Author: Anoush Margaryan

Nature of work tasks in crowdwork: Microwork vs online freelancing

As part of the survey of crowdworkers’ learning practices, I explored the nature of tasks they undertake within crowdwork platforms using two sets of previously validated and published scales: the Classification Structure of Knowledge-intensive Processes and some items from the ‘Knowledge Characteristics’ sub-scale of the Work Design Questionnaire.

In particular, the participants were asked to indicate which of the following 15 task types most closely described their typical crowdwork tasks (they could choose all options that applied):

  1. My crowdwork tasks are mostly routine
  2. My crowdwork tasks are highly reliant on formal processes
  3. My crowdwork tasks don’t give me freedom to decide what should be done in any particular situation
  4. My crowdwork tasks are mostly systematically repeatable
  5. My crowdwork tasks are highly reliant on formal standards
  6. My crowdwork tasks are dependent on integration across functional or disciplinary boundaries
  7. My crowdwork tasks are improvisational/creative
  8. My crowdwork tasks are highly reliant on my deep expertise/personal judgement
  9. My crowdwork tasks are dependent on collaborating with others
  10. My crowdwork tasks are highly reliant on my own individual expertise/experience
  11. My crowdwork tasks involve solving problems that have no obvious correct answer
  12. My crowdwork tasks involve dealing with problems I have not met before
  13. My crowdwork tasks require unique ideas/solutions to problems
  14. My crowdwork tasks require me to use a variety of skills to complete the work
  15. My crowdwork tasks require me to use a number of complex or high-level skills

Here are the results from the two groups: microworkers (Figure 1) and online freelancers (Figure 2). Among microworkers the three most prevalent characterisations of tasks were routine, systematically repeatable and requiring a variety of skills. And among online freelancers the most prevalent tasks were those that required a variety of skills and uniquie ideas and solutions, dealing with novel problem, and were reliant on their own individual expertise and experience. 

NatureOfTasksMW

Figure 1. Microworkers’ perceptions of the nature of their crowdwork tasks 

NatureOfTasksOF

Figure 2. Online freelancers’ perceptions of the nature of their crowdwork tasks

I’m working on a paper comparing these findings with earlier findings from a study of how conventional workers describe the nature of their tasks to see if there are any statistically significant differences and whether a similar typology emerges from the crowdwork settings. Stay tuned!

 

 

Advertisements

Workplace learning activities in crowdwork: 2016 vs 2017 survey

Yesterday, I wrote about the initial findings of a survey scoping the range and frequency of use of workplace learning activities among crowdworkers.   

This is the second iteration of the survey, the first survey was carried out in March 2016 within the same two platforms, but involving different participants. In the 2016 sample a larger proportion of crowdworkers reported using social learning activities, as well as online courses and tutorials. 

I plan to compare the demographic details and other key characteristics of the participants of both survey iterations to identify potential explanations for these differences.

How do crowdworkers learn? Workplace learning activities

This is the first blogpost in the series about learning practices of crowdworkers.

One of the foci of the study is how crowdworkers learn on-the-job: what  types of workplace learning activities they undertake and what learning strategies they use to self-regulate their learning.   The range and frequency of use of learning activities and learning strategies that people undertake in the workplace give us an indication of learning-intensity of a job (that is, the extent to which people need to regularly acquire new skills and knowledge to be able to maintain their job).   Crowdwork is often presumed to be low learning-intensity, low-skill, lacking in professional development opportunities and preventing workers for applying and developing their skills and know-how.  So it’s useful to scope the range and frequency of use of workplace learning activities and strategies among crowdworkers to see what, if any, empirical base there is to these assumptions about crowdwork.

This blogpost is focused on workplace learning activities. The resuts reported here are based on 182 survey responses (167 microworkers and 15 online freelancers). I’m currently collecting more survey responses from online freelancers in order to balance out the sample.

In the survey, I asked the participants to indicate how frequently they used the following 14 workplace learning activities within the last 3 months as part of their work on the crowdwork platforms (never, rarely, frequently or very frequently):

  1. Acquiring new information to complete their crowdwork tasks
  2. Working alone to complete their crowdwork tasks
  3. Collaborating with others to complete their crowdwork tasks
  4. Following new developments in their field
  5. Performing tasks that are new to them
  6. Asking others for advice
  7. Attending a training course/workshop to acquire knowledge/skills for their crowdwork
  8. Taking free online courses or webinars (e.g. Coursera) to acquire knowledge/skills for crowdwork
  9. Using paid online tutorials (e.g. Lynda) to acquire knowledge/skills for crowdwork
  10. Reading articles/books to acquire knowledge/skills for crowdwork
  11. Observing/replicating other people’s strategies
  12. Finding a better way to do a task by trial-and-error
  13. Thinking deeply about their work (e.g. what they could do better next time)
  14. Receiving feedback on their crowdwork tasks (e.g. from a client or peers)

Below are the survey results showing the percentage of the crowdworkers who reported using each learning activity ‘frequently’ or ‘very frequently’.

WLAOverallFigure 1. Percentage of crowdworkers who reported using these workplace learning activities frequently or very frequently

From this chart, crowdworkers most frequently learn by working alone on novel tasks, acquiring new information, following new developments in their fields, seeking better ways to do the tasks by trial-and-error and reflecting deeply on their work.

Crowdworkers reported some other learning activities not included in the list above, such as:

  • Watching YouTube videos
  • Participating in project groups on specific tasks
  • Participating in platform-specific online fora
  • Discussing ideas with others
  • Reading platform-specific blogs
  • Watching news in foreign languages to improve language skills
  • Taking private lessons to improve skills in a particular area
  • Scoping and learning highly-demanded technologies trending on specific crowdwork platforms

It is  interesting that just over a third of the respondents reported observing and replicating other people’s strategies as a key way in which they learn.  Also, some undertake social learning activities such as asking others for advice, collaborating with others, receiving feedback. How does this sort of mimetic and cooperative learning take place in a distributed, digital online workplace? What are the underpinning mechanisms and processes and what is the nature of the connections? These questions will be further explored in the interviews with crowdworkers.

Crowdworkers’ learning practices

I’ve been conducting surveys of crowdworkers’ learning practices as part of the ‘Learning in Crowdwork’ research project I am leading.

In particular, recently (in October 2017) I surveyed two groups of crowdworkers from two platforms: microworkers from CrowdFlower and online freelancers from Upwork.

As the data are being analysed, there are some interesting early findings that I’d like to share. Over the next days and weeks I’ll be publishing a series of blog posts to outline some of the most noteworthy initial results.

The questionnaire I used is based on a validated and published instrument, the Self-regulated Learning at Work Questionnaire (SRLWQ) I co-authored and now adapted for use within crowdwork settings. I’m doing some analyses to further validate the adapted version of the questionnaire hoping to publish it in due course.

On selfishness

The conventional view of ‘selfishness’ takes into account only one dimension of this concept, that is self-interest vs. another’s interest.

At least two other key dimensions are neglected.

First, the nature of the action being carried out in the interest of self or another, that is good deed vs. evil deed (e.g. production, creation vs parasitism, robbery, sacrificing others).

Second, the nature of the desire underpinning the action being carreid out in the interest of self or another, that is rationality vs irrationality.

Slide1

 

 

 

 

 

 

 

 

 

Short talk on Future of Work and Skills

Last week I gave a short talk on the future of work and skills at the 10th anniversary of People Per Hour online freelancing platform in London. Here is the text of the talk.

Good evening and happy birthday, People per Hour! Thank you very much for the opportunity to join the celebration!

I was asked to share brief reflections on the future of work and learning. So I’d like to offer a few general observations about the changing nature of work, then outline a few specific skills and attributes that I think will be crucial moving forward.

Changing nature of work

I want to begin by suggesting that much of the current discourse on the future of work –that intelligent machines will destroy jobs causing mass unemployment and societal upheaval –  seems ahistorical. Historically, introduction of new technologies has always destroyed some jobs and created others. There doesn’t seem to be a compelling reason to expect that the current cycle of technology development will be radically different in terms of this fundamental pattern.

Secondly, just because jobs and work tasks can be automated doesn’t mean that they will be. Technology is not exogenous to our society. It is part of an overall socio-technical system of the society. Historically, the deployment of technology has been enabled as well as constrained by broader forces – economic, legal, ethical, societal and political considerations. We don’t know just exactly how these broader factors will play out and how they may play out differently in different countries.

Thirdly, scholars studying the impact of artificial intelligence have pointed out that currently few if any intelligent machines and systems are truly autonomous, but that rather the role of humans is often obscured in these systems.  Madelene Elish, who is a cultural anthropologist at Columbia University, argued that “when we examine autonomous systems we must look not to the erasures of the human, but to the ways in which we, as humans, are newly implicated”…

Information scientists Bonnie Nardi and Hamid Ekbia coined the term ‘heteromation’ to denote the paradigm of work premised on the division of labour between machines and humans. They argue that unlike the technologies of automation, the aim of which is “to disallow human intervention at nearly all points in the system”, the technologies of heteromation “push critical tasks to end users as indispensable mediators, drawing the humans back into the computational fold”. This sorts of heteromative systems are currently in evidence, from Amazon Mechanical Turk to factory floors in Germany and China where humans increasingly collaborate with robots in our production systems. It seems likely that at least in the medium-term, it is such heteromative arrangements rather than the pure automation paradigm that will predominate.

The historian Yuval Harari has pointed out that “not just the idea of ‘a job for life’, but even the idea of a ‘profession for life’ will soon come to seem antediluvian”. Harari’s argument suggests that we will have to be flexible, not just in terms of our knowledge and skills, but also sector and the location where we work, our professional identity and be prepared to drastically reorient ourselves several times within our careers.

This takes me to the future of learning and skills…

Future of learning and skills

I’d like to suggest to you that, as we undergo a range of these horizontal and vertical reorientations implied in Harari’s argument, the ability to actively initiate rather than only endure these transitions will become a crucial mindset.  Organisational scholars Lynda Gratton and Andrew Scott pointed out that people who are successful in undergoing transitions share at least three important abilities. First, they are self-reliant and possess the necessary self-knowledge to understand who we they are and what they may be in future. Second, they have dynamic and heterogeneous networks to provide them with role models of what they could be and how to become it. And third, they are risk takers, they have tolerance for uncertainty, and openness to experience “to act their way into change”.

I’d argue that presently our public discourse and policy landscape is overly focused on the formal, institutionalised forms of education and training. Limited recognition and value is given to self-directed and self-initiated forms of learning that people undertake throughout their lives, individually or in cooperation with others. The problem is that the educational institutions we have come to rely on for our learning and training are very conservative and slow. And it is questionable that they will be able to provide the rapid and dynamic reskilling and reorientation that individuals will require.

So I’d like to suggest that moving forward the most important attribute that individuals must acquire is the ability to self-regulate and self-direct their learning. This means, being strategic and dynamic in identifying our learning goals and strategies, being proactive in seeking feedback on our learning and work, continuously studying the market to understand and identify the changing skill requirements, strengthening our personal self-efficacy, being self-reflective and able to dynamically change our learning strategies when these are not working. These attributes will be increasingly required of everyone, not just those in managerial or highly-skilled jobs. These individual attributes have always been important psychologically, but I’d argue that they have now become crucially important economically as well.

My last point is that moving forward some radically new skills will also be required. One of these is for us to learn to cooperate effectively with the intelligent machines in our increasingly heteromative workplaces. The emergent forms of algorithmic management of labour necessitate the ability to work productively in a supervisory relationship with a machine. A key question I would like to ask is this: in a workplace where your boss is a machine, where algorithms allocate and dynamically price work, algorithms record and evaluate the quality of the output, algorithms assess workers’ performance and recommend learning activities to improve performance what are the skills, attributes and dispositions needed to work productively with these machines and how can people develop these capabilities?  Presently, our policy landscape appears to overlook these fundamentally new skill requirements altogether and our educational and vocational systems are doing little to prepare people for this new reality.

Let me conclude by observing that to succeed in the future of work, we have a greater need for personal autonomy, self-reliance and self-direction. At least for some time to come there will be no accepted code of rules to absolve us of the challenge of individual decision making in our work and in our learning. Thank you.

Using life course perspective to understand learning practices within crowdwork

I am pleased to have an abstract accepted for the ‘Research Methods for Digital Work: Innovative Methods for Studying Distribute and Multi-modal Working Practices’ conference organised by the Institute of Advanced Studies at the University of Surrey.

In the abstract, Heather Hofmeister and I outline how the Life Course perspective could help advance research into digital work and particularly the study of learning practices within crowdwork.  To my knowledge, this is the first attempt to apply the Life Course perspective in crowdwork research. If you know of other research where the Life Course perspective has been used to study crowdwork or other gig-economy practices please get in touch.

Abstract: Using life course perspective to understand learning practices within crowdwork

Anoush Margaryan and Heather Hofmeister, Department of Sociology, Goethe University Frankfurt 

Background

The unfolding digitalisation of our society has stimulated the development of new types of work practices termed ‘virtual work’ (Huws, 2014). These emerging digitally-mediated work practices challenge traditional patterns of individual agency, organisation, power, responsibility and learning (Littlejohn and Margaryan, 2014). Working under conditions of precarity, digital transformation and changing patterns of agency, workers increasingly have to initiate and regulate their own learning. As the nature of work evolves, understanding how workers learn within these new work practices becomes increasingly important.

One type of virtual work is crowdwork – a form of labour in which a large group of people are brought together within Internet-based platforms for the purpose of performing a task. These platforms act as intermediaries between clients/task requesters and workers, helping oversee the definition, submission, acceptance and payment for the work done (Kuek et al, 2015). Examples are Amazon Mechanical Turk, Upwork and CrowdFlower.  In this highly distributed and fragmented type of work, where workers may not have access to the learning opportunities available within traditional employment (eg training or access to experienced colleagues), how do crowdworkers go about managing their learning? What strategies do crowdworkers use to identify their learning needs, source knowledge, and find others to learn with and from?

In crowdwork research, three key methods have been used – questionnaire survey, interview, and ethnography- all focusing on crowdworkers (Gray et al, 2016; Ipeirotis, 2010; Martin et al, 2016). The perspectives of other key stakeholders e.g. platform providers and task requesters/clients have been overlooked. Also, workers’ experiences have been examined as snapshots rather than being contextualised historically and developmentally.

This paper argues that essential to understanding the learning practices within crowdwork is to analyse both the individual and the historical-environmental factors impacting crowdworkers’ learning. Crowdwork occurs largely online, although it is plausible that crowdworkers’ learning activities span the online and offline realm. Methodological approaches that bridge sociological, psychological, individual, collective, online, offline, and temporal processes and practices of learning within crowdwork are needed.

Life course perspective

We propose the life course perspective as an analytical framework to facilitate a nuanced, contextualised analysis of crowdworkers’ learning (Elder and Giele, 2009; Hofmeister, 2015; Levy, 2013). The life course is an interdisciplinary perspective drawing on sociology, psychology, anthropology, history and biology to help understand human development across the life span (Mortimer and Shanahan, 2004). The life course perspective stresses the importance of the socio-cultural environment in explaining individual behaviour and life history. It focuses on the interplay of the individual, their setting, and the dynamic processes of change individuals undergo within these settings.

Four key elements of the life course perspective would help analyse crowdworkers’ learning: agency, context, linked lives, and timing. Agency refers to an individual’s motives and goals, and the self-regulated activities undertaken to fulfil them (Elder, 1994). Context refers to the setting in which each individual acts, comprising psychological, social, cultural, organisational, technological and physical dimensions, as well as the temporally-constituted patterns that emerge from the interplay between these diverse contexts (Blossfeld, 2009; O’Rand, 2009). Linked lives denotes the interrelations between individuals in their contexts (Moen and Hernandez, 2009). Timing refers to the sequencing of events and pathways of personal activities individuals engage in to reach their goals (Viry et al, 2013).

Quantitative and qualitative methods have been applied in Life Course research (Elder and Giele, 2009). For example, Laub and Sampson (1998) discuss a longitudinal study of juvenile delinquency where data on social, psychological and biological characteristics, family life, school performance, work experience were collected from multiple sources, several points of view and at separate times. They show how the merging of quantitative and qualitative data provides important cues for explaining continuity and change in human behaviour.

Applying life course perspective to crowdworkers’ learning    

The project this paper is based on examines learning strategies and activities, personal motivations, goals, agency beliefs and pathways underpinning crowdworkers’ learning, and the individual and environmental factors impacting upon their learning. Data are collected from two platforms: CrowdFlower and Upwork.

 Several methods are combined to help generate rich descriptions of crowdworkers’ learning practices (Johnson et al., 2004). Crowdworkers’ self-regulatory learning strategies and learning activities are scoped using the Self-Regulated Learning at Work Questionnaire, SRLWQ (Fontana et al., 2015). The survey is supplemented by biographical interviews to ascertain crowdworkers’ professional trajectories and learning pathways, educational and work experiences, current and desired skills, learning goals and motivations to engage in crowdwork and learning.  The interviews are combined with field visits to conduct observations of crowdworkers’ local contexts and to collect data on specific SRL strategies in situ using SRL microanalysis protocols (Cleary, 2011).  Online ethnography is carried out within discussion fora and social networks used by crowdworkers and clients/task requesters to identify learning activities. To contextualise crowdworkers’ perspectives, representatives of crowdwork platforms and selected task requesters are interviewed and training and development provisions offered by the platforms are scoped. Table 1 illustrates how these methods help elucidate the key components of the life course applied to crowdworkers’ learning.

Table 1.  Mapping of methods and the life course framework

Life course components Methods
Human agency

·       Motives to engage in crowdwork

·       Learning and performance goals

·       Career aspirations

·       SRL strategies

·       Learning activities

·       Existing knowledge and skills

 

·       Biographical interviews with crowdworkers

·       Analysis of online fora

·       SRLWQ

·       SRL microanalysis

·       Experience sampling methods (eg diary or tracking devices)

Context

·       Design of crowdwork platforms

·       Task design

·       Physical environment

·       Local infrastructure

·       Local culture

·       Other work/professional commitments

·       Education and training

·       Previous work experiences

·       Local economic conditions, employment and regulatory regimes

·       Training and development provision by the platforms

·       Review of platforms

·       Scoping of sample work tasks

·       Interviews with platform providers and clients

·       Interviews with crowdworkers

·       Interviews with policymakers (trade unions, labour organisations, politicians)

·       Ethnographic observation

·       Document review

·       Analysis of artefacts

Linked lives

·       Family, friends, neighbours

·       Professional networks in and outside crowdwork

·       Clients and employers

·       Online communities

·       Client networks

 ·       Biographical interviews

·       Interviews with platforms and clients

·       Opportunistic interviews with family members/friends

·       Social network analysis

Timing

·       Education

·       Workplace

·       Retirement

·       Disability

·       Immigration

·       Loss of job

·       Decision to freelance

·       Family events

·       Significant other prior events/experiences

·       Biographical interviews

The data collection is in early stages and specific examples of data capture and analytic techniques will be demonstrated at the conference. Opportunities and challenges in mixing methods to study crowdworkers’ learning will be discussed.

The paper contributes an interdisciplinary methodological perspective drawing on Sociology, Learning Sciences, Psychology, Internet Studies and HCI offering insight into how people work and learn within crowdwork and how crowdwork may be shaped to foster learning. 

References

Blossfeld, H. P. (2009). Comparative Life Course research. In G. H. Elder, & J. Z. Giele (Eds.), The craft of Life Course research (pp. 280-306). New York: Guilford Press.

Cleary, T. (2011). Emergence of self-regulated learning microanalysis. In Zimmerman, B., & Schunk, D. (Eds.), Handbook of self-regulation of learning and performance. London: Routledge.

Elder, G. H. (1994). Time, human agency, and social change. Social Psychology Quarterly, 57(1), 4-15.

Elder, G. H., & Giele, J. Z. (Eds.). (2009). The craft of Life Course research. New York: Guilford Press.

Fontana, P. et al (2015). Measuring self-regulated learning in the workplace. International Journal of Training and Development, 19(1), 32-52.

Gray, M. et al (2016). The crowd is a collaborative network. In Proceedings of CSCW 2016 Conference (pp. 134-147). San Francisco: ACM.

Hofmeister, H. (2010). Life Course. In S. Immerfall, & G. Therborn (Eds.), Handbook of European societies (pp. 385-411). New York: Springer.

Hofmeister, H. (2015). Individualisation of the life course. International Social Science Journal, 64(213), 279-290.

Huws, U. (2014). Labour in the global digital economy. New York: Monthly Review Press.

Ipeirotis, P. (2010). Demographics of Mechanical Turk. http://www.ipeirotis.com/wp-content/uploads/2012/02/CeDER-10-01.pdf

Johnson, R., & Onwuegbuzie, A. (2004). Mixed methods research. Educational Researcher, 33(7), 14-26.

Kuek, S. C., et al. (2015). The global opportunity in online outsourcing. Washington, DC: World Bank.

Laub, J. H., & Sampson, R. J. (1998). Integrating quantitative and qualitative data. In J. Z. Giele, & G. H. Elder, Jr. (Eds.), Methods of Life Course research (pp. 213-230). Thousand Oaks, CA: Sage.

Laub, J. H. et al (1998). Trajectories of change in criminal offending. American Sociological Review, 63(2), 225-238.

Levy, R. (2013). Life Course analysis. In R. Levy, & E. D. Widmer (Eds.), Gendered life courses between standardization and individualization (pp. 315-338). Zürich: LIT.

Littlejohn, A., & Margaryan, A. (2014). Technology-enhanced professional learning. London: Routledge.

Martin, D., et al. (2016). Turking in a global labour market. Computer-Supported Cooperative Work, 25(1), 39-77.

Moen, P., & Hernandez, E. (2009). Social convoys. In Elder, G. H., & Giele, J. Z. (Eds.), The craft of Life Course research (pp. 258-79). New York: Guilford Press.

Mortimer, J.T., & Shanahan, M.J. (2004) (Eds.). Handbook of the life course. New York: Springer.

O’Rand, A. M. (2009). Cumulative processes in the Life Course. In G. H. Elder, & J. Z. Giele (Eds.), The craft of Life Course research (pp. 121-140). New York: Guilford Press.

Viry, G. et al (2013). Residential trajectories in the early life course and their effects. In Levy, R., & Widmer, E. D. (Eds.), Gendered life courses between standardization and Individualization (pp. 141-160). Zürich: LIT.