Crowdwork: Work outside of firms and employment contracts

Crowdwork can be defined in general as: ‘[…] the strategy of outsourcing a job that would normally be performed by an employee to an organisation or private person via an open call to a mass of unknown actors’ (Reference PapsdorfPapsdorf, 2009: 69). As Reference Hertwig and PapsdorfHertwig and Papsdorf (2017: 526) have argued, crowdwork exhibits ‘only partial overlaps’ with the sharing economy. In the literature, various attempts have been made to give a typology of the various forms of crowdwork. Reference Mrass, Leimeister, Fortmann and KolocekMrass and Leimeister (2018), for instance, distinguish between seven different platform types: Context/text creation platforms, design platforms, innovation platforms, marketplace platforms, microtask platforms, testing platforms, and customer service/market research/sales platforms (p. 142). Alongside platform type, other commentators have introduced additional classificatory criteria, such as labour force pool, contract type, form of algorithmic control, and sources and mechanisms of establishing ‘digital trust’ (Reference Howcroft and Bergvall-KårebornHowcroft and Bergvall-Kåreborn, 2019). The authors essentially distinguish between four platform types on the basis of two criteria: whether the work is reliably paid (i.e. paid vs non-paid or speculative work), and whether the worker (such as a specialist freelancer) or the client is to be considered the initiating actor. A classification undertaken on the basis of qualitative expert interviews in German-speaking countries, meanwhile, distinguished four key platform types: innovation platforms, testing platforms, microjob platforms, and design platforms (Reference KawalecKawalec, 2019).

The difficulty of giving a clear definition of crowdworking makes it hard to assess how widespread the practice is, particularly if we wish to quantify this global phenomenon for the German context. In contrast to the USA, the economic significance of crowdworking platforms in Germany remains low. At the beginning of 2017, 32 crowdworking platforms either had their headquarters or at least one (physical) branch in Germany (Reference Mrass, Leimeister, Fortmann and KolocekMrass and Leimeister, 2018: 142). On the global scale, ‘both supply and demand are relatively low’ in Germany (Reference Pongratz and BormannPongratz and Bormann, 2017: 165).

The crowdworking platforms themselves do not provide reliable figures on the number of active crowdworkers using their systems; they either treat these figures as trade secrets or advertise their services on the back of figures that are likely exaggerated. We can also assume that many crowdworkers are active on more than one platform, so that any attempt to quantify the total number of crowdworkers in Germany or German-speaking countries using such figures would be of limited value, even assuming a greater level of transparency on the part of the platforms. Two studies have nonetheless attempted to produce as comprehensible and reliable an estimate as possible of the statistical significance of crowdworkers in Germany or German-speaking countries:

• • On the basis of a representative online survey (N = 2180), a comparative study by Reference Huws, Spencer and SyrdalHuws et al. (2017) estimates that 1.45 million or 2.5% of working-age Germans earn at least 50% of their income from crowdworking. On the basis of a narrower definition, the study also identifies a ‘core group’ of crowdworkers consisting of those who also use a special app to stay informed of job offers. The study estimates that circa 1.07 million of these ‘professional’ crowdworkers live in Germany, which equates to 1.9% of the working-age population

• • In their own study, Reference Pongratz and BormannPongratz and Bormann (2017) proceed more carefully and criticise the Huws et al. study as ‘misleading’ (p. 180) on methodological grounds. They estimate that 500,000 to a million people have set up a profile (that still exists) on a platform for online working in Germany in the last 10 years. Of these, 100,000–300,000 take on at least one job per month, while only 1000–5000 people earn a secure income from this work (in comparison with the average German wage (Reference Pongratz and BormannPongratz and Bormann, 2017: 167).

Crowdwork can certainly be advantageous for workers, particularly in regions with a poorly developed labour market. This was affirmed by a long-term study on crowdwork in Sub-Saharan Africa and South-East Asia, which stated that ‘There is no simple story of exploitation […]’ (Reference Graham, Hjorth and LehdonvirtaGraham et al., 2017: 153). Nevertheless, the study also shows that the risks of this working model increase as more workers depend on it to earn a living. Similarly, a study on platform operators in Germany (Reference Mrass, Leimeister, Fortmann and KolocekMrass and Leimeister, 2018: 145–148) has shown that crowdworkers can benefit from greater working flexibility and from an additional source of income, but have to cope with relatively low pay, a lack of social security benefits, and income fluctuations. Crowdsourcers, on the other hand, not only benefit from greater flexibility and external expertise; they can also reduce their outgoings and save on social security contributions.

New forms of digital work thus seem to resurrect the ‘old’ conflicts of interest over income levels and income security evident in in traditional labour-market institutions and negotiation processes. Where crowdworking is concerned, the lack of regulation is a systemic rather than an accidental feature of the sector, or so to speak: a feature, not a bug. For unlike other forms of platform-mediated work that are undertaken in the real world, purely online crowdwork is characterised by the ‘de-territorialisation’ and ‘de-personalisation’ of work (Reference Menz, Tomazic, Dabrowski and WolfMenz and Tomazic, 2017: 14–15). On the one hand, this has an impact on distributive performance-related justice, since it serves to uncouple the global wage differential from local living costs. On the other hand, it affects procedural performance-related justice, since it leads to a lack of transparency in the relevant evaluation criteria and the possibility of workers influencing these, while also removing the social context provided by the (commercial) organisation. For crowdworkers, the consequences of these changes are of emotional quality (Reference Petriglier, Ashford and WrzesniewskiPetriglier et al., 2018).

Performance and the perception of injustice in the workplace

Reference DubetDubet’s (2008) qualitative study served to renew scholarly interest in the question of worker perceptions of injustice in the workplace, which he conceptualised as ‘normative activity’ in the dual sense of the ‘totality of principles that are considered legitimate’ and the ‘autonomous decision of an independent judgement’ (Reference DubetDubet, 2008: 17). On the basis of 350 interviews, Dubet argued that three basic principles underlie workers’ perceptions of injustice: equality, performance, and autonomy. For crowdworkers, the principles of equality and autonomy are less relevant, because crowdworkers have no possibility to compare themselves with other workers and are per definition fully autonomous. For this reason, we focus on performance. Performance-related justice means that one receives a just reward for work performed (Reference DubetDubet, 2008: 24). For Dubet, the concept of performance encompasses both ‘the objective result of the action’ and the ‘engagement on the part of the author of this action’. The ‘ambivalence of performance’ thus continually oscillates ‘between objective usefulness and effort expended’ (Reference DubetDubet, 2008: 127). Building on Dubet, two large qualitative studies were conducted in the last few years on justice expectations among employees in Germany.

• • Reference Hürtgen and VoswinkelHürtgen and Voswinkel (2014) carried out 42 biographical interviews with a cohort composed of ‘average’ employees, that is, middle-aged workers with permanent employment contracts and mid-level qualifications (Reference Hürtgen and VoswinkelHürtgen and Voswinkel, 2014: 39–40). In the interviews, the workers expressed ‘values of universal normative force’ on the basis of fundamental conceptions of themselves as human beings and social beings with a productive capacity. From the perspective of their productive capacity, their income-related expectations were articulated in terms of adequate remuneration for their efforts (Reference Hürtgen and VoswinkelHürtgen and Voswinkel, 2014: 139–140). From the perspective of their social existence, a good income was seen as one that ‘fit’ their lifestyle and their expectations concerning planning security and their rightful social security benefits and pension provision (Reference Hürtgen and VoswinkelHürtgen and Voswinkel, 2014: 145–150). In general, income was conceived both ‘as a reward for an […] accomplishment’ and as a ‘symbolic recognition of one’s productive capacity’ (Reference Hürtgen and VoswinkelHürtgen and Voswinkel, 2014: 149).

• • Another sociological study (Reference Kratzer, Menz and TulliusKratzer et al., 2015; similarly Reference Tullius and WolfTullius and Wolf, 2016) examined various forms of the legitimation of work on the basis of 320 qualitative interviews with non-managerial employees in Germany. As well as expecting opportunities for self-realisation in the workplace, involvement in decisions in their immediate working environment, and working conditions that upheld their dignity, the interviewees also expressed subjective expectations concerning performance-related justice as an effort-centred concept. The authors interpret the employees’ normative expectations, which are ‘explicitly or implicitly formulated in the context of particular organisational regulations (such as distributive rules, decision-making procedures, and crisis management measures)’ as demands for legitimacy (Reference Kratzer, Menz and TulliusKratzer et al., 2015: 14). Among skilled workers, the measure of performance-related justice is linked here to a notion of exertion in the workplace, and among ‘knowledge workers’ to a readiness to exert oneself (Kratzer et al., 2015: 50–51).

Alongside these qualitative studies, which illuminate the complexity of the various expectations placed on work, a quantitative study (Reference SchneiderSchneider, 2018) has shown on the basis of the German Socioeconomic Panel (SOEP) that the majority of traditional employees in Germany feel they are fairly paid. Around 61% of workers (64% including those covered by collective bargaining agreements) consider their gross earnings fair, though only 55% (56% including those covered by collective bargaining agreements) consider their net earnings fair. Among almost all groups, gross earnings were considered fairer than net earnings. Interestingly, only 59% of full-time employees consider their gross earnings fair, compared with 62% of part-time employees and 79% of casually employed workers (Schneider, 2018: 366).

Even among those performing ‘mini-jobs’ which like crowdworking jobs are often considered an additional source of income but are usually low paid, provide little to no social security protection and involve zero-hour contracts, 77% of employees consider a good income important or very important, while 82% feel that an income that is fair with respect to their performance is likewise important or very important (Reference BeckmannBeckmann, 2019: 251–342). In both of these respects, the 2016 found a poor match between employees’ expectations and reality: this match was only 47% for a good income and 48% for a fair performance-related income (Reference BeckmannBeckmann, 2019: 293).

Such a results-based orientation is by no means restricted to the crowdworking sector; in more traditional employment relationships, performance-related policies have increasingly shifted from an effort-oriented approach to a results-based approach in recent years (Reference BreisigBreisig, 2018). As Menz and Niess (2017) note, this shift in managerial approach from effort to results is not considered ‘fundamentally illegitimate’ by employees, yet neither does it constitute ‘an independent principle of justice (p. 133; see footnote 12). What is unique to crowdwork, however, is that in contrast to the types of work (including ‘mini-jobs’) considered by the studies noted above, it does not take place within (commercial) organisations. It is nonetheless within such organisations that the ‘question of the formation of this principle’ (Reference Menz, Nies, Behrmann, Eckert and GefkenMenz and Nies, 2018: 132–133) is addressed and that the demand for performance-related justice persists as an ‘enduring source of conflict and critique’ due to its repeated infringement (Reference Tullius and WolfTullius and Wolf, 2016: 496–497). Because this organisational context (and the attendant space for conflict and critique) is lacking in the case of crowdwork, we do not know what to expect regarding crowdworkers’ perceptions of injustice related to their work performance. In the following two sections, we pursue this question empirically. We first present the results of a qualitative study of perceptions of performance-related justice. We used the results of this study to generate items for a survey instrument used in a quantitative study, described in the subsequent section, of attitudes towards performance-related justice among crowdworkers.

Understanding justice and crowdwork: Methodological approach

Whether the results of prior studies on performance-related justice among traditional employees and commercial organisations can also be applied to crowdworkers is an open question. For exploring this question and gaining information on aspects of performance-related justice from the various perspectives of those involved, we undertook our own empirical investigation in two stages. The first stage, detailed in this section, was qualitative and explorative.

Qualitative findings: Understanding justice expectations

Extrapolating from the crowdworker interviews, we mapped four areas of crowdwork in which expectations were articulated regarding performance-related justice. One central area is planning security. Crowd workers can never be sure how many jobs they will receive and whether they will ultimately be paid for them, and this lack of security is a frequently raised issue, one that is only mitigated when crowdwork is undertaken as an additional source of income rather than as the main source of income. The following represent two typical responses on this point:

The issue of planning security is closely tied to the fairness of performance evaluations. On this point, the interviewees complain that there is a lack of transparency in the evaluation criteria, and that they do not have the possibility of influencing or raising objections to these or the evaluation process itself. Attempts to raise such objections can come up against quite straightforward barriers, as in the following example:

A further problematic area is the evaluation history, which can lead to the accumulation of injustices in the evaluation process:

Crowdworkers not only experience a sense of injustice in relation to their evaluations, however. They also experience it in the lack of clarity in their task descriptions. Gerda Grass, for instance, states that the tasks are often ‘a bit too unspecific’ or ‘wishy-washy’ (Design Platform, 16: 64). Since the task descriptions ultimately affect whether crowdworkers are paid or not, ambiguous or contradictory descriptions are not only frustrating for workers; they also affect their broader sense of fairness. The following represents a typical example:

Regardless of the platform in question, the interviewees all find the level of remuneration inadequate. Some compare their potential earnings with other forms of secondary employment, e.g. ‘There I’ve earned two, three euros so far. It’s not worth it. The time you spend on it, you’d be better off mowing lawns’ (‘Pawel Polanski’, Microtask Platform, 14: 22). Others compare the low potential earnings with the profits of the clients and platform operators, e.g. ‘The whole thing is a multi-million Euro business […] it’s a form of modern slavery’ (‘Fritz Freudig’, Innovation Platform, 15: 45). Many interviewees also complain that the pay for the work does not match the effort required, as in the following case: ‘You do a great deal for little pay, and maybe for no pay at all. So that’s a form of exploitation’ (‘Sigmund Schlecht’, Ideas Platform, 19: 41). Note that the crowdworkers interviewed did not feel the lack of social insurance benefits to be problematic, a finding that seems puzzling. However, the interviewees worked in German-speaking countries, all of which have low unemployment and high social welfare benefits independent of employment status. Thus, these crowdworkers are not dependent on their crowdwork gigs for securing health insurance or retirement benefits.

Quantitative findings: Measuring justice expectations

Before presenting the quantitative findings, it is important to understand the limits of our research methods. We collected data on a battery of 28 attitudinal items using a five-point Likert-type scale ((-2) = strongly disagree; (-1) = disagree; 0 = undecided; 1 = agree; 2 = strongly agree). The majority of interviewees either ‘agreed’ or ‘strongly agreed’ with the majority of the items, and this limited variation makes the identification of response patterns unreliable. Also, because questions were not randomised, the interpretation of any covariance may be hindered by sequential artefacts and the altered item formulations. Conclusive comparisons between questions, groups and conditions, however, can be excluded on various methodological reasons. Furthermore, a quantitative validation of the dimensions in the form of a confirmatory factor analysis, for example, is not possible with the available data. In keeping with the exploratory character of the study, we shall therefore now give a cursory presentation of the descriptive figures:

Table 1A in the appendix presents a breakdown of the results for the dimensions of planning security, performance assessment, task briefing, and remuneration level. Figure 1 compares workers’ expectations concerning crowdwork (fair_c_*; lighter colours/grey) and their expectations concerning regular employment (fair_w_*; darker colours/grey). Our qualitative results had already shown that workers experience the lack of planning security in the crowdwork sector as a specific source of injustice. This is why most interviewees consider crowdwork as a form of secondary employment only. Nevertheless, interviewees’ expectations of a steady flow of jobs and tenders (‘fair_*_1’) is only slightly lower for crowdwork than for conventional employment ( Figure 1 , first bars on top).

Figure 1. Dimensions of justice. Crowdwork (fair_c_*; lighter colours/grey) versus work in regular employment (fair_w_*; darker colours/grey).

Source: Authors

As a result of the structural anonymity and lack of transparency of crowdworking platforms, the absence of opportunities to influence or object to performance assessment and their criteria is a fundamental problem. This problem was raised in the qualitative interviews. The quantitative data on the items ‘fair_*_2’ to ‘fair_*_6’ underscore the relevance of this issue, although the possibility of influencing performance evaluations seems to be considered less important than other aspects of performance-related justice ( Figure 1 , following bars).

Like the qualitative data, the quantitative data shows that unclear or contradictory task briefings very often constitute a further source of perceived injustice. The items ‘fair_*_7’ and ‘fair_*_8’ are considered important, and potentially even more important than in regular employment (Figure 1). The quantitative results also confirm the qualitative results concerning the importance of remuneration for the items ‘fair_*_9’ and ‘fair_*_10’. Nevertheless, the interviewees consider fair pay just as important in crowdwork (_c_) as work in traditional employment (_w_) ( Figure 1 , last bars):

An index that summarises all 10 dimensions of perceived justice (Table 1) also reveals a strong similarity between expectations towards crowdwork (N = 188, mean = 1.04, SD = 0.673) and gainful employment (N = 180, mean = 1.06, SD = 0.550). Hardly any variation is observable when comparing index values by current employment status. Medians and means differ only slightly, the differences are more evident in the distribution. The highest expectations are found among respondents who are currently not in employment with regard to normal employment ( Figure 2 , blue boxplots). All differences are generally small, however, and it seems that one’s own work experience hardly makes a difference in the reported expectations. Moreover, gender or one’s monthly workload in crowdwork – operationalised as average or less than average versus above average using group median as the cut-off – make little difference (Table 1 and Figure 3):

Table 1. Index values for crowdwork by employment status, gender and crowdwork workload.

SD: standard deviation.

Source: Authors

Figure 2. Justice expectations. Boxplots for crowdwork (above) and regular employment (below) differentiated by employment status (see also Table 1).

Source: Authors

Figure 3. Indices of fairness expectations. Boxplots for crowdwork (above) and regular employment (below) by gender and crowdwork workload (see also Table 1).

Source: Authors

In sum, our quantitative survey provides further evidence affirming the results of the qualitative interviews. Crowdworkers’ justice expectations differ little between crowdwork and conventional employment. Note, however, that the survey explorative only, with no claim to be representative of crowdworkers on the platforms investigated or for crowdworkers in general. Because we do not know the socio-demographics of either population, analysis of how the study sample may differ is not possible.