paper media, differences that hinge on the use of code (Lessig’s “perfect copies, freely made”) and network effects. (It is for this reason that I use the prefix “digital” instead of “e-” in this paper.)

Bottom Line: The four tactics selected for the study ignore the most popular current platforms, and have questionable representative salience.

Theory Before Evidence

Though Earl and Kimport’s book is founded on an empirical study, theirs is a deductive rather than inductive logic. They propose interesting theories yet their evidence does not really fit them. In other cases evidence they ignore challenges the validity of their theoretical claims.

There are the main theoretical arguments in the book:

• Tactics that maximally leverage digital affordances will lead to theory 2.0 changes in pre-digital social movement theory while tactics that only minimally leverage those affordances will lead to supersize changes where the mechanics of pre-digital social movement theory remains intact, only at a larger scale.
• That leveraged affordances should be the principal lens through which to evaluate digital tactics.
• That new types of online organizing structures exhibit a power-law function in that the most engaged organizers is twice as active as the second most active and so on.

Supersize / Theory 2.0

Earl and Kimport’s contention that the use of digital technology in activism will require the updating of pre-digital social movement theories is a good one. However, the evidence they provide does not support this contention. This is likely a result of the types of tactics they decided to study. Since they clearly have an interest in showing how digital tactics require the re-working of pre-digital theories, it is peculiar that they choose to focus on tactics that are merely digital forms of pre-digital tactics, a fact they acknowledge when they write that “each of these e-tactical forms has an offline progenitor and a long offline legacy” (p 202). These seem like the tactics least likely to reveal paradigm shifts. In fact, I often refer to e-petitions as the consummate example of a supersize tactic in that it achieves the exact same function as an offline petition – collecting signatures – only it allows more signatures to be collected more quickly and from a wider geographic area.

To find evidence of model change Earl and Kimport could have chosen tactics, like the Google-bomb, that have no pre-digital form. For example, in one high-profile instance of Google bombing, gay sex columnist Dan Savage led a campaign to redefine the name of conservative politician Rick Santorum as a “frothy mix” associated with anal sex, a tactic meant to publicly shame Santorum for his homophobia. An analog equivalent of this tactic – changing the definition of a word in every dictionary – simply doesn’t exist. They could also have studied tactics that don’t quite line up with their pre-digital precursors. For example, though DDoS attacks have been referred to as digital sit-ins, they are really more like ordering a million pizzas to arrive at MasterCard headquarters. These “false cognate” digital tactics could also reveal areas where old theory needs to be updated. The reason why Earl and Kimport chose supersize tactics in their effort to demonstrate theory 2.0 effects is never explained as they never reveal their process for selecting the tactics in the first place.

Leveraged Affordances

Their second theoretical point, that the leveraged affordances (fully utilizing capacities of digital tools) of digital technology is “critical to understanding Web activism” (p 177) is certainly true, but it is not as central as they make it out to be. In their analysis they equate maximal leveraging of affordances with “skill” and the altering of past social movement theory while they equate a lack of maximal leveraging with the junk food metaphor “supersizing” (p 177). They are drawing an implicit hierarchy of tactics here, with the tactics that maximally leverage digital affordances on top.

Earl and Kimport acknowledge that both types of tactics are likely to coexist in practice, which is clearly the case. Yet the reason they give for this integration is flawed:

Reality is likely to always be a mix of supersize effects and theory 2.0 effects because some people don’t notice key affordances, other don’t want to or can’t leverage them even if they do notice them, and still others notice and leverage these affordances quite skillfully (p 177).

In this analysis they set up a duality where skill is equated with leveraged affordances and ignorance, refusal, and lack of capacity are equated with a failure to leverage. However, they miss one reason why someone might choose not to maximally leverage digital affordances: they have noticed the affordance and understand it, but skillfully realize that a digital tactic will not be effective in their particular context. That is, they make a skillful decision not to maximally leverage digital affordances.

There are many examples of the phenomenon of skilled minimal leveraging. The case of the Stop Stock-Outs campaign, which took place in 2009 in the southern Africa countries of Kenya, Malawi, Uganda, Zambia and Zimbabwe, is one such example.2 The key digital tactic of the campaign was a Pill Check Week in which volunteers visited public health facilities and then submitted reports of out-of-stock essential medicines via FrontlineSMS from their mobile phones. Those messages were used to create a stock-out map using the software Ushahidi to create a compelling visual which, in combination with other tactics, successfully drew the attention of local governments and international media.

By Earl and Kimport’s logic the campaign should have maximally leveraged digital affordances by using mostly (or exclusively) online tactics, since these tactics maximally leverage the digital affordances of low-cost collective action without copresence. According to this pure leveraged affordance analysis, the organizers should have called the facilities to check pill levels rather than sending volunteers out into the field to collect reports. A volunteer could call a facility in a few minutes for a few cents but it likely took hours to visit the facilities in person, plus the cost of transportation. However, if volunteers had only called facilities, they would have been forced to receive their information from facility staff who may not have been motivated to be truthful about the shortages at their facility, assuming any of the overworked staff even had the inclination to pick up at all. By using the higher-cost offline tactic of visiting the facilities, the volunteers could interview a range of patients, circumventing the necessity of getting information from facility staff and making their final reports multi-sourced and more reliable. In this way organizers increased their effectiveness by choosing a mixed offline-digital tactic rather than a purely digital tactic that maximally leveraged digital affordances. Though less theoretically elegant that a pure affordances framework, analysis of the effects of digital tactics reveals that no single measure can determine the value of a tactics, especially where information about effects is absent.

Power-Law Dynamics

The final theoretical argument, which Earl and Kimport introduce at the end of their paper, is that new digital organizations are likely to follow a power-law function because of the ease of digital organizing. They quote Clay Shirky’s definition of a power law: data “in which the nth position has 1/nth of the first position’s rank” such that the 2nd position has a quantity of 1/2 that of the 1st position and so on. They argue:

Innovative uses of the Web can make organizing inexpensive enough that it can begin to follow power-law dynamics in some situations. When that happens, one person will bear the majority of the costs, the active organizer has to bear substantially fewer costs, and so on down the line so that quickly there are no organizing costs left to bear at all…. Our power-law explanation of organizing is certainly consistent with the findings that we cited above of online protests being led by single individuals, pairs, or drastically small teams (p 152 and 153).

However, “consistent” is quite different from mathematically verified. Saying that, in the thirty-eight digital organizations they interviewed, a lead organizer did the lion’s share of the work is far from being able to demonstrate a power law. On could easily imagine that the division of labor followed this patten without following the power law.

In fact, they do not even attempt to show quantitative evidence for their claim. It would not have been difficult to collect data during these interviews about the number of hours per week each member of the organization worked, to see how closely this data matched a power-law graph. However, there is no evidence that they attempted to prove or disprove their claim, even based on their own limited sample. Perhaps they thought that the theory was so attractive that it did not require evidence to support it.

But Does it Work?

Why do Earl and Kimport’s theoretical conclusions seem so detached from the real life practice of digital organizing and activism? It is likely because they blithely eschew the key evaluative question of any organizer: “does it work?”. Earl and Kimport do not consider the effectiveness of any of the tactics in their study. “Ours is not a study of the effectiveness of e-tactics,” they write, “so although we are aware of many successful online campaigns, including efforts in our data set, we cannot empirically address” claims challenging the effectiveness of digital tactics (p 94). Their focus on claims over evidence may be a direct result of the fact that their otherwise methodologically strong study ignores the effects of their digital tactics.

Bottom Line: The authors present arguments in favor of theories of value that are unsupported – or substantially contradicted – by evidence.

Conclusion

The book is not without interesting ideas. The idea that digital tactics will force a re-working of some elements of social movement theory is spot-on and leveraged affordances is certainly one valuable way to evaluate a digital tactic. Changes in organizational structure, made possible by digital tools, are also important, though without reference the the effectiveness of these new types of organizations, their ultimate impact is in question. Earl and Kimport are also right to note that online privacy norms may change expectations about what it means to act in public and I agree that these quick-start organizations, created by new activists, will likely lead to more episodic activism. It also seems that the ease of online participation may also not require previous feelings of collective identity to motivate participation, another interesting challenge to existing theory.

These positive points do not save the book, though. Its disregard for effects, choice of outdated tactics as the focus of study, and attention to theory over evidence lays the whole field of the digital contention open to charges of disconnected abstraction, cyber-utopianism, and techno-fetishism that threaten this young field’s legitimacy.