Archive

Posts Tagged ‘Digital Labour’

Social Media Analytics for Better Understanding of the Digital Gig Economy

27 April 2018 1 comment

Owing to the proliferation of digital platforms facilitating online freelance work such as Upwork, Fiverr and Amazon Mechanical Turk, the number of digital gig workers has been continuously increasing worldwide. In 2015, there were as many as 48 million digital gig workers [1]; between 2016 and 2017, a 25% increase in the number of such workers was reported [2].

Digital gig work is indeed attractive to many, with a number of benefits that such independent workers are perceived to enjoy, e.g., flexible working hours, reduced transportation costs, wide range of projects to choose from. However, there exist potentially distressing issues, e.g., lack of job security, tough competition, substandard wages, which are especially pronounced in developing country settings [3]. Whereas traditional media such as news were unable to pinpoint or bring attention to these concerns, social media analysis–done manually by Cision in 2017–provided a window to the thoughts of independent workers which led to the fine-grained identification of issues that they are faced with [4].

As part of the currently ongoing Social Media Analytics Research and Teaching @ Manchester (SMART@Manchester) project funded by the University of Manchester Research Institute (UMRI), we aim to automatically gain insight into people’s perceptions of digital gig work, based on their posts on social media platforms such as Twitter and Facebook, as well as on review sites such as Glassdoor.

Specifically, we wish to test the currently prevailing assumption that digital gig work is experienced differently in the Global South compared to the Global North. Workers tend to make comparisons with their local benchmarks (i.e., office-based work), and it is believed possible that in the Global North, digital gig work is worse than prevailing benchmarks, whereas in the Global South it is better.

The following are some of the research questions that will be addressed as part of this case study.

  1. How do digital gig workers feel about their jobs?
  2. Which topics pertaining to decent work standards do they frequently talk about?
  3. Are there any differences—in terms of sentiments and topics—across different geographic locations, or across genders?

The first question can be answered by opinion mining while the second is addressable by topic identification. To determine whether there are differences with respect to opinions and topics, between the Global North and South or between genders, results from opinion mining and topic identification need to be combined with social media content metadata (e.g., geographic locations). 

In the way of opinion mining, we are currently investigating the use of an automatic emotion identification tool called Illuemotion which was developed by University of Manchester final-year Computer Science student, Elitsa Dimova. The web-based tool, a screenshot of which is provided below, is underpinned by a neural network model that analyses tweets to determine the most dominant emotions expressed, which can be any of anger, fear, joy, love, sadness, surprise and thankfulness.

The image below shows one of the tweets directly fetched by the tool from Twitter (via their API) when supplied with “#upwork” as input query. The tweet, which speaks of hidden dangers of being a digital gig worker, was detected by Illuemotion as expressing sadness and fear. One of our next steps is to apply the tool on a collection of thousands of tweets to allow us to analyse them across different geographic regions as well as genders.

As we are analysing data that pertains to human emotion, ethical considerations are being taken into account, especially bearing in mind that we also do not wish to compromise any of the digital gig workers who are social media users. For example, many Twitter users are unaware that what they post publicly can be used to identify or (reverse) look them up. They also have a right to be forgotten (i.e., they can delete their posts as well as their accounts). Overall what this means for us researchers who make use of their data is that in scholarly publications, we should provide only aggregated results and ensure that we do not include any identifiable information. These and other ethical considerations were discussed in detail in the recently concluded symposium in the Academy of Management Specialised Conference on Big Data entitled, “Ethical and Methodological Considerations for Management Research in the Digital Economy” held at the University of Surrey from the 18-20th April.

As well as two other SMART@Manchester case studies, the above described research questions on perceptions of digital gig work and our proposed approaches will be presented in the upcoming 4th International Workshop on Social Media World Sensors (Sideways 2018) co-located with the 15th European Semantic Web Conference to be held in Heraklion, Crete, Greece from the 3rd-7th June.

References:

[1] Kuek, S.C. et al. (2015) The Global Opportunity in Online Outsourcing. World Bank, Washington, DC. Available at: http://documents.worldbank.org/curated/en/138371468000900555/The-global-opportunity-in-online-outsourcing

[2] Lehdonvirta, V. (2017) The online gig economy grew 26% over the past year, The iLabour Project, Oxford Internet Institute. Available at: http://ilabour.oii.ox.ac.uk/the-online-gig-economy-grew-26-over-the-past-year/

[3] Heeks, R. (2017) Decent Work and the Digital Gig Economy: A Developing Country Perspective on Employment Impacts and Standards in Online Outsourcing, Crowdwork, etc, Centre for Development Informatics, Global Development Institute, University of Manchester. Available at: http://hummedia.manchester.ac.uk/institutes/gdi/publications/workingpapers/di/di_wp71.pdf

[4] Rubec, J. (2017) Study: The Dark Side of the Gig Economy, Cision. Available at: https://www.cision.com/us/2016/12/the-dark-side-of-the-gig-economy/

Advertisements

Do Outsourcing Clients Want Decent Digital Work?

22 December 2017 Leave a comment

There are growing concerns that digital gig work – supplied by platforms like Mechanical Turk, Upwork, Freelancer, etc – falls short of decent work standards.  (For further details see the working paper, “Decent Work and the Digital Gig Economy”.)  To address this, and as discussed previously in this blog, there are plans to encourage new ethical standards.

But almost all evidence on this to date comes from workers.  The voices of only a few platforms have been heard, and there seems to be no evidence from clients.  Yet clients are central to decent digital work standards: if they create incentives for platforms to improve, that will be a powerful motivation.  Conversely, if clients don’t care, it removes a key driving force from the gig economy ecosystem.

So, what evidence can be found?

Here, I summarise Babin, R., & Myers, P. (2015) Social responsibility trends and perceptions in global IT outsourcing, Proceedings of the Conference on Information Systems Applied Research, v8, n3663.  This in turn summarises results from surveys conducted during 2009-2014 by the International Association of Outsourcing Professionals.

The survey was specifically about corporate social responsibility (CSR) in IT outsourcing.  So: a) it is not exactly about digital gig work but a broader category of outsourcing; b) the survey may encourage some level of “virtue signalling”: respondents wanting to appear more socially-responsible than they are in reality.  Nonetheless, it offers some relevant guidance about client attitudes to decent digital work.

In general terms, half the respondents were US-based; half were non-US; a fair reflection of gig work clients.  They ranged from SMEs to multinationals and just over half had a written CSR policy.  They are thus larger and more formally-CSR-inclined than the modal micro-enterprise client for digital gig work, but important given the increasing involvement of firms in gig outsourcing.

Key findings include the following:

– Nearly half “often” or “always” gave preference to outsourcing providers who had demonstrable CSR capability.

– Nearly two-thirds expected CSR consideration to become “more” or “much more” important in their future IT outsourcing.

– The largest factor in evaluating CSR capabilities of an outsourcing provider was its labour practices (see figure below).

Figure: Key factors in evaluating the CSR capabilities of an outsourcing provider, survey median (IAOP, 2009-14)

At least for this group of clients, then, the type of labour practices covered by proposed decent digital work standards were the top CSR issue; and CSR was quite widespread as a determinant in digital-related outsourcing (only 5% said they never used CSR as a determinant).

This gives some basis for believing – at least among larger clients for digital gig work – that an appetite exists for better employment and working conditions; an appetite that can encourage platforms to change.

Decent Digital Work and the FairWork Foundation

31 October 2017 1 comment

How can we improve standards for digital gig workers: those undertaking micro-work and online freelancing via platforms like Upwork and Mechanical Turk?

The recent research paper – “Decent Work and the Digital Gig Economy” – explains why such standards are needed.  With up to 70m workers worldwide registered for online work and growth rates of 20-30% per year, this is already a sizeable activity.  It is especially popular with the c.80% of workers based in middle- and low-income countries, who often see online work as better than local alternatives.  However, this ignores the chronic precarity and structural inequality associated with such work: damaging outcomes that will only spread if nothing is done.

But what should be done?

The paper develops an inventory of “Decent Digital Work” standards.  This is a comprehensive set of guidelines that integrates two things: first, the global decent work standards set by the ILO; second, the actions needed to address specific digital gig economy problems.

A key value for this inventory is as a comparator with other decent work initiatives.  For example, the paper analyses the way in which two major initiatives – SA8000, and the Ethical Trading Initiative – do and do not cover the requirements for decent digital work.

Below, a further comparison is undertaken, between the Decent Digital Work standards, and the criteria adopted by the FairWork Foundation; an initiative aiming to rate and certify gig economy platforms.  The table indicates those elements which are the same in both standards; those where a completely-different element is included; and those where there is some variation in the element.

From this, three things can be concluded:

a) A number of Decent Digital Work standards are absent in the FairWork Foundation certification criteria. Several of these relate to the broader context for work, would be outwith the scope of an individual platform, and therefore are not relevant to platform certification. However, those identified under “Employment” and “Work Conditions” can form part of a further discussion to consider their relevance to certification.

b) Some elements (e.g. around access to digital work opportunities, and accounting for worker costs other than unpaid time) speak to the particular conditions of gig workers from the global South. This is the location for the great majority of gig workers: already for digital gig workers; increasingly for physical gig workers. As such, the FairWork Foundation must ensure its global North origins do not skew its focus.

c) The FairWork Foundation should review how prevalent the non-competition and non-disclosure agreement problems are, and whether they are worth including. (Human review of task instructions – something unlikely to be practicable for most platforms – appears to have been dropped from later versions of the certification criteria; hence, its inclusion in brackets.)

As noted in the Decent Work and the Digital Gig Economy paper’s action research agenda, next steps here would be:

– Survey of worker, client and platform views about identified standards.

– A multi-stakeholder dialogue to agree a minimum set of certification standards and evaluation methods.

– Parallel research on the impact of standards and certifications in the gig economy, and analysing the costs and benefits of interventions such as standards and certifications at micro- and macro-level.

This is just one example of the application of the Decent Digital Work standards.  We hope you can identify other uses . . .

%d bloggers like this: