OII – The Policy and Internet Blog

Sexism Typology: Literature Review

The Everyday Sexism Project catalogues instances of sexism experienced by women on a day to day basis. We will be using computational techniques to extract the most commonly occurring sexism-related topics.

As Laura Bates, founder of the Everyday Sexism project, has recently highlighted, “it seems to be increasingly difficult to talk about sexism, equality, and women’s rights” (Everyday Sexism Project, 2015). With many theorists suggesting that we have entered a so-called “post-feminist” era in which gender equality has been achieved (cf. McRobbie, 2008; Modleski, 1991), to complain about sexism not only risks being labelled as “uptight”, “prudish”, or a “militant feminist”, but also exposes those who speak out to sustained, and at times vicious, personal attacks (Everyday Sexism Project, 2015). Despite this, thousands of women are speaking out, through Bates’ project, about their experiences of everyday sexism. Our research seeks to draw on the rich history of gender studies in the social sciences, coupling it with emerging computational methods for topic modelling, to better understand the content of reports to the Everyday Sexism Project and the lived experiences of those who post them. Here, we outline the literature which contextualizes our study.

Studies on sexism are far from new. Indeed, particularly amongst feminist theorists and sociologists, the analysis (and deconstruction) of “inequality based on sex or gender categorization” (Harper, 2008) has formed a central tenet of both academic inquiry and a radical politics of female emancipation for several decades (De Beauvoir, 1949; Friedan, 1963; Rubin, 1975; Millett, 1971). Reflecting its feminist origins, historical research on sexism has broadly focused on defining sexist interactions (cf. Glick and Fiske, 1997) and on highlighting the problematic, biologically rooted ‘gender roles’ that form the foundation of inequality between men and women (Millett, 1971; Renzetti and Curran, 1992; Chodorow, 1995).

More recent studies, particularly in the field of psychology, have shifted the focus away from whether and how sexism exists, towards an examination of the psychological, personal, and social implications that sexist incidents have for the women who experience them. As such, theorists such as Matteson and Moradi (2005), Swim et al (2001) and Jost and Kay (2005) have highlighted the damaging intellectual and mental health outcomes for women who are subject to continual experiences of sexism. Atwood, for example, argues in her study of gender bias in families, that sexism combines with other life stressors to create significant psychological distress in women, resulting in them needing to “seek therapy, most commonly for depression and anxiety” (2001, 169).

Given its increasing ubiquity in every day life, it is hardly surprising that the relationship between technology and sexism has also sparked interest from contemporary researchers in the field. Indeed, several studies have explored the intersection between gender and power online, with Susan Herring’s work on gender differences in computer-mediated communication being of particular note (cf. Herring, 2008). Theorists such as Mindi D. Foster have focused on the impact that using digital technology, and particularly Web 2.0 technologies, to talk about sexism can have on women’s well being. Foster’s study found that when women tweeted about sexism, and in particular when they used tweets to a) name the problem, b) criticise it, or c) to suggest change, they viewed their actions as effective and had enhanced life satisfaction, and therefore felt empowered (Foster, 2015: 21).

Despite this diversity of research on sexism, however, there remain some notable gaps in understanding. In particular, as this study hopes to highlight, little previous research on sexism has considered the different ‘types’ of sexism experienced by women (beyond an identification of the workplace and the education system as contexts in which sexism often manifests as per Barnett, 2005; Watkins et al., 2006; Klein, 1992). Furthermore, research focusing on sexism has thus far been largely qualitative in nature. Although a small number of studies have employed quantitative methods (cf. Brandt 2011; Becker and Wright, 2011), none have used computational approaches to analyse the wealth of available online data on sexism.

This project, which will apply a natural language processing approach to analyse data collected from the Everyday Sexism Project website, seeks to fill such a gap. By providing much needed analysis of a large-scale crowd sourced data set on sexism, it is our hope that knowledge gained from this study will advance both the sociological understanding of women’s lived experiences of sexism, and methodological understandings of the suitability of computational topic modelling for conducting this kind of research.

Find out more about the OII’s research on the Everyday Sexism project by visiting the webpage or by looking at the other Policy & Internet blog posts on the project – post 1 and post 2.

Taha Yasseri is a Research Fellow at the OII who has interests in analysis of Big Data to understand human dynamics, government-society interactions, mass collaboration, and opinion dynamics.

Kathryn Eccles is a Research Fellow at the OII who has research interests in the impact of new technologies on scholarly behaviour and research, particularly in the Humanities.

Sophie Melville is a Research Assistant working at the OII. She previously completed the MSc in the Social Science of the Internet at the OII.

References

Atwood, N. C. (2001). Gender bias in families and its clinical implications for women. Social Work, 46 pp. 23–36.

Barnett, R. C. (2005). Ageism and Sexism in the workplace. Generations. 29(3) pp. 25 30.

Bates, Laura. (2015). Everyday Sexism [online] Available at: http://everydaysexism.com [Accessed 1 May 2016].

Becker, Julia C. & Wright, Stephen C. (2011). Yet another dark side of chivalry: Benevolent sexism undermines and hostile sexism motivates collective action for social change. Journal of Personality and Social Psychology, Vol 101(1), Jul 2011, 62-77

Brandt, Mark. (2011). Sexism and Gender Inequality across 57 societies. Psychiological Science. 22(11).

Chodorow, Nancy (1995). “Becoming a feminist foremother”. In Phyllis Chesler, Esther D. Rothblum, Ellen Cole,. Feminist foremothers in women’s studies, psychology, and mental health. New York: Haworth Press. pp. 141–154.

De Beauvoir, Simone. (1949). The second sex, woman as other. London: Vintage.

Foster, M. D. (2015). Tweeting about sexism: The well-being benefits of a social media collective action. British Journal of Social Psychology.

Friedan, Betty. (1963). The Feminine Mystique. W. W. Norton and Co.

Glick, Peter. & Fiske, Susan T. (1997). Hostile and Benevolent Sexism. Psychology of Women Quarterly, 21(1) pp. 119 – 135.

Harper, Amney J. (2008). The relationship between sexism, ambivalent sexism, and relationship quality in heterosexual women. PhD Auburn University.

Herring, Susan C. (2008). Gender and Power in Online Communication. In Janet Holmes and Miriam Meyerhoff (eds) The Handbook of Language and Gender. Oxford: Blackwell.

Jost, J. T., & Kay, A. C. (2005). Exposure to benevolent sexism and complementary gender stereotypes: Consequences for specific and diffuse forms of system justification. Journal of Personality and Social Psychology, 88 pp. 498–509.

Klein, Susan Shurberg. (1992). Sex equity and sexuality in education: breaking the barriers. State University of New York Press.

Matteson, A. V., & Moradi, B. (2005). Examining the structure of the schedule of sexist events: Replication and extension. Psychology of Women Quarterly, 29 pp. 47–57.

McRobbie, Angela (2004). Post-feminism and popular culture. Feminist Media Studies, 4(3) pp. 255 – 264.

Millett, Kate. (1971). Sexual politics. UK: Virago.

Modleski, Tania. (1991). Feminism without women: culture and criticism in a “postfeminist” age. New York: Routledge.

Renzetti, C. and D. Curran, 1992, “Sex-Role Socialization”, in Feminist Philosophies, J. Kourany, J. Sterba, and R. Tong (eds.), New Jersey: Prentice Hall.

Rubin, Gayle. (1975). The traffic in women: notes on the “political economy” of sex. In Rayna R. Reiter (ed.), Toward and anthropology of women. Monthly Review Press.

Swim, J. K., Hyers, L. L., Cohen, L. L., & Ferguson, M. J. (2001). Everyday sexism: Evidence for its incidence, nature, and psychological impact from three daily diary studies. Journal of Social Issues, 57 pp. 31–53.

Watkins et al. (2006). Does it pay to be sexist? The relationship between modern sexism and career outcomes. Journal of Vocational Behaviour. 69(3) pp. 524 – 537.

Exploring the Ethics of Monitoring Online Extremism

(Part 2 of 2) The Internet serves not only as a breeding ground for extremism, but also offers myriad data streams which potentially hold great value to law enforcement. The report by the OII’s Ian Brown and Josh Cowls for the VOX-Pol project: Check the Web: Assessing the Ethics and Politics of Policing the Internet for Extremist Material explores the complexities of policing the web for extremist material, and its implications for security, privacy and human rights. In the second of a two-part post, Josh Cowls and Ian Brown discuss the report with blog editor Bertie Vidgen. Read the first post.

Surveillance in NYC's financial district. Photo by Jonathan McIntosh (flickr). — Surveillance in NYC’s financial district. Photo by Jonathan McIntosh (flickr).

Ed: Josh, political science has long posed a distinction between public spaces and private ones. Yet it seems like many platforms on the Internet, such as Facebook, cannot really be categorized in such terms. If this correct, what does it mean for how we should police and govern the Internet?

Josh: I think that is right – many online spaces are neither public nor private. This is also an issue for some for privacy legal frameworks (especially in the US).. A lot of the covenants and agreements were written forty or fifty years ago, long before anyone had really thought about the Internet. That has now forced governments, societies and parliaments to adapt these existing rights and protocols for the online sphere. I think that we have some fairly clear laws about the use of human intelligence sources, and police law in the offline sphere. The interesting question is how we can take that online. How can the pre-existing standards, like the requirement that procedures are necessary and proportionate, or the ‘right to appeal’, be incorporated into online spaces? In some cases there are direct analogies. In other cases there needs to be some re-writing of the rule book to try figure out what we mean. And, of course, it is difficult because the internet itself is always changing!

Ed: So do you think that concepts like proportionality and justification need to be updated for online spaces?

Josh: I think that at a very basic level they are still useful. People know what we mean when we talk about something being necessary and proportionate, and about the importance of having oversight. I think we also have a good idea about what it means to be non-discriminatory when applying the law, though this is one of those areas that can quickly get quite tricky. Consider the use of online data sources to identify people. On the one hand, the Internet is ‘blind’ in that it does not automatically codify social demographics. In this sense it is not possible to profile people in the same way that we can offline. On the other hand, it is in some ways the complete opposite. It is very easy to directly, and often invisibly, create really firm systems of discrimination – and, most problematically, to do so opaquely.

This is particularly challenging when we are dealing with extremism because, as we pointed out in the report, extremists are generally pretty unremarkable in terms of demographics. It perhaps used to be true that extremists were more likely to be poor or to have had challenging upbringings, but many of the people going to fight for the Islamic State are middle class. So we have fewer demographic pointers to latch onto when trying to find these people. Of course, insofar as there are identifiers they won’t be released by the government. The real problem for society is that there isn’t very much openness and transparency about these processes.

Ed: Governments are increasingly working with the private sector to gain access to different types of information about the public. For example, in Australia a Telecommunications bill was recently passed which requires all telecommunication companies to keep the metadata – though not the content data – of communications for two years. A lot of people opposed the Bill because metadata is still very informative, and as such there are some clear concerns about privacy. Similar concerns have been expressed in the UK about an Investigatory Powers Bill that would require new Internet Connection Records about customers, online activities. How much do you think private corporations should protect people’s data? And how much should concepts like proportionality apply to them?

Ian: To me the distinction between metadata and content data is fairly meaningless. For example, often just knowing when and who someone called and for how long can tell you everything you need to know! You don’t have to see the content of the call. There are a lot of examples like this which highlight the slightly ludicrous nature of distinguishing between metadata and content data. It is all data. As has been said by former US CIA and NSA Director Gen. Michael Hayden, “we kill people based on metadata.”

One issue that we identified in the report is the increased onus on companies to monitor online spaces, and all of the legal entanglements that come from this given that companies might not be based in the same country as the users. One of our interviewees called this new international situation a ‘very different ballgame’. Working out how to deal with problematic online content is incredibly difficult, and some huge issues of freedom of speech are bound up in this. On the one hand, there is a government-led approach where we use the law to take down content. On the other hand is a broader approach, whereby social networks voluntarily take down objectionable content even if it is permissible under the law. This causes much more serious problems for human rights and the rule of law.

Read the full report: Brown, I., and Cowls, J., (2015) Check the Web: Assessing the Ethics and Politics of Policing the Internet for Extremist Material. VOX-Pol Publications.

Ian Brown is Professor of Information Security and Privacy at the OII. His research is focused on surveillance, privacy-enhancing technologies, and Internet regulation.

Josh Cowls is a a student and researcher based at MIT, working to understand the impact of technology on politics, communication and the media.

Josh and Ian were talking to Blog Editor Bertie Vidgen.

P-values are widely used in the social sciences, but often misunderstood: and that’s a problem.

P-values are widely used in the social sciences, especially ‘big data’ studies, to calculate statistical significance. Yet they are widely criticized for being easily hacked, and for not telling us what we want to know. Many have argued that, as a result, research is wrong far more often than we realize. In their recent article P-values: Misunderstood and Misused OII Research Fellow Taha Yasseri and doctoral student Bertie Vidgen argue that we need to make standards for interpreting p-values more stringent, and also improve transparency in the academic reporting process, if we are to maximise the value of statistical analysis.

“Significant”: an illustration of selective reporting and statistical significance from XKCD. Available online at http://xkcd.com/882/ — “Significant”: an illustration of selective reporting and
statistical significance from XKCD. Available online at
http://xkcd.com/882/

In an unprecedented move, the American Statistical Association recently released a statement (March 7 2016) warning against how p-values are currently used. This reflects a growing concern in academic circles that whilst a lot of attention is paid to the huge impact of big data and algorithmic decision-making, there is considerably less focus on the crucial role played by statistics in enabling effective analysis of big data sets, and making sense of the complex relationships contained within them. Because much as datafication has created huge social opportunities, it has also brought to the fore many problems and limitations with current statistical practices. In particular, the deluge of data has made it crucial that we can work out whether studies are ‘significant’. In our paper, published three days before the ASA’s statement, we argued that the most commonly used tool in the social sciences for calculating significance – the p-value – is misused, misunderstood and, most importantly, doesn’t tell us what we want to know.

The basic problem of ‘significance’ is simple: it is simply unpractical to repeat an experiment an infinite number of times to make sure that what we observe is “universal”. The same applies to our sample size: we are often unable to analyse a “whole population” sample and so have to generalize from our observations on a limited size sample to the whole population. The obvious problem here is that what we observe is based on a limited number of experiments (sometimes only one experiment) and from a limited size sample, and as such could have been generated by chance rather than by an underlying universal mechanism! We might find it impossible to make the same observation if we were to replicate the same experiment multiple times or analyse a larger sample. If this is the case then we will mischaracterise what is happening – which is a really big problem given the growing importance of ‘evidence-based’ public policy. If our evidence is faulty or unreliable then we will create policies, or intervene in social settings, in an equally faulty way.

The way that social scientists have got round this problem (that samples might not be representative of the population) is through the ‘p-value’. The p-value tells you the probability of making a similar observation in a sample with the same size and in the same number of experiments, by pure chance In other words, it is actually telling you is how likely it is that you would see the same relationship between X and Y even if no relationship exists between them. On the face of it this is pretty useful, and in the social sciences we normally say that a p-value of 1 in 20 means the results are significant. Yet as the American Statistical Association has just noted, even though they are incredibly widespread many researchers mis-interpret what p-values really mean.

In our paper we argued that p-values are misunderstood and misused because people think the p-value tells you much more than it really does. In particular, people think the p-value tells you (i) how likely it is that a relationship between X and Y really exists and (ii) the percentage of all findings that are false (which is actually something different called the False Discovery Rate). As a result, we are far too confident that academic studies are correct. Some commentators have argued that at least 30% of studies are wrong because of problems related to p-values: a huge figure. One of the main problems is that p-values can be ‘hacked’ and as such easily manipulated to show significance when none exists.

If we are going to base public policy (and as such public funding) on ‘evidence’ then we need to make sure that the evidence used is reliable. P-values need to be used far more rigorously, with significance levels of 0.01 or 0.001 seen as standard. We also need to start being more open and transparent about how results are recorded. It is a fine line between data exploration (a legitimate academic exercise) and ‘data dredging’ (where results are manipulated in order to find something noteworthy). Only if researchers are honest about what they are doing will we be able to maximise the potential benefits offered by Big Data. Luckily there are some great initiatives – like the Open Science Framework – which improve transparency around the research process, and we fully endorse researchers making use of these platforms.

Scientific knowledge advances through corroboration and incremental progress, and it is crucial that we use and interpret statistics appropriately to ensure this progress continues. As our knowledge and use of big data methods increase, we need to ensure that our statistical tools keep pace.

Read the full paper: Vidgen, B. and Yasseri, T., (2016) P-values: Misunderstood and Misused, Frontiers in Physics, 4:6. http://dx.doi.org/10.3389/fphy.2016.00006

Bertie Vidgen is a doctoral student at the Oxford Internet Institute researching far-right extremism in online contexts. He is supervised by Dr Taha Yasseri, a research fellow at the Oxford Internet Institute interested in how Big Data can be used to understand human dynamics, government-society interactions, mass collaboration, and opinion dynamics.

Facts and figures or prayers and hugs: how people with different health conditions support each other online

Online support groups are being used increasingly by individuals who suffer from a wide range of medical conditions. OII DPhil Student Ulrike Deetjen‘s recent article with John Powell, Informational and emotional elements in online support groups: a Bayesian approach to large-scale content analysis uses machine learning to examine the role of online support groups in the healthcare process. They categorise 40,000 online posts from one of the most well-used forums to show how users with different conditions receive different types of support.

Online forums are important means of people living with health conditions to obtain both emotional and informational support from this in a similar situation. Pictured: The Alzheimer Society of B.C. unveiled three life-size ice sculptures depicting important moments in life. The ice sculptures will melt, representing the fading of life memories on the dementia journey. Image: bcgovphotos (Flickr)

Online support groups are one of the major ways in which the Internet has fundamentally changed how people experience health and health care. They provide a platform for health discussions formerly restricted by time and place, enable individuals to connect with others in similar situations, and facilitate open, anonymous communication.

Previous studies have identified that individuals primarily obtain two kinds of support from online support groups: informational (for example, advice on treatments, medication, symptom relief, and diet) and emotional (for example, receiving encouragement, being told they are in others’ prayers, receiving “hugs”, or being told that they are not alone). However, existing research has been limited as it has often used hand-coded qualitative approaches to contrast both forms of support, thereby only examining relatively few posts (<1,000) for one or two conditions.

In contrast, our research employed a machine-learning approach suitable for uncovering patterns in “big data”. Using this method a computer (which initially has no knowledge of online support groups) is given examples of informational and emotional posts (2,000 examples in our study). It then “learns” what words are associated with each category (emotional: prayers, sorry, hugs, glad, thoughts, deal, welcome, thank, god, loved, strength, alone, support, wonderful, sending; informational: effects, started, weight, blood, eating, drink, dose, night, recently, taking, side, using, twice, meal). The computer then uses this knowledge to assess new posts, and decide whether they contain more emotional or informational support.

With this approach we were able to determine the emotional or informational content of 40,000 posts across 14 different health conditions (breast cancer, prostate cancer, lung cancer, depression, schizophrenia, Alzheimer’s disease, multiple sclerosis, cystic fibrosis, fibromyalgia, heart failure, diabetes type 2, irritable bowel syndrome, asthma, and chronic obstructive pulmonary disease) on the international support group forum Dailystrength.org.

Our research revealed a slight overall tendency towards emotional posts (58% of posts were emotionally oriented). Across all diseases, those who write more also tend to write more emotional posts—we assume that as people become more involved and build relationships with other users they tend to provide more emotional support, instead of simply providing information in one-off interactions. At the same time, we also observed that older people write more informational posts. This may be explained by the fact that older people more generally use the Internet to find information, that they become experts in their chronic conditions over time, and that with increasing age health conditions may have less emotional impact as they are relatively more expected.

The demographic prevalence of the condition may also be enmeshed with the disease-related tendency to write informational or emotional posts. Our analysis suggests that content differs across the 14 conditions: mental health or brain-related conditions (such as depression, schizophrenia, and Alzheimer’s disease) feature more emotionally oriented posts, with around 80% of posts primarily containing emotional support. In contrast, nonterminal physical conditions (such as irritable bowel syndrome, diabetes, asthma) rather focus on informational support, with around 70% of posts providing advice about symptoms, treatments, and medication.

Finally, there was no gender difference across conditions with respect to the amount of posts that were informational versus emotional. That said, prostate cancer forums are oriented towards informational support, whereas breast cancer forums feature more emotional support. Apart from the generally different nature of both conditions, one explanation may lie in the nature of single-gender versus mixed-gender groups: an earlier meta-study found that women write more emotional content than men when talking among others of the same gender – but interestingly, in mixed-gender discussions, these differences nearly disappeared.

Our research helped to identify factors that determine whether online content is informational or emotional, and demonstrated how posts differ across conditions. In addition to theoretical insights about patient needs, this research will help practitioners to better understand the role of online support groups for different patients, and to provide advice to patients about the value of online support.

The results also suggest that online support groups should be integrated into the digital health strategies of the UK and other nations. At present the UK plan for “Personalised Health and Care 2020” is centred around digital services provided within the health system, and does not yet reflect the value of person-generated health data from online support groups to patients. Our research substantiates that it would benefit from considering the instrumental role that online support groups can play in the healthcare process.

Read the full paper: Deetjen, U. and J. A. Powell (2016) Informational and emotional elements in online support groups: a Bayesian approach to large-scale content analysis. Journal of the American Medical Informatics Association. http://dx.doi.org/10.1093/jamia/ocv190

Ulrike Deetjen (née Rauer) is a doctoral student at the Oxford Internet Institute researching the influence of the Internet on healthcare provision and health outcomes.

Topic modelling content from the “Everyday Sexism” project: what’s it all about?

We recently announced the start of an exciting new research project that will involve the use of topic modelling in understanding the patterns in submitted stories to the Everyday Sexism website. Here, we briefly explain our text analysis approach, “topic modelling”.

At its very core, topic modelling is a technique that seeks to automatically discover the topics contained within a group of documents. ‘Documents’ in this context could refer to text items as lengthy as individual books, or as short as sentences within a paragraph. Let’s take the idea of sentences-as-documents as an example:

Document 1: I like to eat kippers for breakfast.
Document 2: I love all animals, but kittens are the cutest.
Document 3: My kitten eats kippers too.

Assuming that each sentence contains a mixture of different topics (and that a ‘topic’ can be understood as a collection of words (of any part of speech) that have different probabilities of appearance in passages discussing the topic), how does the topic modelling algorithm ‘discover’ the topics within these sentences?

The algorithm is initiated by setting the number of topics that it needs to extract. Of course, it is hard to guess this number without having an insight on the topics, but one can think of this as a resolution tuning parameter. The smaller the number of topics is set, the more general the bag of words in each topic would be, and the looser the connections between them.

The algorithm loops through all of the words in each document, assigning every word to one of our topics in a temporary and semi-random manner. This initial assignment is arbitrary and it is easy to show that different initializations lead to the same results in long run. Once each word has been assigned a temporary topic, the algorithm then re-iterates through each word in each document to update the topic assignment using two criteria: 1) How prevalent is the word in question across topics? And 2) How prevalent are the topics in the document?

To quantify these two, the algorithm calculates the likelihood of the words appearing in each document assuming the assignment of words to topics and topics to documents.

Of course words can appear in different topics and more than one topic can appear in a document. But the iterative algorithm seeks to maximize the self-consistency of the assignment by maximizing the likelihood of the observed word-document statistics.

We can illustrate this process and its outcome by going back to our example. A topic modelling approach might use the process above to discover the following topics across our documents:

Document 1: I like to eat kippers for breakfast. [100% Topic A]
Document 2: I love all animals, but kittens are the cutest. [100% Topic B]
Document 3: My kitten eats kippers too. [67% Topic A, 33% Topic B]

Topic modelling defines each topic as a so-called ‘bag of words’, but it is the researcher’s responsibility to decide upon an appropriate label for each topic based on their understanding of language and context. Going back to our example, the algorithm might classify the underlined words under Topic A, which we could then label as ‘food’ based on our understanding of what the words mean. Similarly the italicised words might be classified under a separate topic, Topic B, which we could label ‘animals’. In this simple example the word “eat” has appeared in a sentence dominated by Topic A, but also in a sentence with some association to Topic B. Therefore it can also be seen as a connector of the two topics. Of course animals eat too and they like food!

We are going to use a similar approach to first extract the main topics reflected on the reports to the Everyday Sexism Project website and extract the relation between the sexism-related topics and concepts based on the overlap between the bags of words of each topic. Finally we can also look into the co-appearance of topics in the same document. This way we try to draw a linguistic picture of the more than 100,000 submitted reports.

As ever, be sure to check back for further updates on our progress!

The limits of uberization: How far can platforms go?

Platforms that enable users to come together and buy/sell services with confidence, such as Uber, have become remarkably popular, with the companies often transforming the industries they enter. In this blog post the OII’s Vili Lehdonvirta analyses why the domestic cleaning platform Homejoy failed to achieve such success. He argues that when buyer and sellers enter into repeated transactions they can communicate directly, and as such often abandon the platform.

Homejoy CEO Adora Cheung appears on stage at the 2014 TechCrunch Disrupt Europe/London, at The Old Billingsgate on October 21, 2014 in London, England. Image: TechCruch (Flickr)

Homejoy was slated to become the Uber of domestic cleaning services. It was a platform that allowed customers to summon a cleaner as easily as they could hail a ride. Regular cleanups were just as easy to schedule. Ratings from previous clients attested to the skill and trustworthiness of each cleaner. There was no need to go through a cleaning services agency, or scour local classifieds to find a cleaner directly: the platform made it easy for both customers and people working as cleaners to find each other. Homejoy made its money by taking a cut out of each transaction. Given how incredibly successful Uber and Airbnb had been in applying the same model to their industries, Homejoy was widely expected to become the next big success story. It was to be the next step in the inexorable uberization of every industry in the economy.

On 17 July 2015, Homejoy announced that it was shutting down. Usage had grown slower than expected, revenues remained poor, technical glitches hurt operations, and the company was being hit with lawsuits on contractor misclassification. Investors’ money and patience had finally ran out. Journalists wrote interesting analyses of Homejoy’s demise (Forbes, TechCrunch, Backchannel). The root causes of any major business failure (or indeed success) are complex and hard to pinpoint. However, one of the possible explanations identified in these stories stands out, because it corresponds strongly with what theory on platforms and markets could have predicted. Homejoy wasn’t growing and making money because clients and cleaners were taking their relationships off-platform: after making the initial contact through Homejoy, they would simply exchange contact details and arrange further cleanups directly, taking the platform and its revenue share out of the loop. According to Forbes, only 15-20 percent of customers came back to Homejoy within a month to arrange another cleanup.

According to the theory of platforms in economics and management studies literature, platforms solve coordination problems. Digital service platforms like Uber and Airbnb solve, in particular, the problem of finding another party to transact with. Through marketing and bootstrapping efforts they ensure that both buyers and sellers sign up to the platform, and then provide match-making mechanisms to bring them together. They also provide solutions towards the problem of opportunism, that is, how to avoid being cheated by the other party. Rating systems are their main tool in this.

Platforms must compete against the existing institutional arrangements in their chosen industry. Uber has been very successful in taking away business from government-licensed taxicabs. Airbnb has captured market share from hotels and hotel booking sites. Both have also generated lots of new business: transactions that previously didn’t happen at all. It’s not that people didn’t already occassionally pay a highschool friend to give them a ride home from a party, or rent a room for the weekend from a friend of a friend who lives in New York. It’s that platforms make similar things possible even when the highschool friend is not available, or if you simply don’t know anyone with a flat in New York. Platforms coordinate people to turn what is otherwise a thin market into a thick one. Not only do platforms help you to find a stranger to transact with, but they also help you to trust that stranger.

Now consider the market for home cleaning services. Home cleaning differs from on-demand transport and short-term accommodation in one crucial way: the service is typically repeated. Through repeated interactions, the buyer and the seller develop trust in each other. They also develop knowledge capital specific to that particular relationship. The buyer might invest time into communicating their preferences and little details about their home to the seller, while the seller will gradually become more efficient at cleaning that particular home. They have little need for the platform to discipline each individual cleanup; relationships are thus soon taken off-platform. Instead of an all-encompassing Uber-style platform, all that may be needed is a classifieds site or a conventional agency that provides the initial introduction and references. Contrast this with on-demand transport and short-term accommodation, where each transaction is unique and thus each time the counterparty is a stranger — and as such a potential cheat or deadbeat. Here the platform continues to provide security after the parties have been introduced.

The case of Homejoy and the economic theory on platforms thus suggest that there are fundamental limits to the uberization of the economy. Digital service platforms can be very successful at mediating one-off transactions, but they are much less useful in industries where the exact same service is repeated many times, and where buyers and sellers develop assets specific to the relationship. Such industries are more likely to continue to be shaped by hierarchies and networks of personal relationships.

There are probably other dimensions that are also pivotal in predicting whether an industry is susceptible to uberization. Geographical span is one: there are efficiencies to be had from particular cleaners specializing in particular neighbourhoods. Yet, at the same time, online labour platforms like Upwork cater to buyers and sellers of software development (and other digitally mediated contract work) across national boundaries. I will discuss this dimension in detail in a future blog post.

Vili Lehdonvirta is a Research Fellow at the OII. He is an economic sociologist who studies the design and socioeconomic implications of digital marketplaces and platforms, using conventional social research methods as well as novel data science approaches. Read Vili’s other Policy & Internet Blog posts on Uber and Airbnb:

Uber and Airbnb make the rules now – but to whose benefit?

Why are citizens migrating to Uber and Airbnb, and what should governments do about it?

Should we love Uber and Airbnb or protest against them?

The limits of uberization: How far can platforms go?

Uber and Airbnb make the rules now – but to whose benefit?

Why are citizens migrating to Uber and Airbnb, and what should governments do about it?

Should we love Uber and Airbnb or protest against them?

Assessing the Ethics and Politics of Policing the Internet for Extremist Material

The Internet serves not only as a breeding ground for extremism, but also offers myriad data streams which potentially hold great value to law enforcement. The report by the OII’s Ian Brown and Josh Cowls for the VOX-Pol project: Check the Web: Assessing the Ethics and Politics of Policing the Internet for Extremist Material explores the complexities of policing the web for extremist material, and its implications for security, privacy and human rights. Josh Cowls discusses the report with blog editor Bertie Vidgen.*

*please note that the views given here do not necessarily reflect the content of the report, or those of the lead author, Ian Brown.

In terms of counter-speech there are different roles for government, civil society, and industry. Image by Miguel Discart (Flickr).

Ed: Josh, could you let us know the purpose of the report, outline some of the key findings, and tell us how you went about researching the topic?

Josh: Sure. In the report we take a step back from the ground-level question of ‘what are the police doing?’ and instead ask, ‘what are the ethical and political boundaries, rationale and justifications for policing the web for these kinds of activity?’ We used an international human rights framework as an ethical and legal basis to understand what is being done. We also tried to further the debate by clarifying a few things: what has already been done by law enforcement, and, really crucially, what the perspectives are of all those involved, including lawmakers, law enforcers, technology companies, academia and many others.

We derived the insights in the report from a series of workshops, one of which was held as part of the EU-funded VOX-Pol network. The workshops involved participants who were quite high up in law enforcement, the intelligence agencies, the tech industry civil society, and academia. We followed these up with interviews with other individuals in similar positions and conducted background policy research.

Ed: You highlight that many extremist groups (such as Isis) are making really significant use of online platforms to organize, radicalize people, and communicate their messages.

Josh: Absolutely. A large part of our initial interest when writing the report lay in finding out more about the role of the Internet in facilitating the organization, coordination, recruitment and inspiration of violent extremism. The impact of this has been felt very recently in Paris and Beirut, and many other places worldwide. This report pre-dates these most recent developments, but was written in the context of these sorts of events.

Given the Internet is so embedded in our social lives, I think it would have been surprising if political extremist activity hadn’t gone online as well. Of course, the Internet is a very powerful tool and in the wrong hands it can be a very destructive force. But other research, separate from this report, has found that the Internet is not usually people’s first point of contact with extremism: more often than not that actually happens offline through people you know in the wider world. Nonetheless it can definitely serve as an incubator of extremism and can serve to inspire further attacks.

Ed: In the report you identify different groups in society that are affected by, and affecting, issues of extremism, privacy, and governance – including civil society, academics, large corporations and governments

Josh: Yes, in the later stages of the report we do divide society into these groups, and offer some perspectives on what they do, and what they think about counter-extremism. For example, in terms of counter-speech there are different roles for government, civil society, and industry. There is this idea that ISIS are really good at social media, and that that is how they are powering a lot of their support; but one of the people that we spoke to said that it is not the case that ISIS are really good, it is just that governments are really bad!

We shouldn’t ask government to participate in the social network: bureaucracies often struggle to be really flexible and nimble players on social media. In contrast, civil society groups tend to be more engaged with communities and know how to “speak the language” of those who might be vulnerable to radicalization. As such they can enter that dialogue in a much more informed and effective way.

The other tension, or paradigm, that we offer in this report is the distinction between whether people are ‘at risk’ or ‘a risk’. What we try to point to is that people can go from one to the other. They start by being ‘at risk’ of radicalization, but if they do get radicalized and become a violent threat to society, which only happens in the minority of cases, then they become ‘a risk’. Engaging with people who are ‘at risk’ highlights the importance of having respect and dialogue with communities that are often the first to be lambasted when things go wrong, but which seldom get all the help they need, or the credit when they get it right. We argue that civil society is particularly suited for being part of this process.

Ed: It seems like the things that people do or say online can only really be understood in terms of the context. But often we don’t have enough information, and it can be very hard to just look at something and say ‘This is definitely extremist material that is going to incite someone to commit terrorist or violent acts’.

Josh: Yes, I think you’re right. In the report we try to take what is a very complicated concept – extremist material – and divide it into more manageable chunks of meaning. We talk about three hierarchical levels. The degree of legal consensus over whether content should be banned decreases as it gets less extreme. The first level we identified was straight up provocation and hate speech. Hate speech legislation has been part of the law for a long time. You can’t incite racial hatred, you can’t incite people to crimes, and you can’t promote terrorism. Most countries in Europe have laws against these things.

The second level is the glorification and justification of terrorism. This is usually more post-hoc as by definition if you are glorifying something it has already happened. You may well be inspiring future actions, but that relationship between the act of violence and the speech act is different than with provocation. Nevertheless, some countries, such as Spain and France, have pushed hard on criminalising this. The third level is non-violent extremist material. This is the most contentious level, as there is very little consensus about what types of material should be called ‘extremist’ even though they are non-violent. One of the interviewees that we spoke to said that often it is hard to distinguish between someone who is just being friendly and someone who is really trying to persuade or groom someone to go to Syria. It is really hard to put this into a legal framework with the level of clarity that the law demands.

There is a proportionality question here. When should something be considered specifically illegal? And, then, if an illegal act has been committed what should the appropriate response be? This is bound to be very different in different situations.

Ed: Do you think that there are any immediate or practical steps that governments can take to improve the current situation? And do you think that there any ethical concerns which are not being paid sufficient attention?

Josh: In the report we raised a few concerns about existing government responses. There are lots of things beside privacy that could be seen as fundamental human rights and that are being encroached upon. Freedom of association and assembly is a really interesting one. We might not have the same reverence for a Facebook event plan or discussion group as we would a protest in a town hall, but of course they are fundamentally pretty similar.

The wider danger here is the issue of mission creep. Once you have systems in place that can do potentially very powerful analytical investigatory things then there is a risk that we could just keep extending them. If something can help us fight terrorism then should we use it to fight drug trafficking and violent crime more generally? It feels to me like there is a technical-military-industrial complex mentality in government where if you build the systems then you just want to use them. In the same way that CCTV cameras record you irrespective of whether or not you commit a violent crime or shoplift, we need to ask whether the same panoptical systems of surveillance should be extended to the Internet. Now, to a large extent they are already there. But what should we train the torchlight on next?

This takes us back to the importance of having necessary, proportionate, and independently authorized processes. When you drill down into how rights privacy should be balanced with security then it gets really complicated. But the basic process-driven things that we identified in the report are far simpler: if we accept that governments have the right to take certain actions in the name of security, then, no matter how important or life-saving those actions are, there are still protocols that governments must follow. We really wanted to infuse these issues into the debate through the report.

Read the full report: Brown, I., and Cowls, J., (2015) Check the Web: Assessing the Ethics and Politics of Policing the Internet for Extremist Material. VOX-Pol Publications.

Josh Cowls is a a student and researcher based at MIT, working to understand the impact of technology on politics, communication and the media.

Josh Cowls was talking to Blog Editor Bertie Vidgen.

New Voluntary Code: Guidance for Sharing Data Between Organisations

Many organisations are coming up with their own internal policy and guidelines for data sharing. However, for data sharing between organisations to be straight forward, there needs to a common understanding of basic policy and practice. During her time as an OII Visiting Associate, Alison Holt developed a pragmatic solution in the form of a Voluntary Code, anchored in the developing ISO standards for the Governance of Data. She discusses the voluntary code, and the need to provide urgent advice to organisations struggling with policy for sharing data.

Collecting, storing and distributing digital data is significantly easier and cheaper now than ever before, in line with predictions from Moore, Kryder and Gilder. Organisations are incentivised to collect large volumes of data with the hope of unleashing new business opportunities or maybe even new businesses. Consider the likes of uber, Netflix, and Airbnb and the other data mongers who have built services based solely on digital assets.

The use of this new abundant data will continue to disrupt traditional business models for years to come, and there is no doubt that these large data volumes can provide value. However, they also bring associated risks (such as unplanned disclosure and hacks) and they come with constraints (for example in the form of privacy or data protection legislation). Hardly a week goes by without a data breach hitting the headlines. Even if your telecommunications provider didn’t inadvertently share your bank account and sort code with hackers, and your child wasn’t one of the hundreds of thousands of children whose birthdays, names, and photos were exposed by a smart toy company, you might still be wondering exactly how your data is being looked after by the banks, schools, clinics, utility companies, local authorities and government departments that are so quick to collect your digital details.

Then there are the companies who have invited you to sign away the rights to your data and possibly your privacy too – the ones that ask you to sign the Terms and Conditions for access to a particular service (such as a music or online shopping service) or have asked you for access to your photos. And possibly you are one of the “worried well” who wear or carry a device that collects your health data and sends it back to storage in a faraway country, for analysis.

So unless you live in a lead-lined concrete bunker without any access to internet connected devices, and you don’t have the need to pass by webcams or sensors, or use public transport or public services; then your data is being collected and shared. And for the majority of the time, you benefit from this enormously. The bus stop tells you exactly when the next bus is coming, you have easy access to services and entertainment fitted very well to your needs, and you can do most of your bank and utility transactions online in the peace and quiet of your own home. Beyond you as an individual, there are organisations “out there” sharing your data to provide you better healthcare, education, smarter city services and secure and efficient financial services, and generally matching the demand for services with the people needing them.

So we most likely all have data that is being shared and it is generally in our interest to share it, but how can we trust the organisations responsible for sharing our data? As an organisation, how can I know that my partner and supplier organisations are taking care of my client and product information?

Organisations taking these issues seriously are coming up with their own internal policy and guidelines. However, for data sharing between organisations to be straight forward, there needs to a common understanding of basic policy and practice. During my time as a visiting associate at the Oxford Internet Institute, University of Oxford, I have developed a pragmatic solution in the form of a Voluntary Code. The Code has been produced using the guidelines for voluntary code development produced by the Office of Community Affairs, Industry Canada. More importantly, the Code is anchored in the developing ISO standards for the Governance of Data (the 38505 series). These standards apply the governance principles and model from the 38500 standard and introduce the concept of a data accountability map, highlighting six focus areas for a governing body to apply governance. The early stage standard suggests considering the aspects of Value, Risk and Constraint for each area, to determine what practice and policy should be applied to maximise the value from organisational data, whilst applying constraints as set by legislation and local policy, and minimising risk.

I am Head of the New Zealand delegation to the ISO group developing IT Service Management and IT Governance standards, SC40, and am leading the development of the 38505 series of Governance of Data standards, working with a talented editorial team of industry and standards experts from Australia, China and the Netherlands. I am confident that the robust ISO consensus-led process involving subject matter experts from around the world, will result in the publication of best practice guidance for the governance of data, presented in a format that will have relevance and acceptance internationally.

In the meantime, however, I see a need to provide urgent advice to organisations struggling with policy for sharing data. I have used my time at Oxford to interview policy, ethics, smart city, open data, health informatics, education, cyber security and social science experts and users, owners and curators of large data sets, and have come up with a “Voluntary Code for Data Sharing”. The Code takes three areas from the data accountability map in the developing ISO standard 38505-1; namely Collect, Store, Distribute, and applies the aspects of Value, Risk and Constraint to provide seven maxims for sharing data. To assist with adoption and compliance, the Code provides references to best practice and examples. As the ISO standards for the Governance of Data develop, the Code will be updated. New examples of good practice will be added as they come to light.

[A permanent home for the voluntary code is currently being organised; please email me in the meantime if you are interested in it: Alison.holt@longitude174.com]

The Code is deliberately short and succinct, but it does provide links for those who need to read more to understand the underpinning practices and standards, and those tasked with implementing organisational data policy and practice. It cannot guarantee good outcomes. With new security threats arising daily, nobody can fully guarantee the safety of your information. However, if you deal with an organisation that is compliant with the Voluntary Code, then at least you can have assurance that the organisation has at least considered how it is using your data now and how it might want to reuse your data in the future, how and where your data will be stored, and then finally how your data will be distributed or discarded. And that’s a good start!

Alison Holt was an OII Academic Visitor in late 2015. She is an internationally acclaimed expert in the Governance of Information Technology and Data, heading up the New Zealand delegations to the international standards committees for IT Governance and Service Management (SC40) and Software and Systems Engineering (SC7). The British Computer Society published Alison’s first book on the Governance of IT in 2013.

Controlling the crowd? Government and citizen interaction on emergency-response platforms

There is a great deal of interest in the use of crowdsourcing tools and practices in emergency situations. Gregory Asmolov‘s article Vertical Crowdsourcing in Russia: Balancing Governance of Crowds and State–Citizen Partnership in Emergency Situations (Policy and Internet 7,3) examines crowdsourcing of emergency response in Russia in the wake of the devastating forest fires of 2010. Interestingly, he argues that government involvement in these crowdsourcing efforts can actually be used to control and regulate volunteers from the top down — not just to “mobilize them”.

RUSSIA, NEAR RYAZAN - 8 MAY 2011: Piled up woords in the forest one winter after a terribly huge forest fires in Russia in year 2010. Image: Max Mayorov. — RUSSIA, NEAR RYAZAN – 8 MAY 2011: Piled up wood in the forest one winter after a terribly huge forest fire in Russia in year 2010. Image: Max Mayorov (Flickr).

My interest in the role of crowdsourcing tools and practices in emergency situations was triggered by my personal experience. In 2010 I was one of the co-founders of the Russian “Help Map” project, which facilitated volunteer-based response to wildfires in central Russia. When I was working on this project, I realized that a crowdsourcing platform can bring the participation of the citizen to a new level and transform sporadic initiatives by single citizens and groups into large-scale, relatively well coordinated operations. What was also important was that both the needs and the forms of participation required in order to address these needs be defined by the users themselves.

To some extent the citizen-based response filled the gap left by the lack of a sufficient response from the traditional institutions.[1] This suggests that the role of ICTs in disaster response should be examined within the political context of the power relationship between members of the public who use digital tools and the traditional institutions. My experience in 2010 was the first time I was able to see that, while we would expect that in a case of natural disaster both the authorities and the citizens would be mostly concerned about the emergency, the actual situation might be different.

Apparently the emergence of independent, citizen-based collective action in response to a disaster was considered as some type of threat by the institutional actors. First, it was a threat to the image of these institutions, which didn’t want citizens to be portrayed as the leading responding actors. Second, any type of citizen-based collective action, even if not purely political, may be an issue of concern in authoritarian countries in particular. Accordingly, one can argue that, while citizens are struggling against a disaster, in some cases the traditional institutions may make substantial efforts to restrain and contain the action of citizens. In this light, the role of information technologies can include not only enhancing citizen engagement and increasing the efficiency of the response, but also controlling the digital crowd of potential volunteers.

The purpose of this paper was to conceptualize the tension between the role of ICTs in the engagement of the crowd and its resources, and the role of ICTs in controlling the resources of the crowd. The research suggests a theoretical and methodological framework that allows us to explore this tension. The paper focuses on an analysis of specific platforms and suggests empirical data about the structure of the platforms, and interviews with developers and administrators of the platforms. This data is used in order to identify how tools of engagement are transformed into tools of control, and what major differences there are between platforms that seek to achieve these two goals. That said, obviously any platform can have properties of control and properties of engagement at the same time; however the proportion of these two types of elements can differ significantly.

One of the core issues for my research is how traditional actors respond to fast, bottom-up innovation by citizens.[2]. On the one hand, the authorities try to restrict the empowerment of citizens by the new tools. On the other hand, the institutional actors also seek to innovate and develop new tools that can restore the balance of power that has been challenged by citizen-based innovation. The tension between using digital tools for the engagement of the crowd and for control of the crowd can be considered as one of the aspects of this dynamic.

That doesn’t mean that all state-backed platforms are created solely for the purpose of control. One can argue, however, that the development of digital tools that offer a mechanism of command and control over the resources of the crowd is prevalent among the projects that are supported by the authorities. This can also be approached as a means of using information technologies in order to include the digital crowd within the “vertical of power”, which is a top-down strategy of governance. That is why this paper seeks to conceptualize this phenomenon as “vertical crowdsourcing”.

The question of whether using a digital tool as a mechanism of control is intentional is to some extent secondary. What is important is that the analysis of platform structures relying on activity theory identifies a number of properties that allow us to argue that these tools are primarily tools of control. The conceptual framework introduced in the paper is used in order to follow the transformation of tools for the engagement of the crowd into tools of control over the crowd. That said, some of the interviews with the developers and administrators of the platforms may suggest the intentional nature of the development of tools of control, while crowd engagement is secondary.

[1] Asmolov G. “Natural Disasters and Alternative Modes of Governance: The Role of Social Networks and Crowdsourcing Platforms in Russia”, in Bits and Atoms Information and Communication Technology in Areas of Limited Statehood, edited by Steven Livingston and Gregor Walter-Drop, Oxford University Press, 2013.

[2] Asmolov G., “Dynamics of innovation and the balance of power in Russia”, in State Power 2.0 Authoritarian Entrenchment and Political Engagement Worldwide, edited by Muzammil M. Hussain and Philip N. Howard, Ashgate, 2013.

Read the full article: Asmolov, G. (2015) Vertical Crowdsourcing in Russia: Balancing Governance of Crowds and State–Citizen Partnership in Emergency Situations. Policy and Internet 7,3: 292–318.

Gregory Asmolov is a PhD student at the LSE, where he is studying crowdsourcing and emergence of spontaneous order in situations of limited statehood. He is examining the emerging collaborative power of ICT-enabled crowds in crisis situations, and aiming to investigate the topic drawing on evolutionary theories concerned with spontaneous action and the sustainability of voluntary networked organizations. He analyzes whether crowdsourcing practices can lead to development of bottom-up online networked institutions and “peer-to-peer” governance.