A dissertation about RuneScape, Venezuela, and human psychology

What do you do after finishing a dissertation? Publish it!

Reproduced here raw - mistakes and all.

This isn't very readable. A later post will distil this 45 minute read into a 3 minute read!

The moral psychology of digital migration: from Venezuela to RuneScape

Contents

Acknowledgements

Abstract

Chapter 1: Background, Literature Review, Rationale

Background

Literature review

Moral circles and communication technology

 RuneScape and Venezuela’s economic crisis

The age of empathy

Moral foundations theory

Chapter 2: Methods

Overall study design

Method 1: Questionnaire

Participants

Materials

Procedure

Method 2: Content analysis

2.3.1. Participants

2.3.2. Materials

2.3.3. Procedure

2.3.4. Item codes

2.3.5. Hypothetical coding example

Chapter 3: Results

Questionnaire results

Data treatment

Reliability

Convergent validity between scales

Moral foundations

Content analysis results

Data treatment

Chi-square statistics

Chapter 4: Discussion

Null results

The problem with context based moral foundations. 34

Main findings

Implications

Limitations

Future research

References

Appendix

Acknowledgements

There are many people who contributed to this journey that I’d like to thank - I’ll try and keep it short. Whenever someone says this, what follows is always long.

First, I would like to thank my family, Ann, Declan, and Rosie Hughes. My therapist Maggie is next. Indeed, without these four people, I doubt that I’d ever have even attended university, let alone execute a research project like this.

Next, I want to thank my supervisor, Paul Richardson, who was both supportive and flexible on the journey, who gave me the space to create my own independent project, an attitude that not only aligned to my personality, but also that incentivised me to make this work my own, which I really enjoyed doing!

To some good friends - Carl and Tom - who, both through good conversation, listening, and questions, helped to open my world up intellectually. To Hazel Gordon, for listening to, being critical of, and helping to develop the ideas presented in this work. And to Helen Chambers, whose speedy proofreading was immensely helpful.

Thanks to the weird and wonderful community of RuneScape, who not only provided me with many great memories, but also directly inspired this work.

Finally, to the many academics and authors who put complex ideas into understandable language. When I finish reading their books, I only realise how interesting and complicated the world is, how much there is to know, how much I don’t, and how many good questions there to ask! These authors are scattered throughout this work.

Abstract

Introduction. Venezuelan people living in extreme poverty earn an income from the online game RuneScape, known as gold farming. Normal players of RuneScape learn about, interact with, and discover the Venezuelans real-life circumstances either in game or via YouTube videos. Exploiting the opportunity, this study tests associations from both the empathy-altruism hypothesis and moral foundations theory.

Methods.A mixed method design is applied, using two quantitative methods. First, members of the RuneScape community and non-RuneScape community members are surveyed with a questionnaire. Second, a content analysis is performed on four YouTube videos which interview Venezuelan players.

Results. Being a member of the RuneScape community is moderately associated with knowledge of gold farming (r (90) = .51, p < .001). In relation to four YouTube videos, a large majority (n = 253; 72.1%) of commenters expressed sentiment broadly supportive of Venezuelan gold farming. Independent of each other, codes were unlikely to have arisen by chance variation, Care χ2 (1, N= 109) = 166, p < .001, Fairness χ2 (3, N = 50) = 329, p< .001, and Loyalty χ2 (1, N = 73) = 259, p < .001.

Discussion. Associative results are interpreted in the context of empathy-mechanics and ideas about what makes moral circles expand. Digital communication technology, especially when actors are members of the same team or community (RuneScape), may make the abstract process of learning about the suffering of people in foreign lands more concrete. Under certain conditions, this may help to incentivise prosocial outcomes.

Chapter 1: Background, Literature Review, Rationale

1.1 Background

Moral circles determine who and what is worthy of our interest, care, and attention. Indeed, understanding what makes people expand or constrict their moral circle is an important area of research in moral psychology (Crimston, Bain, Hornsey & Bastian, 2016; Graham, Waytz, Meindl, Iyver & Young, 2017). Should people care about the environment? About racial minorities? What about people suffering in foreign lands? This study addresses the latter question. Our moral circles expand, say philosophers, because of reason (Singer, 1981). Adam Smith (2002/1759) argued that an important distinction exists between concrete and abstract experience, suggesting that a person tends to feel more psychological distress from concrete experience, such as hurting one’s little finger, compared to abstract experience, such as learning about the deaths of others in foreign lands. But despite this, Smith maintained, reason instructs that – if given a choice between minor pain to oneself and the death of others faraway – the pain to oneself would be incurred. Although it is not possible to test Smith’s hypothesis directly, it is possible to test which specific mechanisms make learning about the suffering of others in foreign lands less abstract and more concrete.

Imagine that a Venezuelan person lives in extreme poverty. Suppose they begin playing an online game, RuneScape. By playing this game, not only do they earn a real income that helps to alleviate their poverty somewhat, but they also discover that this income is a better alternative compared to economic options in their local area. Next, consider how a normal person playing RuneScape might react upon learning that this other person lives in extreme poverty. In fact, this is not a hypothetical. It describes a real scenario that continues today where Venezuelan people are “migrating digitally” to online games like RuneScape. This phenomenon is fertile ground for testing psychological hypotheses about empathy and morality.

1.2. Literature review

1.2.1. Moral circles and communication technology

Without tools, communication between people is severely limited. The world map for ancient Greeks and Chinese was largely dark (Wells, 1932). Missing appropriate tools, both societies were unaware of many other people and places across planet Earth. Without awareness, how can anyone know that help is needed, much less expand their moral circle. Communication technologies are tools that permit moral circle expansion. Indeed, argues Pinker (2011), not only are expanding moral circles associated with increasingly sophisticated communication technologies, but they also make the abstract experience more concrete

Consider ancient innovation: the dawn of writing 4,500 years ago (Diamond, 1999). Older still is the suggested cause of language itself, and thus of all subsequent communication, the Cognitive Revolution, approximately 70,000 years ago (Harai, 2014). Modern technological innovations include the printing press, international transportation, and television media. Many examples exist that demonstrate how communication technology (language, writing, travel) helps to expand moral circles. Eighteenth-century authors distributed letters arguing against the commonplace practice of public executions (Edelstein, Findlen, Ceserani, Winterer & Coleman, 2017). Media reports about a famine in British India caused an international fundraiser (Brewis, 2010). A novel exposes the illogic of slavery (Twain, 1885), another tells of the waste of youth in war (Remarque,1929). In contemporary culture, communication booms. Countries in which women are more influential have less domestic violence (Archer, 2006), people who know gay people are less homophobic (Gallup, 2021), American counties along coasts and waterways are more liberal than rural conservative communities (Haidt & Graham, 2007), and effective altruists voluntarily organise to alleviate the destitution of complete strangers (Singer, 2019). Today, communication technology generates not only more data per day than what amassed in the entire written history of human civilisation, but that 18.1 million texts, 188 million emails, and 4.5 million videos are sent and viewed every 60 seconds (Marr, 2018; Kwik, 2020). Combining both patterns (expanding moral circles and communication technology) is what this study is interested in. Specifically, the interaction between moral circles and digital communication technology.

1.2.2. RuneScape and Venezuela’s economic crisis

The internet - a series of “inter-networked” computers able to communicate with each other and distributed around the world - permits digital communication (Ryan, 2013). Both the online game RuneScape and the video sharing platform YouTube are built on this infrastructure; both permit direct, peer-to-peer interaction in real time; both are digital communication technologies. This “interaction mechanism” allows an individual sitting at a computer who may be living comfortably, to meet, interact with, and to learn about the suffering of another person in a foreign land. Moreover, digital communications mirror real world interactions whenever human actors are involved, reflecting both human nature and culture (Chalmers, 2022). Moreover, in digital contexts, there is evidence suggesting that people respond similarly to real-world disasters as they would in the ‘real-world’ (Miller, 2015, 2019; Pearce, 2017). In other words, the technology may make what would otherwise be abstract more concrete.

The online game RuneScape is entangled with a real-world economic crisis. Approximately 96% of Venezuelans live on less than $1.90 a day, a cut-off level that defines extreme poverty (Thomas, 2019; UN, 2015). It is, perhaps, the worst economic crisis ever experienced during peacetime (Gallegos, 2016). Consequently, tens of millions of people across Venezuela have experienced hyper-inflation (meaning worthless money), food shortages, and millions of refugees. Because work cannot generate a sufficient income, Venezuelans sought (and continue seeking) online work: in the online game, RuneScape.

Players roam the “point-and-click” pixelated world, complete quests, trade, kill monsters, and in doing so, acquire in-game “gold”. Gold can be sold for real-world money, although this is against game rules (Jagex, 2022). For most normal players, it is not worth the opportunity cost – that is, the risk of receiving a permanent ban for their account. However, for Venezuelans whose fiat currency has collapsed, playing RuneScape for real-world money has become a strategy for generating an income and feeding families. The opportunity cost is worthwhile (Powell, 2019). Consider the numbers. A monthly salary in Venezuela in 2019 was between $5-15 USD, whereas an income on RuneScape yielded $50-400 USD a month (Weiss, 2021; Aronczyk & Beras, 2021). Playing a game for cash has a term: gold farming.

Gold farming disrupted prices in the RuneScape economy, provoking a moral response (Ombler, 2020). The moral response is either: farmers are an out-group of rule-breakers, or an out-group trying to get by. This is the language of immigration debate, unfairly taking jobs, or fairly seeking a better life (Sowell, 2011). In-group-outgroup distinctions are robust psychological phenomena that vary depending on how the group is perceived, and which probably feature both within democratic debate and inside computer-generated worlds (Fiske, 2015; Everett, Faber & Crockett, 2015). Previous research with RuneScape supports this notion (Crowe & Bradford, 2006; Bilir, 2009; Walker, 2015; Robe, 2018; Smith, Rauwolf, Intrillagator, Rogers, 2020). Indeed, the literature suggests that the distinction between digital and physical worlds is porous - often not clear cut. Another theme is that human nature and culture runs through community life no matter the context, including RuneScape (Feenberg & Bakardjieva, 2004).

In summary, the average RuneScape player may be, at times, interacting with gold farmers who are probably playing to feed themselves and their family. YouTube videos capture this dynamic. Farmers are interviewed who confirm the opportunity cost idea: they play RuneScape for work. In short, these videos communicate. In doing so, they may alter how one group perceives the other. One group of people (gamers) can learn about another group in poverty (Venezuelans). Moreover, both groups are members of the same team (RuneScape). Before discussing explanatory theory, a curious empirical fact is of note. Two videos by RuneScape content creators (who curate RuneScape videos for RuneScape gamers) raised money for Venezuelan charities. More than $15,000 USD was raised. This empirical fact – and the related psychological factors at play - needs explaining. First, empathy. Then, morality.

1.2.3. The age of empathy

In psychological research empathy is a popular topic. More empathy makes the world a better place, resolves conflicts, and aids relationships (de Waal, 2010; Baren-Cohen, 2012). Two psychologists (Pinker, 2011; Bloom, 2016), however, are sceptical. Empathy research, they suggest, is often poorly defined, suffers measurement problems, both conformation and publication bias, and typically uses small sample sizes, increasing volatile results. Empathy has many different definitions. For instance, Batson (2011), a key researcher in the area, identified seven different definitions for empathy, encouraging definitions to be clearly operationalised, and for researchers to be consistent. In this study, empathy is defined as “experiencing other-oriented emotions elicited by and congruent with the perceived welfare of someone in need”, which is the working definition that Batson (2011, p.11) also uses. Empathy, then, is about feeling the feelings of others. Critical analysis follows.

The relationship between empathy and prosocial action is not straightforward. As Pinker (2011, p. 575) describes:

“For many years a charity called Save the Children ran magazine ads with a heartbreaking photograph of a destitute child and the caption ‘You can save Juan Ramos for five cents a day. Or you can turn the page.’ Most people turn the page.”

Why? Mirror neurons exist in the brain, suggesting the hardwired home of empathy (Pellegrino, Fadiga, Fogassi, Gallese, & Rizzolatti, 1992). And although real neurological overlap exists between when people empathise with others and when they themselves experience pain, there is also much neurological difference (Bloom, 2016). Seeing someone else suffer is not simply enough for moral action. Instead, empathic concern, asserts Bloom (2016), is determined by moral judgement. Put differently, who you empathise with, and to what extent, is not driven by how much of a “good person” you are, or how much empathy you possess, but by how you first morally judge others. For instance, using functional magnetic resonance imaging (fMRI) to detect how empathy responses differ, Decety, Echols & Correll (2010) had subjects watch videos of people who had contracted AIDS but were told different stories: the victims with AIDS contracted it via injecting drugs, or via blood transfusion. In general, subjects reported – and fMRIs confirmed - feeling less empathy when the victims contracted the disease via drug-taking behaviour. More fMRI scans in a different study reveal that empathic responses of football fans differ depending on who is in pain (Hein, Silani, Preuschoff, Batson & Singer, 2010). In this manipulation, fans experienced more empathy towards watching fans of the same team in pain compared to fans from alternative teams.

Another study indirectly tested empathy when subjects viewed pictures of homeless people and drug addicts (Harris & Fiske, 2006). In this study, instead of empathising with these people’s suffering, the brain regions associated with disgust, not empathy, were most active. Finally, another experiment examines empathy another way. If subjects are guided towards empathising strongly with somebody, such as a young girl in need of medical treatment, they are prone towards unfair preferential treatment (Batson, Tricia, Klein, Highberger & Shaw, 1995). In this experiment, participants tend to move this girl ‘up the medical queue’, even at the expense of others who were in greater need. Together, these findings suggest why people may turn the page; they do so because empathy is a spotlight that favours some and not others, dependent on prior beliefs and preferences.

Empathy may, despite the preceding discussion, have positive applications. Three decades of experimental research culminated in the empathy-altruism hypothesis, which posits that specific conditions exist that maximise the probability for prosocial outcomes (Krebs, 1975; Batson, 1981; Batson & Moran, 1999; Batson & Ahmad, 2001; Batson, 2011). Three conditions have been identified. First, perspective taking. Second, valuing the other. Third, perceiving need in the other. In essence, these findings are the ‘feel-good’ inverse of empathic costs. The empathy-altruism hypothesis is not blind to the costs of empathy, however; it predicts that empathy can help to create prosocial outcomes only under these specific conditions. At least one condition needs to be active, though two or three may interact. Concluding this section, to a large degree, this study accepts the sceptics view of empathy, which argues that empathy is a weak moral guide, a narrow spotlight prone to biases. However, experimental research also suggests that prosocial outcomes can become more likely under certain conditions, which this study seeks to exploit. And where there is empathy, there is also morality.

1.2.4. Moral foundations theory

An additional framework to analyse the RuneScape-Venezuela situation is moral foundations theory (MFT; Haidt, & Joseph, 2004; Graham, Nosek, Haidt, Iyer, Koleva & Ditto, 2011; Haidt, 2013). Even a critic of MFT credits it with changing the way many researchers look at, think about, and study morality (Curry, 2019). Morality is explained naturalistically. It is something that human beings come into the world equipped to do, a series of evolutionary and cultural adaptations to the problems of social life. Morality is innate, defined as organised in advance of experience (Marcus, 2004). But experience then revises this, which is why cultures proliferate around the world, use different moral codes, and morals change (Shweder, 1991; Bloom, 2010). If experience could not revise what is innate, for instance, Western Europe would still host public executions. Conversely, something is common across cultures: human nature (Brown, 1991; Pinker, 2002). Hence, MFT bridges this gap. The theory posits five foundations that claim to model universal moral concerns. These foundations are care/harm, fairness/cheating, loyalty/betrayal, authority/subversion, and sanctity/degradation.

Different people rely on different foundations, which then predict what issues people care about (Graham, Haidt & Nosek, 2009; Kivikangas, Fernández-Castilla, Järvelä Ravaja & Lönnqvist, 2021). Political liberals are, for example, more concerned with the care and fairness foundations as compared to political conservatives, who tend to value all foundations (roughly) equally. Empirical research indicates two foundations then emerge, “individualising” and “binding”. Moreover, there is evidence - ecological and experimental - that stronger individualising foundations are associated with volunteering and donating to out-group (international) charities, while binding foundations predict the allocation of resources towards helping in-group (domestic) causes (Nilsson, Erlandsson & Västfjäll, 2016; Nilsson, Erlandsson & Västfkäll, 2020). Where possible, this study seeks to replicate these associations in relation to the RuneScape-Venezuela situation.

MFT, however, has flaws. One team of researchers (Curry, Chesters & Lissa, 2019) agree with the overall evolutionary-cultural approach, but instead build their theory, morality-as-cooperation (MAC), in line with game theory, which is crucial for hypothesis testing in evolutionary biology (Dawkins, 2016). MAC has two main improvements on MFT. First, MAC makes specific, mathematical predictions. Second, MAC differentiates between established evolutionary concepts (e.g., kin, reciprocal and competitive altruism, and respect for prior possession). Nevertheless, the theory is new; it remains an empirical question whether it can improve upon MFT. Hence, MFT was employed.

1.2.5. Rationale

Communication technology is associated with a widening moral circle, reducing the physical limitations inherent to communication, and making it more concrete. A rationale follows: as digital communication pervades daily life exponentially, it is important to gain a balanced perspective of the potential costs and benefits inherent to it. Getting more specific, perhaps the RuneScape-Venezuela situation fuels only more division between people. But the preliminary evidence from YouTube videos cited here suggest that hard money, $15,000 USD, have already been donated to appropriate charities, pointing away from division and toward an expanding moral circle. The empathy-altruism hypothesis stipulates three conditions for which prosocial outcomes occur, which are all present within the RuneScape-Venezuela scenario. Moreover, moral foundations theory can test how the moral response of RuneScape and non-RuneScape players differ. To the author’s knowledge, nor empathy, nor morality has been studied in the context of RuneScape. Thus, this study provides a novel contribution to the empirical literature.

1.2.6. Hypotheses

The following hypotheses correspond to the two methods the study uses.

H1 = questionnaire.

H1a = content analysis of YouTube comments.

· H1, RuneScape players will have a stronger moral response to gold farming than non-RuneScape players

· H2, Data will replicate the individualizing and binding pattern

· H3, RuneScape players will have donated more money to Venezuelan charities than non-RuneScape players

· H4, RuneScape players will be more aware about gold farming than non-RuneScape players

· H1a, Comments will include more reference to individualising than binding foundations

· H2a, Comments will include more reference to Care than other foundations

· H3a, Comments will include Care, Fairness, and Loyalty more than Authority and Sanctity foundations

· H4a. Because the videos create the empathy-altruism hypothesis conditions, comments will reflect this

Chaper two: Methods

2.1. Overall study design

Overall, the study design is between-subjects and quasi-experimental. In the questionnaire section, the design is between-subjects because two separate groups are tested on the same independent variables. In the content analysis, the design is quasi-experimental because statistical analyses are performed on pre-existing textual data. Therefore, this study employs convergent and holistic triangulation (Turner, Cardinal & Burton, 2017). This permits different methods to analyse the same phenomenon. Moreover, Yin (2006) argues that mixed methods research can use two quantitative methods rather than ‘quantitative/qualitative’ only. Hence, two quantitative methods are employed. Perhaps the fundamental consideration of mixed methods research is that different methods get used because they aim to shed light on the same issue rather than multiple issues (otherwise, two separate studies are more appropriate). Instead, mixed methods ought to converge to capture the same central issue. Although the methods suggested here will measure different sub-areas of digital communication (RuneScape, YouTube), the aim is to converge these findings to have a broader explanation of the same phenomenon – in this case the empathic and moral response that gamers have toward a Venezuelan out-group. Finally, additional materials, data, and tables reside in the appendices.

2.2 Ethics

Before data collection, ethical approval was granted by the ethics committee at Sheffield Hallam University. Data collection was contingent on this ethical agreement. Originally, a third method was proposed (a social network analysis of Tweets), but in data collection this method was dropped due to word-count and time constraints. For the questionnaire, participants were asked to read an information sheet, consent form, and then agree to having understood the terms and conditions of the study. In the study itself, no sensitive or harmful material was presented to participants. After the study, participants were debriefed about the nature of the study. None of their data is reproduced here, is shared only with the project supervisor, is securely stored, anonymous, and will be deleted after the dissertation’s completion. The ethical considerations of a content analysis of YouTube comments presents unique challenges, because, while the videos and their comments are in the public domain, commenters themselves are unaware that their data may be used for academic study purposes (Reilly, 2014). Moreover, it is not possible to individually contact > 500 individual commenters, nor the video makers themselves (video makers are essentially popular public figures with busy email inboxes, making it difficult to communicate with them). Thus, to collect data, anonymity is key. When collecting commenters data, all information associated with the comment (except the comment itself) is deleted prior to analysis, and no comments are reproduced here. Similarly, data is shared only with the project supervisor, stored securely, anonymous, and will be deleted after the completion of the dissertation. Because ethical considerations have been discussed here, the methods section does not refer again to any ethical step (content forms, anonymity etc), though this was the procedure adhered to throughout.

2.2. Method 1: Questionnaire

2.2.1. Participants

Before data cleaning, the sample consisted of 152 individuals. After data cleaning (discussed in the results section) the sample consisted of 90 individuals (69 male, 23 female, 5 non-binary, 1 preferred not to report). Participants were sampled opportunistically via Facebook, Twitter, and email, with ages ranging from 18 to 71 (M = 27.8, SD = 10.6).  It was first proposed that data could be collected from three distinct groups: non-RuneScape players, Old-school RuneScape, and RuneScape 3 players. This was because gold farming happens on Old-school, but not on RuneScape 3. However, data collection only sampled 2 players from RuneScape 3. To navigate this, a new variable was created which dichotomised these three into two groups: member of the RuneScape community, which includes people who no longer play but do take an active interest in RuneScape affairs, players from both games, and non-RuneScape players. If more RuneScape 3 participants had been recruited, they would have been analysed separately. However, this was not the case, and homogenising the data somewhat (by including two extra participants) seemed more reasonable than data deletion. Thus, regarding RuneScape, 41 participants (46%) had never played the game, while 49 (54%) considered themselves members of the RuneScape community. 66 participants (73%) were aware about the current crisis in Venezuela, though only 45 (50%) were aware that Venezuelans played for an income, and 34 (37%) were aware of associated charities. In appendix 1 are Table 1 and Table 2, which display additional descriptive statistics.

2.2.2. Materials

The short version of the Moral Foundations Questionnaire (MFQ) was administered to all participants, which has 20-items in two sections: Relevance and Judgements. This has reduced psychometric properties compared to the MFQ-30, although it is still psychometrically sufficient (Graham et al., 2011). The shorter version was used to reduce attrition. The Relevance section measures the relevance individuals associated with each of the foundations and is rated along a 6-point Likert scale ranging from 0 (Not at all relevant) to 5 (Extremely relevant). The Judgement section is said to be contextualised items and again is rated along a 6-point Likert scale ranging from 0 (Strongly disagree) to 5 (Strongly agree). This study also included a second MFQ-20, but that was altered so that the questions were context relevant to Venezuelan gold farming, as recommended by Smith Alford, Hibbing, Martin & Hatemi (2017). By testing the genetic heritability of moral foundations, they provided evidence, and explicitly suggested, that future research with moral foundations should focus more on specific moral issues because heritability estimates were low (thus, specific contexts will matter much more). This context-MFQ attempted to be as similar as possible to the original MFQ, while directing the items towards the moral issue of gold farming. For example, the original MFQ asks participants to rate the item: “Compassion for those who are suffering is the most crucial virtue”. In this case, the context-MFQ asked: “Compassion for Venezuelan’s who are suffering is the most crucial virtue”. This method was replicated for 19 other questions relating to other foundations.

2.2.3. Procedure

As recommended by MFT researchers, copies of the MFQ were downloaded from the moral foundations website www.yourmorals.org. This MFQ and the context-MFQ were then built into the Qualtrics survey platform, and participants were asked to take part. Upon completing the survey, participants were finally asked to agree to submit their response in line with the conditions of the study. After completion, participants were greeted with a debrief form which included more information about the study. To a large degree, this study attempted to replicate the MFT methodology, the only exception being the inclusion of an additional measure, the context-MFQ.

2.3. Method 2: Content analysis

2.3.1. Participants

Four publicly available YouTube videos and their comments (the participants) were selected. In total, these videos had been viewed 803,959 times with 2,046 comments associated with them. However, there is not a 1:1 relationship between when an individual watches a YouTube video and how the YouTube algorithms chooses what counts as a view (Funk, 2020). Based on an analysis of how YouTube algorithms for views function, it is estimated that between 160,791 and 803,959 individual users have watched these videos. Associated with these views were 509 comments, which were randomly selected for analysis.

2.3.2. Materials

This study followed advice from Miller (2015; 2018) who has outlined guidelines on how to conduct a content analysis. For example, he recommends that researchers use a 95% confidence level with a confidence interval of 4, which was calculated at https://www.surveysystem.com/sscalc.htm. Because there was a total of 2,046 comments to choose from, applying this algorithm meant that 509 were selected for analysis. To randomly select these 509 comments, every 4th comment was systematically selected across all videos, which is aligned to Miller’s recommendations.

2.3.3. Procedure

Comments were manually scraped from the relevant YouTube videos according to the correct sample size, pasted into a Word document, and were printed out ready for analysis with pen and paper. Next, comments were coded according to specific values (see item codes section). Because the present author worked alone, two measures of reliability were not included which ought to have been, according to the literature (Miller, 2018). These missed analyses were interrater reliability and Cohen’s kappa, which are both statistical practices that aim to mitigate subjectivity in research. Nevertheless, once coded, two levels were introduced for the five moral foundations to indicate which direction a comment expressed a particular foundation, either low or high. This measurement reflects how the MFQ is scored. For example, lower scores on the fairness foundation suggests that the participant thinks that something is unfair, whereas higher scores indicate that something is perceived as fair. This follows for all other foundations.

2.3.4. Item codes

In total, 14 codes were created.

Care/harm low. Coded if commenter expressed sentiment expressing harm or indifference towards gold farmers.

Care/harm high. If warmth or concern was expressed towards gold farmers.

Fairness/cheating low. If commenter expressed their belief that gold farming was wrong, unfair, or cheating.

Fairness/cheating high. If commenter expressed their view that gold farming was permissible and fair.

Loyalty/betrayal low. If commenter expressed negative sentiment about farmers as an unwelcome out-group or a feeling of betrayal towards the video maker

Loyalty/betrayal high. If commenter expressed positive relating to farmers as a welcome in-group or loyalty to the video maker

Authority/subversion low. If commenter believed that there were exceptions to the game rules for farming gold.

Authority/subversion high. If game rules were more important than farming gold.

Sanctity/degradation low. If commenter dismissed concern about sanctity or disgust as irrelevant.

Sanctity/degradation high. If commenter expressed concern about sanctity or disgust issues

Empathy-altruism. If commenter exhibited Batson’s (2011) criteria for the empathy-altruism hypothesis: perspective taking, value the other, or perception of another in need.

Sceptical. If commenter asked questions, was doubtful, or demanded more evidence.

Political. If commenter expressed political opinions.

Venezuelan. If commenter self-identified as being a Venezuelan native.

2.3.5. Hypothetical coding example

Previous research (Miller, 2019) includes a hypothetical coding example because it helps to communicate to readers how the present study was conducted and how codes were formed. See example below from (fictitious) YouTube commenter Lumbridgebeast912:

“I don’t believe this video is factually accurate. I’d like to see a more in-depth interviews with the gold farmers. Even if it were true, I don’t think cheating should ever be allowed because it sends the wrong message. Typical socialist agendas. People should always respect the rules of RuneScape, and cheating should be punished.”

This comment includes considerations about moral foundations, fairness (“cheating”) and authority (“respect the rules”). The sentiment is strong and would be coded as low for fairness, which reflects the sentiment of unfairness, and high for authority, because the commenter feels that authority is something to be respected. It also includes scepticism (“I don’t believe”) and a political element (“socialist”).

Chapter 3: Results

3.1. Questionnaire results

3.1.1. Data treatment

Prior to analyses, data were treated. First, it was appropriate to reverse scores for two items in the MFQ so that scoring pointed in the same direction. Next, missing data were identified. The present author, who was unfamiliar with the functionality of Qualtrics software and how questionnaire link distributions work, unintentionally inflated the participants in this study by using the “questionnaire link” himself to check for spelling mistakes and to check the format. Each time this was done, a “participant” took part, which was false. Moreover, real attrition also took place. Nevertheless, the author followed conventions in the literature about how to proceed (Little & Lang, 2018). Where there were swathes of items missing (demographic information was given, but none of the questionnaire items were answered), mean imputation was not possible. The author recognises that deleting participants is often the worst option about how to treat data, however it was a necessary last resort here because deletion mitigated against the combined effects of inflation and attrition. In total, the data from 53 participants were deleted. After deletion, 99 participants remained. Finally, based on the MFQ criteria, which contains two “catch items” to ensure that the responses are genuine, a further 9 participants were deleted, leaving a total of 90 participants.

Of these remaining participants, a pattern of attrition remained. Specifically, this pattern persisted in the final section of the questionnaire, which contained four items pertaining to “donations to Venezuelans” – either via charity, RuneScape ‘in-game’ currency, or Bitcoin. There seems to have been a misunderstanding among participants around these questions. For example, all participants answered 54 previous items consistently, then 7 participants missed item 55, 16 participants missed item 56, and 3 participants missed item 57. Indeed, the simplest explanation seems to be that participants were confused about the question or got bored towards the end of the survey. The pattern in scores for these questions can help solve this, however.
All participants who answered items 56 and 57 answered in the same way: no donations had been made with RuneScape currency. Similarly, for item 58, only one donation had been made with Bitcoin, while the rest had not made donations with Bitcoin. Moreover, this one donation was an outlier. The participant claimed to have donated “$500 USD or less”, were “aware” of the crisis, but “not aware” about the gold farmers. While outliers are an important source of information in scientific research (Taleb, 2007), and there is reason to believe this result might be genuine, given the dearth of other participants who failed to donate Bitcoin, this outlying result seems unlikely to be real. Hence it was removed. Next, mean imputations were given to the missing items in 56 and 57, which effectively continued the pattern of negative responses” to these questions.

Finally, assumptions were checked. The author could not locate the discussion about MFT distributions in the literature. However, as is visible in table 3 in Appendix 1, nearly all the distributions of foundations departed significantly from normality. Of 14 variables, only two, Context-authority (W = .98, p< .1) and Binding (W = .97, p < .06) were normally distributed. The literature on assumption checks also suggests that > ± 1 may be problematic on kurtosis and skewness levels (Orcan, 2020). For the most part, the data did fit these assumptions. However, because most of the variables violate the assumption of normality, non-parametric statistics were chosen.

3.1.2. Reliability

Cronbach’s alpha for the moral foundations were .50 (care), .56 (fairness), .75 (loyalty), .78 (authority), and .83 (sanctity). For the context-MFQ, the alphas .72 (care), .64 (fairness), .81 (loyalty), .80 (authority), and .83 (sanctity). The alpha levels for both care and fairness are well below the .70 cut off that the literature considers reasonable and was likely a function of using a reduced scale (Agbo, 2014). Indeed, other researchers reported similar methodological concerns regarding low alphas in the MFQ, both for the 20-item and even 30-item scales (Curry et al., 2019). However, according to one research team, this issue can be resolved somewhat by averaging alphas into their respective individualising or binding foundations (Doğruyol, Alper & Yilmaz, 2019). However, this resulted in .53 (individualising) and .79 (binding), leaving the former well below the convention in the normal MFQ. The context-MFQ performed slightly better here, .68 individualising and .81 binding.

3.1.3. Convergent validity between scales

The correlation between the total MFT scale and the total context specific MFT scale was both significant and strong (r (90) = .63, p <.001), suggesting broad agreement between the participants when answering items about morality in general and morality regarding gold farming on RuneScape. However, this did not hold true for the fairness foundation, which, when correlated together, indicated no correlation whatsoever, and a nonsignificant result (r (90) = .08, p < .44).

3.1.4. Moral foundations

The presentation of results is organised chronically according to the originally proposed hypotheses.

H1. RuneScape players will have a stronger moral response to the Venezuelan situation than non-RuneScape players. The first hypothesis was not supported; table 4 in Appendix 1 presents the moral foundations of the two groups across the two scales. A further independent Mann-Whitney U test revealed that the likelihood of achieving this pattern of results, assuming the null hypothesis to be true, was within the realms of chance variation, and had a small effect size. For the MFT scale (MWU = 8491.32, p = < .21, d = .16) and the context-MFQ (MWU = .874, p < .29, d= .13). Despite this “non-result”, it does mean that the two samples are well matched in their moral foundations, which helps to draw inferences from successive comparisons.

H2. Data will replicate the “individualizing” and “binding” foundation pattern. The second hypothesis was supported; the pattern of results indeed replicated the finding in the MFT literature that people tend to rely on different moral foundations (Kivikangas et al., 2021). Namely, the individualising and binding foundations displayed in table 4 in Appendix 1.

H3. RuneScape players will have donated more money to charitable ends than non-RuneScape players This hypothesis was not supported. In fact, of the 90 participants sampled, only 8 people (9%) claimed to have donated any money to charity. Of those 9% who did donate, half were from the RuneScape community, while half were not. The correlation between being in the RuneScape community and donating to charity was weak and non-significant, suggesting that being in the RuneScape community is not associated with more money donated to Venezuelan charities (r (8) = .26, p < .53). In support of this result there were two additional questions, which asked about donations made to Venezuelan players “in-game” (in RuneScape). The data suggests that nobody did this.

H4. RuneScape players will self-report as being more aware about the gold farming situation compared to non-RuneScape players. This hypothesis was supported. As expected, the data suggested that members of the RuneScape community were more aware of the specific issue of gold farming, with a moderate correlation and a result that was unlikely to have arisen by chance, when considering the null hypothesis to be true (r (90) = .51, p < .001). Moreover, when comparing knowledge about gold farming between two groups – the RuneScape and non-RuneScape conditions – the effect size is moderate, and the result significant, (MWU= 487, p < .001, d = .52; Figure 1). However, this difference is much smaller when comparing just knowledge of the Venezuelan crisis, (MWU= 777, p < .016, d = .26). Indeed, the RuneScape community was correlated weakly to self-reported awareness of the crisis in general, despite its statistical significance (r = .26, p < .02).

3.2. Content analysis results

3.2.1. Data treatment

Once the comments had been coded, it was discovered that 158 comments (31%) did not align to any of the codes. Such comments had no relevance to the hypotheses of the research and for this reason were discarded. Thus, 351 comments remained. Next, a series of chi-square2) were done for each coded item. Statistically significant (p < .001) chi-square results were found for all 14 of the coded items, suggesting that the pattern of results observed for each independent code was unlikely to have arisen by what is expected by chance alone. The nature of chai-square statistics assumes no distribution, so data was not checked for normality.

3.2.2. Chi-square statistics

Results are again presented chronologically according to the original hypotheses.

H1a.Comments will be associated with more “individualising” rather than “binding” foundations. This first hypothesis was supported. Individualising foundations, Care high (χ2 (1, N = 109) = 166, p < .001), and Fairness high, χ2 (3, N = 50) = 329, p < .001) constituted a small majority (n = 184; 52.4%) of the sample, whereas the binding foundations, Loyalty high, χ2 (1, N = 73) = 259, p < .001, Authority high , χ2 (1, N = 5) = 489, p < .001, and Sanctity high , χ2 (1, N= 4) = 493, p < .001) constituted a moderate minority (n = 109; 31.1%).

H2a,Comments will reflect a range of moral foundations responses, but care/harm will be the most prevalent. This second hypothesis was also supported. The Care foundation accounted for the largest amount of variance of any code (n = 117; 33.3%) in the sample. Moreover, only 2.3% was coded as Care low, which meant that most comments coded as Care were in the other direction, χ2 (1, N= 8) = 478, p < .001. In other words, (n = 109) 31.1% of the comments were coded as Care high, suggesting that a large minority of commenters were sympathetic to the situation of Venezuela, χ2 (1, N = 109) = 166, p < .001.

H3a,Comments will be most associated with care, fairness, and loyalty foundations; authority and sanctity will not feature strongly. This third hypothesis was also supported. Care high (χ2 1, N = 8) = 478, p < .001), Fairness high, χ2 (3, N = 50) = 329, p < .001, and Loyalty high, χ2 (1, N = 73) = 259, p < .001, made up (n = 232) 65.9% of the total comments from the 14 codes total. The remaining two foundations – authority and sanctity – together accounted for 9.1% of the variance, suggesting that moral concerns regarding these foundations were not salient for commenters. Of these, however, authority low was reported most, accounting for 6% of the total variance, thereby suggesting that a minority of commenters supported breaking traditional rules in the context of gold farming behaviour, χ2 (1, N= 21) = 428, p < .001. In sum, then, a large majority (n = 253; 72.1%) of commenters expressed sentiment broadly supportive of Venezuelan gold farming, and each of these codes were (independent of each other) unlikely to have arisen by chance variation.

H4a. Because the videos present the conditions for the empathy-altruism hypothesis, the comments will also reflect this. Finally, this hypothesis was also supported. Batson (2011) identified three conditions under which people will tend to feel empathy for others, which in turn may generate altruistic action (discussed above). Comments associated with the empathy-altruism hypothesis were also high, making up a moderate minority (30.8%), χ2 (1, N = 108) = 169, p < .001. This finding is almost identical to the Care high code finding.

3.2.3. Additional findings

It is worth mentioning that the range of codes extended beyond moral foundations and empathy. For example, 13.4% of comments were sceptical, which meant either those commenters requested more information, engaged in critical reasoning, expressed frank doubt, or a combination of any of these, χ2 (3, N = 47) = 338, p< .001. A further 10.3% discussed political topics, χ2 (1, N = 36) = 375, p < .001, and 8.5% self-identified as being of Venezuelan nationality, χ2 (1, N = 30) = 396, p < .001. These ratios broadly support the ideas discussed earlier in the literature review. Specifically, the ideas that the internet is a place where a range of real human discussion, expression, and culture takes place, including in YouTube comments. There are, however, no hypotheses in this study to be tested regarding this claim.

Chaper 4: Discussion

4.1. Null results

H1. RuneScape players will have a stronger moral response to gold farming than non-RuneScape players. The theoretical justification for this hypothesis was that RuneScape players would likely have had, on average, more concrete interaction (as opposed to abstract interaction) with Venezuelans, either in-game or via YouTube videos, and therefore be more engaged in the events. Empirical studies presented earlier in the literature review gave credence to the idea that communication between people, whether through books, in the media, or in person is associated with broadening morals circles. The reverse can also be true: more interaction can cause a stronger moral response in the other direction, as in the Britain’s Brexit vote, where immigration concern was an important factor in how people formed their opinions (Smith, 2019). Hence, it was a reasonable hypothesis that RuneScape would have a stronger response - in either direction. However, the pattern of results indicated that the differences were minimal, nonsignificant, and the two groups were in fact well-matched their moral foundation adoption, even on the content-MFQ scale. There are at least two possible interpretations of such data. First, this pattern of results may reflect reality, in which case, the study fails in a specific prediction. Second, real differences may exist, but these were not captured due to small sample sizes. If the results reflect reality, the study does gain in credibility elsewhere because it means that the two groups are well matched in their MFQ scores, meaning that this is another variable the groups are similar on, which is useful when further associations on different variables (Sowell, 2011). Related to this was the null correlation between fairness and context-fairness. Perhaps when participants thought about fairness in relation to gold farming their answers changed from when thinking about fairness in an abstract way, though it was not clear to author how to interpret this pattern of results.

H3. RuneScape players will have donated more money to Venezuelan charities than non-RuneScape players. The theoretical justification for this hypothesis was tied up in historical and psychological data. The historical examples include situations where communication between people has happened in some manner (e.g., news media) and a charitable cause has been started because of this, as in 19th century British India (Brewis, 2010). Psychologically, the empathy-altruism hypothesis predicts that if certain conditions are met, like the perceived need of the other, value of the other, and perspective taking of the other, altruistic behaviour is likely to follow (Batson, 2011). Because RuneScape players will have had more perspective taking opportunities to gold farmers, are in the same group as them (RuneScape), which increases the likelihood of valuing the other, these factors may combine leading to an increased perceived need, and ultimately in more donations made to relevant charities. However, the questionnaire data does not support this conclusion. In the content analysis, however, two YouTube videos raised > $15,000 USD. It seems likely, then, that the specific sample surveyed here do not reflect the diversity of RuneScape players and failed to tap into what was a real effect – charitable giving.

4.2. The problem with context based moral foundations

As mentioned in the 3.1.2 Reliability section, Smith et al., (2017) estimated the heritability of moral foundations, and concluded they were weak, which then informed adopting the context-MFQ. When executed correctly, behavioural genetic research designs are powerful because their findings often overturn or challenge conventional wisdom across psychology (Plomin, 2018). Smith et al. (2017) suggested that moral foundations were more like ‘states’ than ‘traits’ found in personality psychology (Ashton, 2013), in the sense that they would change from time and place, rather than remaining more generalised. However, post-data collection, the present author later discovered that Smith et al’s study (2017) was critiqued by Haidt (2017), who pointed out their measurement error. In wave 1 of data collection the study used an earlier version of the MFQ, one that today is psychometrically redundant. Moreover, two new studies with behavioural genetic designs were recently conducted to estimate again the heritability of moral foundations, and their evidence points in the other direction: moral foundations, as expected, do have a substantial genetic basis, around 40% (Kandler, Penner, Richter & Zapko-Willmes, 2019; Zakharin & Bates, 2022). Such a finding is not surprising given the ubiquity of constructs that are genetically heritable (Turkheimer, 2000), however it does mean that the contextual-focused approach of this current studies research design is theoretically misinformed.

4.3. Main findings

First, the individualising/binding pattern was replicated. Second, the main finding from the questionnaire was that RuneScape players self-reported as being more aware about the gold farmers than did non-RuneScape players, meaning that being a member of the RuneScape community granted access to a particular piece of local knowledge (gold farming) compared to the sample not in the RuneScape community. The third main finding came from the content analysis, which showed that 72.1% of commenters expressed positive sentiment towards gold farming activity, a result unlikely to have arisen by chance variation. Not only is positive sentiment expressed in a textual form, but these comments are also associated with a total of $15,000 USD raised for charity. Together, these findings tentatively support the idea that digital communication (RuneScape and YouTube) is associated with a broadening moral circle (hard money donated to charitable ends). Triangulation permitted these findings: without the content analysis, the study would find that donations were sparse; without the questionnaire, it would have been impossible to discover that being in the RuneScape community is associated with unique, local knowledge (gold farming).

4.4. Implications

Because the internet is a decentralised communication system purposely built to withstand nuclear war, to reduce the problem of geography, and is rapidly evolving due to technological advancement, it is increasingly having more impact in daily life around the world (Ryan, 2013). Healthcare, education, and currency are all increasingly digital (Ammos, 2018; Chalmers, 2022). Similarly, the world of work is increasingly affected (e.g., “work from home”). Regarding work, this study looked at people dubbed “digital migrants”, because they are migrating for work not to larger cities or countries, but online. The specific implication of this study is that digital migration can sometimes mean that two groups of people – one in poverty and one not – can meet, learn about each other, and reach prosocial outcomes (like raising $15,000 USD for charity). Assuming this analysis to be correct (at least partly), two implications follow. First, if more high-profile RuneScape content creators were involved in appropriate charitable causes, more money would likely reach Venezuelans in poverty through video makers donating fan bases. Second, if the games company owner Jagex did something similar, like incentivising (by making it easy for) RuneScape players to donate to Venezuelan charities or directly to individual Venezuelans, either with “gold” or fiat currency, probably many more players would.

4.5. Limitations

This study has four main limitations. First, causality was not tested, so no causal conclusions can be drawn. Perhaps the type of people to watch RuneScape YouTube videos are, for example, interested in international affairs or inherently kinder than the average person. Spurious correlations abound and this could be another (Vigen, 2022). Second, the MFQ-20 was unreliable for two foundations. Third, relative to the size of RuneScape players, which is likely over > 300,000 individuals, sampling 49 individuals is small, increasing the likelihood of extreme results (MMO Populations, 2022). Fourth, this research was executed independent of Jagex (owners of RuneScape). There is evidence that research with RuneScape independent of Jagex are associated with higher attrition rates and smaller sample sizes (Bilir, 2009; Adams, 2014). This is probably because if Jagex were on board, they could advertise, bring interest, trust, and legitimacy to the study.

4.6. Future research

Many avenues for studying digital migration exist. One can imagine a qualitative or ethnographical account detailing the experiences of people who are “gaming to live”, which cannot be straightforward or easy. Another may look to manipulate causality, in line with Batson’s (2011) recommendation for studying empathy. A different study might attempt to replicate or challenge the findings here, but with much larger sample sizes. A future research team should try to work in concert with Jagex, because when previous studies have done this, not only have the sample sizes been larger, but the attrition rates have been much lower (Walker, 2015; Smith et al., 2020). Then, one could use the psychometrically more reliable MFQ-30 or other measures. Finally, different textual data could be analysed (like Reddit posts) to test if the observed positive moral response from the RuneScape community holds across different modes of communication.

References

Adams, M., Amanda. (2014). The effects of self-control and social learning variables on deviance and social control in a virtual world (Doctoral thesis). https://ufdc.ufl.edu/UFE0046506/00001

Agbo, A. A. (2010). Cronbach's Alpha: Review of Limitations and Associated Recommendations. Journal of Psychology in Africa, 20(2), 233-239. 10.1080/14330237.2010.10820371

Ammous, S. (2018). The Bitcoin Standard: The Decentralized Alternative to Central Banking. John Wiley & Sons, Incorporated.

Archer, J. (2006). Cross-Cultural Differences in Physical Aggression Between Partners: A Social-Role Analysis. Personality and Social Psychology Review, 10(2), 133-153. 10.1207/s15327957pspr1002_3

Aronczyk, A., & Beras, E. (2021). Video gaming the system. National Public Radio. https://www.npr.org/transcripts/1018915121

Ashton, M. C. (2013). Individual differences and personality (2nd ed.). Amsterdam: Academic Press.

Atari, M., Graham, J., & Dehghani, M. (2020). Foundations of morality in Iran. Evolution and Human Behavior, 41(5), 367-384. 10.1016/j.evolhumbehav.2020.07.014

Baron-Cohen, S. (2012). The science of evil. Cambridge: Tantor Media Inc.

Batson, C. D. (1981). Is empathic emotion a source of altruistic motivation? Journal of Personality and Social Psychology, 40(2), 290-302. 10.1037/0022-3514.40.2.290

Batson, C. D., & Ahmad, N. (2001). Empathy-induced altruism in a prisoner's dilemma II: what if the target of empathy has defected? European Journal of Social Psychology, 31(1), 25-36. 10.1002/ejsp.26

Batson, C. D., Klein, T. R., Highberger, L., & Shaw, L. L. (1995). Immorality from empathy-induced altruism: When compassion and justice conflict. Journal of Personality and Social Psychology, 68(6), 1042-1054. doi:10.1037/0022-3514.68.6.1042

Batson, C. D., Batson, J. G., Slingsby, J. K., Harrell, K. L., Peekna, H. M., & Todd, R. M. (1991). Empathic Joy and the Empathy-Altruism Hypothesis. Journal of Personality and Social Psychology, 61(3), 413-426. 10.1037/0022-3514.61.3.413

Batson, C. D., & Moran, T. (1999). Empathy-induced altruism in a prisoner's dilemma. European Journal of Social Psychology, 29(7), 909-924. 10.1002/(SICI)1099-0992(199911)29:7<909::AID-EJSP965>3.0.CO;2-L

Batson, D. (2011). Altruism in Humans. Oxford: Oxford University Press.

Bilir, T. (2009). Real economics in virtual worlds: a massive multiplayer online game case study, RuneScape (master’s thesis). https://smartech.gatech.edu/handle/1853/31657

Bloom, P. (2016). Against empathy. New York: The Bodley Head.

Bloom, P. (2010). How do morals change? Nature, 464(7288), 490. 10.1038/464490a

Brewis, G. (2010). ‘Fill Full the Mouth of Famine’: Voluntary Action in Famine Relief in India 1896–1901. Modern Asian Studies, 44(4), 887-918. 10.1017/S0026749X0999031X

Brown, Donald E. (1991). Human universals. New York: McGraw-Hill.

Chalmers, D. (2022). Reality+.New York: Penguin.

Correlation and Causation: Tyler Vigen's Spurious Correlations.(2022).[Video/DVD] 10.4135/9781473997714 https://methods.sagepub.com/video/correlation-and-causation-tyler-vigens-spurious-correlations

Crimston, C. R., Bain, P. G., Hornsey, M. J., & Bastian, B. (2016). Moral Expansiveness: Examining Variability in the Extension of the Moral World. Journal of Personality and Social Psychology, 111(4), 636-653. 10.1037/pspp0000086

Crowe, N., & Bradford, S. (2006). 'Hanging out in Runescape': Identity, Work and Leisure in the Virtual Playground. Children's Geographies, 4(3), 331-346. 10.1080/14733280601005740

Curry, C., O. (2019). What’s Wrong with Moral Foundations Theory, and How to get Moral Psychology Right. https://behavioralscientist.org/. https://behavioralscientist.org/whats-wrong-with-moral-foundations-theory-and-how-to-get-moral-psychology-right/

Curry, O. S., Jones Chesters, M., & Van Lissa, C. J. (2019). Mapping morality with a compass: Testing the theory of ‘morality-as-cooperation’ with a new questionnaire. Journal of Research in Personality, 78, 106-124. 10.1016/j.jrp.2018.10.008

Dawkins, R. (2016). The selfish gene (40th ed.). Oxford: Oxford University Press.

Diamond, J. M. (1999). Guns, germs, and steel. New York: W. W. Norton.

de Waal, Frans. (2010). The age of empathy. London: Souvenir Press Ltd.

Decety, J., Echols, S., & Correll, J. (2010). The blame game: The effect of responsibility and social stigma on empathy for pain. Journal of Cognitive Neuroscience, 22(5), 985-997. doi:10.1162/jocn.2009.21266

di Pellegrino, G., Fadiga, L., Fogassi, L., Gallese, V., & Rizzolatti, G. (1992). Understanding motor events: a neurophysiological study. Experimental Brain Research, 91(1), 176-180. 10.1007/BF00230027

Doğruyol, B., Alper, S., & Yilmaz, O. (2019). The five-factor model of the moral foundations theory is stable across WEIRD and non-WEIRD cultures. Personality and Individual Differences, 151, 109547. 10.1016/j.paid.2019.109547

Edelstein, D., Findlen, P., Ceserani, G., Winterer, C., & Coleman, N. (2017). Historical research in a digital age: Reflections from the mapping the republic of letters project. The American Historical Review, 122(2), 400-424. doi:10.1093/ahr/122.2.400

Everett, J. A. C., Faber, N. S., & Crockett, M. (2015). Preferences and beliefs in ingroup favoritism. Frontiers in Behavioral Neuroscience, 9, 15. 10.3389/fnbeh.2015.00015

Feenberg, A., & Bakardjieva, M. (2004). Virtual Community: No ‘Killer Implication’. New Media & Society, 6(1), 37-43. 10.1177/1461444804039904

Fiserman. (2017). Venezuelans Turn to Playing Runescape in Order To Survive. https://steemit.com/. https://steemit.com/steemit/@fiserman/venezuelans-turn-to-playing-runescape-in-order-to-survive

Fiske, S. T. (2015a). Intergroup biases: a focus on stereotype content. Current Opinion in Behavioral Sciences, 3, 45-50. 10.1016/j.cobeha.2015.01.010

Funk, M. (2020). How Does YouTube Count Views? www.tubics.com https://www.tubics.com/blog/what-counts-as-a-view-on-youtube

Gallegos, R. (2016). Crude nation. New York: Potomac Books Inc.

Graham, J., Haidt, J., & Nosek, B. A. (2009). Liberals and Conservatives Rely on Different Sets of Moral Foundations. Journal of Personality and Social Psychology, 96(5), 1029-1046. 10.1037/a0015141

Graham, J., Nosek, B. A., Haidt, J., Iyer, R., Koleva, S., & Ditto, P. H. (2011). Mapping the Moral Domain. Journal of Personality and Social Psychology, 101(2), 366-385. 10.1037/a0021847

Graham, J., Waytz, A., Meindl, P., Iyer, R., & Young, L. (2017). Centripetal and centrifugal forces in the moral circle: Competing constraints on moral learning. Cognition, 167, 58-65. 10.1016/j.cognition.2016.12.001

Haidt, J. (2013). The righteous mind. New York: Penguin Books.

Haidt, J. (2017). Are moral foundations heritable? Probably. www.righteousmind.com. https://righteousmind.com/are-moral-foundations-heritable-probably/

Haidt, J., & Graham, J. (2007). When Morality Opposes Justice: Conservatives Have Moral Intuitions that Liberals may not Recognize. Social Justice Research, 20(1), 98-116. 10.1007/s11211-007-0034-z

Haidt, J., & Joseph, C. (2004). Intuitive Ethics: How Innately Prepared Intuitions Generate Culturally Variable Virtues. Daedalus (Cambridge, Mass.), 133(4), 55-66. 10.1162/0011526042365555

Harari, Y. (2014). Sapiens. New York: Random House.

Harris, L. T., & Fiske, S. T. (2006). Dehumanizing the Lowest of the Low: Neuroimaging Responses to Extreme Out-Groups. Psychological Science, 17(10), 847-853. 10.1111/j.1467-9280.2006.01793.x

Hein, G., Silani, G., Preuschoff, K., Batson, C. D., & Singer, T. (2010). Neural responses to ingroup and outgroup members' suffering predict individual differences in costly helping. Neuron (Cambridge, Mass., 68(1), 149-160. doi:10.1016/j.neuron.2010.09.003

Jagex. (2022). Rules of RuneScape. https://www.jagex.com/en-GB/terms/rules-of-runescape

Kandler, C., Penner, A., Richter, J., & Zapko-Willmes, A. (2019). The Study of Personality Architecture and Dynamics (SPeADy): A Longitudinal and Extended Twin Family Study. Twin Research and Human Genetics, 22(6), 548-553. 10.1017/thg.2019.62

Kivikangas, J. M., Fernández-Castilla, B., Järvelä, S., Ravaja, N., & Lönnqvist, J. (2021). Moral foundations and political orientation: Systematic review and meta-analysis. Psychological Bulletin, 147(1), 55-94. 10.1037/bul0000308

Krebs, D. (1975). Empathy and Altruism. Journal of Personality and Social Psychology, 32(6)

Kwik, J. (2020). Limitless. New York: Hay House Inc.

Lang, K., Little, T. D., & FSW, E. U. (2018). Principled missing data treatments. Prevention Science, 19(3), 284-294. 10.1007/s11121-016-0644-5

Marcus, Gary F. (2004). The birth of the mind. New York: Basic Books.

Marr, B. (2018). How Much Data Do We Create Every Day? The Mind-Blowing Stats Everyone Should Read. https://www.forbes.com/?sh=1ac8fa802254. https://www.forbes.com/sites/bernardmarr/2018/05/21/how-much-data-do-we-create-every-day-the-mind-blowing-stats-everyone-should-read/?sh=5499229c60ba

Miller, E. D. (2018). Content analysis of YouTube comments from differing videos: an overview and key methodological considerations. SAGE Publications Ltd.

Miller, E. D. (2019). Codifying gradients of evil in select YouTube comment postings. Human Behavior and Emerging Technologies, 1(3), 216-222. 10.1002/hbe2.155

Miller, E., D. (2015). Content analysis of select YouTube postings: comparisons of reactions to the Sandy Hook and Aurora shootings and hurricane Sandy. Cyberpsychology, Behavior, and Social Networking, 18(11)

Newport Frank. (2021). Homosexuality. https://news.gallup.com/poll/9916/homosexuality.aspx

Nilsson, A., Erlandsson, A., & Västfjäll, D. (2016). The congruency between moral foundations and intentions to donate, self-reported donations, and actual donations to charity. Journal of Research in Personality, 65, 22-29. 10.1016/j.jrp.2016.07.001

Nilsson, A., Erlandsson, A., & Västfjäll, D. (2020). Moral Foundations Theory and the Psychology of Charitable Giving. European Journal of Personality, 34(3), 431-447. 10.1002/per.2256

Old-school RuneScape Stats. (2022). https://mmo-population.com/. https://mmo-population.com/r/2007scape/stats

Ombler, M. (2020). How RuneScape is helping Venezuelans survive. https://www.polygon.com/. https://www.polygon.com/features/2020/5/27/21265613/runescape-is-helping-venezuelans-survive

Orcan, F. (2020). Parametric or Non-Parametric: Skewness to Test Normality for Mean Comparison.International Journal of Assessment Tools in Education, 7(2), 255.

Pearce, J. S. (2017). Lafayette strong: A content analysis of grief and support online following a theater shooting. Illness, Crisis, and Loss, 28(4), 299-320. doi:10.1177/1054137317742234

Pinker, S. (2011). The better angels of our nature. New York: Viking Books.

Pinker, S. (2003). The blank slate. New York: Penguin.

Plomin, R. (2018). Blueprint. London: Penguin.

Powell, C. (2019, Venezuela’s paper currency is worthless, so its people seek virtual gold. The Economist, https://www.economist.com/the-americas/2019/11/21/venezuelas-paper-currency-is-worthless-so-its-people-seek-virtual-gold

Reilly, P. (2014). The 'battle of Stokes Croft' on YouTube: the development of an ethical stance for the study of online comments. SAGE Publications.

Remarque, & M, E. (1929). All quiet on the western front. London: Brown and Company.

Robe, I. (2018). Inescapably social: dimensions of self-construction in the virtual social world of RuneScape (master’s thesis) https://dc.etsu.edu/etd/3375/

Ryan, J. (2013). A history of the internet and the digital future. New York: Reakiton Books.

Shweder, A. R. (1991). Thinking through cultures. New York: Harvard University Press.

Singer, P. (2019). The life you can save. New York: Random Books House.

Singer, P. (2011). The expanding circle. Princeton: Princeton University Press.

Smith, A., & Haakonssen, K. (2002). The theory of moral sentiments. Cambridge University Press.

Smith, M. C, Rauwolf, P, Intriligator, J & Rogers, D. Robert. (2020). Hostility Is Associated with Self-Reported Cognitive and Social Benefits Across Massively Multiplayer Online Role-Playing Game Player Roles. Cyberpsychology, Behavior, and Social Networking, 23(7) https://doi.org/10.1089/cyber.2019.0349

Smith, D. S. (2019). Shaping the modern world with a stone-age brain: Brexit and the Moral Foundations Theory. Journal of Social and Political Psychology, 7(2), 863-889. 10.5964/jspp.v7i2.1032

Smith, K. B., Alford, J. R., Hibbing, J. R., Martin, N. G., & Hatemi, P. K. (2017). Intuitive Ethics and Political Orientations: Testing Moral Foundations as a Theory of Political Ideology. American Journal of Political Science, 61(2), 424-437. 10.1111/ajps.12255

Sowell, T. (2011). Economic facts and fallacies. New York: Basic books.

Taleb, N. N. (2007). The black swan. London: Allen Lane.

The 17 goals. (2015). United Nations. https://sdgs.un.org/goals

Thomas, M. (2019). Poverty and Politics in Venezuela. https://borgenproject.org/. https://borgenproject.org/tag/poverty-in-venezuela/

Turner, S. F., Cardinal, L. B., & Burton, R. M. (2017). Research Design for Mixed Methods: A Triangulation-based Framework and Roadmap. Organizational Research Methods, 20(2), 243-267. 10.1177/1094428115610808

Turkheimer, E. (2000). Three Laws of Behavior Genetics and What They Mean. Current Directions in Psychological Science, 9(5), 160-164. 10.1111/1467-8721.00084

Twain, M. (1885). The adventures of Huckleberry Finn. New York: Charles L. Webster And Company.

Walker, B. (2015). Notions of community in a Massively Multiplayer Online Role-Playing Game (Unpublished master’s thesis). University of London

Weiss, B. (2021). The Venezuelans Trying to Escape Their Country Through Video Game Grunt Work. https://slate.com/. https://slate.com/technology/2021/08/venezuelans-old-school-runescape-tasks.html

Wells., G. H. (1932). The outline of history (7th ed.). London: Cassell and Company, Ltd.

Yin, K. R. (2006). Mixed Methods Research: Are the Methods Genuinely Integrated or Merely Parallel?1. Research in the Schools, 13(1), 41.

Zakharin, M., & Bates, T. C. (2022). Testing heritability of moral foundations: Common pathway models support strong heritability for the five moral foundations. European Journal of Personality, 89020702211039. 10.1177/08902070221103957

Appendix

Appendix 1

Statistical output, tables, distributions

Table 1

 Table 1

Group differences between members of the RuneScape community and non-RuneScape players in education

Member of the RuneScape community?

Education

        Yes

        No

Completed high school/secondary school

7

2

Completed sixth form/pre-university

8

6

Bachelor’s degree from university

24

19

Master’s degree from university

10

10

PhD completed

0

4

Table 2

 Table 2

Group differences between RuneScape community members and non-RuneScape players in employment

                            Member of the RuneScape community?

Employment

                                                                    Yes

No

Yes

34

26

No

3

3

I am a student in full time education

10

9

Other

2

3

Table 3

Table 3

Distributions of moral foundations and associated descriptive statistics

Skewness

Kurtosis

Shapiro-Wilk

Moral foundations and context-moral foundations

N

Mean

SD

Skewness

SE

Kurtosis

SE

W

p

Care/harm

90

4.61

0.757

-0.5689

0.254

-0.2865

0.503

0.950

0.002

Fairness/Cheating

90

4.61

0.762

-0.7673

0.254

0.6565

0.503

0.953

0.002

Loyalty/Betrayal

90

2.83

1.032

0.5779

0.254

-0.4660

0.503

0.949

0.001

Authority/Subversion

90

2.83

1.087

0.0227

0.254

-1.1490

0.503

0.953

0.003

Sanctity/Degradation

90

2.93

1.180

0.0724

0.254

-0.7868

0.503

0.966

0.018

Context-Care/harm

90

4.65

0.946

-0.5432

0.254

-0.0796

0.503

0.951

0.002

Context-Fairness/cheating

90

4.63

0.924

-1.0737

0.254

2.2342

0.503

0.932

< .001

Context-Loyalty/betrayal

90

2.87

0.993

0.6105

0.254

-0.2444

0.503

0.954

0.003

Context-Authority/subversion

90

3.11

1.093

0.1771

0.254

-0.7178

0.503

0.976

0.096

Context-Sanctity/degradation

90

1.48

0.724

1.6953

0.254

2.4189

0.503

0.713

< .001

Individualising

90

4.61

0.632

-0.4383

0.254

-0.3128

0.503

0.972

0.050

Binding

90

2.86

0.978

0.2246

0.254

-0.8189

0.503

0.973

0.055

Context-Individualising

90

4.64

0.818

-0.9178

0.254

1.5956

0.503

0.945

< .001

Context-Binding

90

2.49

0.781

0.5865

0.254

-0.4306

0.503

0.952

0.002

 

 

Table 4

 

 

 

 

Table 4

 Mean scores and standard deviations for moral foundations

Moral foundation metrics 

N

Mean

SD

Care/harm

90

4.61

0.757

Fairness/cheating

90

4.61

0.762

Loyalty/betrayal

90

2.83

1.032

Authority/subversion

90

2.83

1.087

Sanctity/degradation

90

2.93

1.180

Context-Care/harm

90

4.65

0.946

Context-Fairness/cheating

90

4.63

0.924

Context-Loyalty/betrayal

90

2.87

0.993

Context-Authority/subversion

90

3.11

1.093

Context-Sanctity/degradation

90

1.48

0.724

Individualising

90

4.61

0.632

Binding

90

2.86

0.978

Context-Individualising

90

4.64

0.818

Context-Binding

90

2.49

0.781

Chi-square results:

Care low:

Proportion Test (N Outcomes)

Proportions - Care low

Level

 

Count

Proportion

0

Observed

501

0.9843

 

Expected

255

0.500

1

Observed

8

0.0157

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

478

1

< .001

Care high:

Proportion Test (N Outcomes)

Proportions - Care high

Level

 

Count

Proportion

0

Observed

400

0.786

 

Expected

255

0.500

1

Observed

109

0.214

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

166

1

< .001

Fair low:

Proportion Test (N Outcomes)

Proportions - Fair low

Level

 

Count

Proportion

0

Observed

492

0.9666

 

Expected

255

0.500

1

Observed

17

0.0334

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

443

1

< .001

Fair high:

Proportion Test (N Outcomes)

Proportions - Fair high

Level

 

Count

Proportion

0

Observed

459

0.9018

 

Expected

255

0.500

1

Observed

50

0.0982

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

329

1

< .001

Proportion Test (N Outcomes)

Proportions - Loyalty low

Level

 

Count

Proportion

0

Observed

505

0.99214

 

Expected

255

0.500

1

Observed

4

0.00786

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

493

1

< .001

Proportion Test (N Outcomes)

Proportions - Loyalty high

Level

 

Count

Proportion

0

Observed

436

0.857

 

Expected

255

0.500

1

Observed

73

0.143

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

259

1

< .001

Proportion Test (N Outcomes)

Proportions - Authority low

Level

 

Count

Proportion

0

Observed

488

0.9587

 

Expected

255

0.500

1

Observed

21

0.0413

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

428

1

< .001

Proportion Test (N Outcomes)

Proportions - Authority high

Level

 

Count

Proportion

0

Observed

504

0.99018

 

Expected

255

0.500

1

Observed

5

0.00982

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

489

1

< .001

Proportion Test (N Outcomes)

Proportions - Sanctity low

Level

 

Count

Proportion

0

Observed

507

0.99607

 

Expected

255

0.500

1

Observed

2

0.00393

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

501

1

< .001

Proportion Test (N Outcomes)

Proportions - Sanctity high

Level

 

Count

Proportion

0

Observed

505

0.99214

 

Expected

255

0.500

1

Observed

4

0.00786

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

493

1

< .001

Proportion Test (N Outcomes)

Proportions - Empathy-altruism

Level

 

Count

Proportion

0

Observed

401

0.788

 

Expected

255

0.500

1

Observed

108

0.212

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

169

1

< .001

Proportion Test (N Outcomes)

Proportions - Sceptical

Level

 

Count

Proportion

0

Observed

462

0.9077

 

Expected

255

0.500

1

Observed

47

0.0923

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

338

1

< .001

Proportion Test (N Outcomes)

Proportions - Political

Level

 

Count

Proportion

0

Observed

473

0.9293

 

Expected

255

0.500

1

Observed

36

0.0707

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

375

1

< .001

Proportion Test (N Outcomes)

Proportions - Venezuelan

Level

 

Count

Proportion

0

Observed

479

0.9411

 

Expected

255

0.500

1

Observed

30

0.0589

 

Expected

255

0.500

χ² Goodness of Fit

χ²

df

p

396

1

< .001