Approaches to digital piracy research

Introductory remarks

In response to the February 2013 special issue of Convergence on ‘Online Piracy’, this article reviews some of the conventional research methodologies used to further our understanding of this phenomenon. With its focus on novel aspects of piracy and its ‘ethically neutral position’ (Jewitt and Yar, 2013: 4), the special issue included six articles that covered a variety of research areas. To briefly focus on just a few: Harvey (2013) uses ethnographic observation – an approach seldom adopted in digital piracy (DP) research – from online forums to explore the album leak; Parkes (2013) provides an illuminating case study of the changing practices of the UK organisation Industry Trust in addressing DP and in doing so makes a convincing case for a behavioural change approach to tackling piracy; while contributions from Newman (2013) and Skågeby (2013) position DP in an everyday context by uncovering video game piracy and guitar shred videos, respectively.

As well as adding substantively to the literature to date by focusing on such novel research areas, the articles featured in the special issue draw attention to the generally weak approaches that dominate the bulk of the research into DP. One of the reasons for this is the conventional focus, particularly amongst criminologists and economists, on the commercial issues raised by DP (e.g. does music piracy have a negative impact on the recorded music industry), which demand quantitative methodologies and are subject to limitations (explained in depth later). Breaking with this tradition, the special issue of Convergence captures the benefits of exploring the phenomenon of DP as a means to enhance our understanding of how technology has impacted on culture.

Regardless of the specific research questions posed when empirically researching DP, a number of particular difficulties are ever present. Just a few of these are mapped out below, followed by a discussion of conventional approaches to researching DP, which forms the main focus of this article. In the process, the limitations of such approaches are exposed and used to inform the remainder of the article, which discusses alternative approaches that can be drawn upon to better serve future research in this field. A brief overview of one of the most pervasive issues in DP research offers a good starting point to begin this discussion.

Is DP good or bad?

As with much criminal activity, hard figures on DP rates are elusive. This, along with inconsistent and often vague accounts of the negative economic impact of piracy, makes it difficult for rights holders to convincingly demonstrate a need to address piracy behaviours. As Brown et al. (2007) explain, the public response to crime is largely dependent on views concerning the seriousness and scale of a particular crime. In the case of music piracy, for instance, little attention is paid in the media to artists who may be struggling as a result of that piracy, ¹ and hence, individuals engaging in DP may be subject to confirmation bias – that is, they will tend to favour information consistent with their existing belief that piracy is a victimless crime. Indeed, research into the neutralisations and rationalisations of DP show that denial of injury is the most commonly utilised argument to justify piracy (see Brown, 2013a, for a review).

Academic researchers face a variety of difficulties in trying to adequately explore DP, including human error. In a working paper, Andersen and Frenz (2007) conclude, against conventional wisdom and against the findings of previous research (e.g. Rochelandet and Le Guel, 2005; Zentner, 2006), that peer-to-peer (P2P) file-sharing (where individuals with networked computers can exchange files with each other) encourages rather than discourages legal music purchasing. Controversially, however, when Barker (2012) conducts his analysis of the data set used by Andersen and Frenz, he reaches the opposite conclusion, maintaining that a 10% increase in P2P downloads reduces legitimate demand by 0.4%. Although Andersen and Frenz’s original research has since been largely discredited, as has any apparent empirical evidence to support the notion that piracy increases legal sales of recorded music, Hammond’s (2012) recent research does suggest that new releases (marginally) benefit from file-sharing. Using a combination of BitTorrent data (the most common mode of file-sharing today) and sales data (via Neilsen SoundScan), Hammond traced the performances of new releases over a period of 5 weeks, in 2010.

Curiously, Andersen and Frenz (2010), in what appears to be a peer-reviewed version of the aforementioned 2007 working paper, reach the conclusion that there is no association between P2P file-sharing and compact disc sales. This echoes the conclusion reached by Oberholzer-Gee and Strumpf (2007) who, in a summary of this pervasive research area (2010: 49), explain that ‘papers using actual file-sharing data, suggest that piracy and music sales are largely unrelated’. Common sense dictates that these researchers cannot all be correct. And to complicate matters, many research articles on this particular topic (including Barker, 2012 and Hammond, 2012 cited above) are working papers and have not been subject to peer review. Yet such research is freely accessible on the Web, and hence, it is the research that is most likely to be read by the general public.

Commenting on the scarcity of reliable data, Hargreaves (2011: 73) notes that ‘measurement of any area of unlawful activity presents statistical challenges’. Commissioned by UK Prime Minster David Cameron to investigate the potential need for copyright reform in the United Kingdom, ² Hargreaves outlines some of the particular complications in measuring DP, including (and with an emphasis on) weak research methods and confusion over legalities. Hargreaves identifies a need for evidence-based claims, whilst Karaganis (2011) criticises the attention-seeking focus on numbers in industry research. Put simply, confusion and conflicting opinions exist over what is actually known about DP.

Cutting across the (often overlapping) issues and research discussed above is an overarching assumption that all DP occur online. Yet, before widespread broadband Internet penetration, enterprising individuals would shoulder the risk and download digital media illegally to distribute to others offline, often for payment. With the risks of engaging in DP increasing as technological and legislative changes continue to intensify, DP in the future may in fact revert to these earlier practices.

Conventional DP research methods and their shortcomings

A substantial amount of research on DP uses various forms of (economic) modelling – that is, simplified theoretical constructs that quantitatively explain relationships between variables. Whilst relevant, emerging empirically based research demonstrates that this approach has limitations, since it cannot take into account the significant differences in individual behaviour that can be observed across different DP practices. Yet, of the 11 articles in the December 2010 special issue of Information Economics and Policy on ‘Digital Piracy’, for example, only 3 draw from empirical data.

Employing a modelling approach, Minniti and Vergari (2010) show that companies benefit from DP in emerging markets since piracy allows them to reach a broader customer base, whilst in established markets they argue this effect is absent. This leads the authors to conclude that companies in emerging economies are in theory unlikely to take a firm anti-piracy stance. However, the report Media Piracy in Emerging Economies (Karaganis (2011) does not support this claim, showing instead that anti-media piracy measures are prevalent globally.

Whilst there is a significant volume of research adopting this methodology, there is another even more common approach that arguably saturates the scientific literature: self-report methodology. The majority of research in the field of criminology (mostly in the United States), for example, uses self-report methodology to collect data from college students. Results using this approach must also be interpreted with caution. In their invaluable (but now dated) systematic review of literature on DP drawing from over 400 sources, Williams et al. (2010) also identify an over-reliance on the use of student samples, whilst Gunter et al. (2010) utilise a sample of even younger participants. Adults are virtually ignored in the literature. Criminological research also falls victim to focusing on DP only as a criminal activity, with Higgins et al. (2006) – using a sample of undergraduate students – defining DP as a ‘white-collar crime’ and Higgins et al. (2012: 413) proposing that ‘music piracy is not a social science issue, but a legal issue’. To a certain extent, given the central concerns of criminology, this is hardly surprising, but such statements nevertheless highlight the often one-sided approach to researching DP. This is why the 2013 special issue of Convergence with its eclectic approach is such a welcome addition to the existing literature.

In another recent article, Robertson et al. (2012) make assumptions about illegal downloading using a (crude) scale that asks for self-reported rates of downloading. Whilst the scale does not distinguish between illegal downloading and downloading generally, the authors nonetheless infer that illegal downloaders are less ethical in their behaviour. Exploring the weaknesses of self-report methodology, Brown et al. (2007) explain that most self-reported surveys of DP tend to focus on occasional and frequent offenders – not persistent criminals. Surely from a regulatory point of view, the latter justify the greatest attention.

Research using self-report methodology also suffers from a seldom-discussed problem: the word ‘piracy’. To start with, it is inherently subject to bias, and Agnete Haaland (2010), the Former President of the International Federation of Actors, argues ‘we should change the word piracy. To me, piracy is something adventurous, it makes you think about Johnny Depp’. But the word also introduces another problem. As Morris and Higgins (2009) reveal, piracy behaviours are neutralised and rationalised differently for music and for movies, whilst more recently Siponen et al. (2012) have demonstrated that software piracy is also neutralised and justified in very particular ways.

Put simply, this means individuals actively seek out different types of digital media for different reasons. This is an important insight. Jacobs et al. (2012: 959) explain that ‘in the discussion of downloading activity, music, movie and software downloading are often not differentiated, which can naturally lead to inaccurate conclusions’. In their book Digital Piracy: An Integrated Theoretical Approach, Higgins and Marcum (2011) explore the relevance of various theories across software, music and movie piracy independently of one another and thus undertake a consideration of ‘multiple piracies’.

Indeed, as far back as 2004, Marshall cautions against the one-size-fits-all approach in the process of championing the benefits of bootlegging (an unauthorised audio or video recording, often of live performances). An often overlooked aspect of music piracy, he explains that bootlegs do not substitute for existing recordings and argues that artists are therefore not in fact victims of piracy. Kernfeld (2011) argues that bootlegs afford young or emerging musicians the chance to have their material promoted early in their career, echoing a common sentiment voiced amongst pirates that downloading media illegally is beneficial rather than harmful (although as evidenced above, some literature suggests otherwise).

Given the existence of ‘multiple piracies’, the singular term piracy is manifestly inadequate to describe DP. The term ‘pirate’ is similarly problematic. For example, Holt and Copes’ (2010: 642) mixed-methods investigation into piracy subcultures (discussed in more depth later) reveals that ‘profiting from piracy was strongly and uniformly condemned’ (my emphasis). This suggests that someone wanting simply to avoid paying for digitised media content is not considered on a par with individuals who engage in DP in order to actively gain financially from it and that the latter are not considered to be part of the wider community of pirates.

As emerges through the work of Oberholzer-Gee and Strumpf (2010) and Tschmuck (2012), differing research methodologies make it almost impossible to achieve consensus across research findings and especially so in another pervasive area of DP research: morality. Chen et al. (2008), for instance, employed a widely accepted theoretical approach among psychologists to measure moral development and found no link between moral reasoning and piracy. However, Gupta et al. (2004) employed a different approach in their study – simply asking participants if they considered software piracy to be unethical – and found that consumers who are less concerned with ethics are more likely to pirate. Indeed, as Dejean (2009) observes, the variety of data collection methods makes comparisons across different studies difficult.

Irrespective of the approach, a substantial volume of research poses questions relating to behavioural frequency or speculates about hypothetical scenarios. However, the former requires an analysis of previous behaviour, whilst the latter demands a projection of anticipated future behaviour, and each has scope for bias. Although criminological research should be applauded for its frequent use of theory – perhaps best illustrated in Higgins and Marcum’s aforementioned 2011 book (which also discusses the limitations of self-report methodology) – Perlmutter (2010) raises compelling arguments that suggest criminological theory is in fact poorly equipped to explore copyright infringement.

Alternative approaches to researching piracy: The need for innovation

Although the proliferation of relatively weak research on DP could be attributed to the difficulties in measuring piracy objectively, qualitative research shows the potential for simply talking to individuals engaging in piracy. The few articles that have adopted a qualitative approach demonstrate the willingness of these individuals to discuss their behaviours openly – something that researchers have rarely seized upon. An interesting insight into book piracy motivation (with research in this area still in its infancy), for instance, comes from author Lloyd Shepherd’s (2012a) account of having his first novel, The English Monster (2012b), pirated and published online. He requested clarification about why his book had been pirated and met with frank responses explaining that pirated books create great publicity for the authors – this is of course analogous to the argument that pirated recorded music motivates live concert attendance. Thus, pirates do not appear to be shy about discussing their activities, despite the increasing risks involved.

The March 2011 edition of the Journal of Retailing and Consumer Services – a special issue on ‘The Future of Music Retailing’ – is particularly notable for its use of qualitative research methods (McIntyre, 2011; Nutall et al., 2011) and consideration of live music (Dilmperi et al., 2011), an area seldom considered in the exploration of music piracy. The work of McIntyre and Nutall et al. adds a level of depth to research on predictive factors by exploring identities, an expanding area of research in music psychology (see Hallam et al., 2011), which quantitative research would struggle to capture using conventional methodologies. As a cultural phenomenon, research into identity is revealing, with Nutall et al. uncovering the effect of fan loyalty on attitudes and downloading behaviour and antipathy towards the record industry.

Nutall et al. (2011), like Tade and Akinleye’s noteworthy 2012 study, are among a minority of researchers using focus groups. Drawing on focus group interviews conducted among Nigerians, Tade and Akinleye adopted the rare approach of exploring music piracy from the point of view of pirates, musicians and producers. Nutall et al. (2011), in their study of music consumption, identify six different ‘tribes’ of consumers: loyalist; experience seeker; preacher; revolutionist; techy; and conventionalist. Whilst the authors acknowledge overlaps between ‘tribes’, their findings serve to demonstrate the potential for a segmented approach to music marketing in a world where technology has created different types of music listener (e.g. Lamont and Webb, 2010). This demonstrates that qualitative research can serve to inform practical matters on how to understand and address DP just as much as quantitative research can. Without being hypothesis driven, the former can also reveal aspects of DP behaviours that might otherwise remain hidden.

Another important qualitative study is Holt and Copes’ (2010) aforementioned mixed-methods study that employs ethnographic observation of a discussion forum and interviews to investigate forms of subcultural exchange in an online setting. Their rich findings show how the risks associated with piracy are minimised by exchanging tips via the online discussion forum. This is now also easily observable on the microblogging service Twitter, where individuals engaging in DP openly request help to locate illegal content online. Crucially, Yu (2012) demonstrated inconsistent findings across qualitative and quantitative research, which suggests there is a gap in the research methodology, which could perhaps be filled by mixed-method approaches.

Whilst a more consistent use of theory would facilitate comparisons across quantitative and qualitative studies, this is difficult to achieve with researchers from such a broad variety of disciplines exploring the phenomenon. Crucially however, many have focussed on the economic aspects of piracy, largely ignoring the underlying cultural dimensions that may prove most revealing and more productive in helping facilitate policymaking. To this end, more ethnographic research would provide valuable additions to the literature, with the additional benefit of avoiding the limitations of self-report methodology.

Although fierce critics of conventional quantitative methodologies are likely to retain their stance, such methods are not entirely without merit. Mostafa (2011), for instance, employed a neurocomputational intelligence analysis of software piracy rates and found the rate to be increasing globally. While this approach is technically complex, in earlier work Mostafa and Nataraajan (2009) demonstrated that this approach outperformed traditional statistical approaches when they used it to analyse and predict the ecological footprint of nations. Such alternatives to conventional quantitative approaches may similarly have their limitations, but they do exist.

In their 2010 systematic review of DP literature, Williams et al. explain that research findings are often generalised globally and draw attention to the lack of cross-cultural research (with not a single study cited in their review). Whilst still scarce, cross-cultural research on DP does exist. Karaganis and Renkema (2013) demonstrate that although Germans are less likely than Americans to believe that their Web activity is monitored, they are more than twice as likely to protect their online privacy. This, along with Cinnirella and Green’s (2007) finding (using an experimental paradigm) that individuals from collectivist cultures are greater ‘cyber-conformists’, suggests that different norms of cultural behaviour will play a role in determining piracy behaviours. A shift away from an ethnocentric approach towards more cross-cultural research initiatives might therefore prove productive in both determining the degree to which differing cultural norms, beliefs and attitudes can contribute to the relative successes and failures of piracy deterrent measures and developing more effective anti-piracy strategies.

Williams et al. (2010) also noted a lack of longitudinal research, and even in the intervening years only a handful of studies have adopted this methodology. One notable example is Danaher et al.’s (2012) investigation into the effect of the ‘three strikes’ HADOPI law in France. This controversial law, aimed at encouraging compliance with copyright laws, required Internet service providers to terminate the Internet access of repeat piracy offenders but was revoked in 2013 for being too harsh. Drawing on weekly iTunes sales data from July 2008 to May 2011, Danaher et al. show there was a marked increase in both song and album sales during the period and argued this was a result of increased awareness of the law. Their study demonstrates the possibilities for a longitudinal approach and the potential for its use in other areas of DP research.

Williams et al. (2010) were, however, able to report more novel research methods, including content analysis (Lau, 2006) and experimental design (Tang and Farn, 2005), and one study that utilised analysis of actual piracy behaviour (Bhattacharjee, et al., 2006). Since Williams et al. undertook their review, more studies that draw on piracy behaviour have been published (although they are not without their shortcomings), but they still represent only a small proportion of the research carried out to date.

Concluding remarks

This article exposes a few of the ‘contradictions and complexities in this sensitive and controversial subject’ (Williams et al., 2010: 298), drawing attention in particular to methodological issues, the contentious use of the word piracy and the need for different forms of piracy to be considered independently of each other. Whilst the last of these three further complicates how best to conceptualise DP, its importance is demonstrated in the growing area of research that exposes how a musician’s ‘stage in the game’ largely determines the best means to distribute their music (see Brown, 2013b, for a review). There also needs to be greater awareness of the specific challenges of researching DP and the potential benefits of engaging in cross-discipline research.

With regard to the over-reliance on student samples in data collection, Popham (2011) does actually demonstrate the usability of undergraduate-based research as predictive of music piracy in the general public. Furthermore, Taylor (2012) shows that the intention to pirate is predictive of piracy behaviours. He devised an innovative method for measuring file-sharing activity, which was both unobtrusive and provided participants with anonymity. Taylor does however note, that this approach introduces new limitations. Nevertheless, these studies are valuable additions to the literature, and more research in this vein would help expose the shortcomings of much DP research as outlined above.

However, the issue of student samples in DP research raises a broader concern. Ingram and Hinduja’s (2008) reliance on student samples is criticised by Williams et al. (2010) because it is justified by citing the International Federation of Phonographic Industry’s Digital Music Report (IFPI, 2005). Williams et al. found no empirical support for the IPFI’s claim that young people are most likely to source music illegally, which suggests Ingram and Hinduja have simply relied on the assumed credibility of the organisation. A matter for concern in its own right, it is all the more so, given that Hinduja (2005) is one of the most published (and cited) DP researchers. It is also all the more surprising, given that Hinduja acknowledges many of the limitations of DP research in his published work, including in his 2005 book Music Piracy and Crime Theory. Williams et al. also criticise Sims et al. (1996) for not citing any literature to support their claim that universities are ‘hotbeds’ for piracy and, more recently, Altschuller and Benbunan-Fich (2009) for their reliance on two newspaper articles – both written 6 years prior to their own article – to inform their similar aged-based assumptions.

In his book How Music Works, Talking Heads’ frontman David Byrne (2012: 136) explains how:

A century of technological innovation and the digitization of music has inadvertently had the effect of emphasizing its social function. Not only do we still give friends copies of music that excites us, but increasingly we have come to value the social aspect of a live performance more than we used to.