[Proposal] The moral weight of electronic waste: modelling moralisation over time

Carolyn B. McNabb1, Paul Vanags1, & Kathryn B. Francis2

1 University of Reading, UK
2 Keele University, UK


Living in the 21st century brings with it an unprecedented number of social and environmental issues, each demanding global scale behavioural change. Although many of these issues (e.g. climate change, poverty and disease) threaten to harm the world we live in or the people in it, not every issue will be considered a moral one. This may be because human morality is poorly equipped to deal with these kinds of challenges. After all, we’ve evolved to solve problems on a much smaller scale, like surviving in a hunter-gatherer environment. However, assigning moral significance to an issue can have powerful influences on human behaviour (Francis & McNabb, 2020; Prosser et al., 2020) so understanding how issues become moralised is crucial to formulating effective approaches for tackling major societal issues.

To date, research investigating moralisation (when an issue becomes morally loaded) has been limited to the subject of meat-eating (Feinberg et al., 2019). While this work has provided insight into how different cognitive and emotional experiences contribute to the process of moralisation, meat-eating is distinct from other issues in many ways. It is therefore important to understand whether these characteristics are relevant for the moralisation of other issues, not just meat-eating.

Our research will investigate how different cognitive and emotional factors contribute to the moralisation of a comparatively unacknowledged but fast-growing issue, the disposal of electronic waste (e-waste). Each year, humans generate over 50 million tonnes of e-waste; this includes old computers, phones and many household appliances. Most of this is destined for landfills where it leaches toxic chemicals such as mercury and lead into soil and water supplies. Even more alarmingly, much of the e-waste that is “recycled” is shipped to the world’s poorest countries, where raw materials are extracted for resale using hazardous techniques that are harmful to workers; sadly, these workers are often children. These environmental and anti-humanitarian consequences, along with its relative unfamiliarity, make e-waste an ideal issue for studying the process of moralisation.


The purpose of this study is to evaluate whether moralisation is driven by differences/changes in moral cognition, moral emotion and other features such as personal pleasure (hedonic motivation) and denial (dissonance reduction strategies). Building on previous work by Feinberg et al., 2019 on the moralisation of meat-eating, our study will investigate whether the same factors contribute to the moralisation of e-waste, a comparatively neutral issue that we anticipate will become moralised over time.

Work from our pilot study comparing four issues, including meat-eating, e-waste, self-driving cars and ocean plastic, demonstrated that e-waste became moralised to the greatest extent when compared to the other issues, partially owing to its relative lack of familiarity at baseline (Fig. 1).

Figure 1. Familiarity and mean moral score before and after viewing an educational video about each issue.

The proposed study will be an online experiment (Gorilla) including four sessions (Fig. 2), three of which will include an educational video about the environmental and humanitarian consequences of e-waste. Participants will first rate statements related to their moral identity and their trait reactance. They will then be asked about their behavioural intentions relating to e-waste. They will also rate statements related to moral emotions, cognition/attitudes, dissonance reduction, hedonic motivations, and the moralisation of e-waste. Participants will then watch a ~6 min video about e-waste. After 5 days, participants respond to the set of questionnaires outlined in Fig. 2 and view another video about e-waste (repeated at Time 3). At Time 4, participants respond to the same questionnaires, and are then asked about their behavioural intentions related to e-waste.

Figure 2. Study design, created with BioRender.com

We hypothesise that different cognitive and emotional factors will contribute different amounts to the moralisation of e-waste. To test this hypothesis, we will use elastic-net regression (a type of prediction model) and cross-lagged analysis (to measure how these factors interact over time). As we expect that moralisation also leads to changes in behaviour, we will investigate behavioural intentions related to e-waste at Time 1 and 4. By using these types of analyses, we can learn more about the individual contributions of emotion, cognition, dissonance reduction and hedonic motivation to the moralisation of previously neutral issues and whether moralisation really does lead to changes in behaviour.


In order to truly understand how cognitive and emotional processes (such as those outlined above) influence moralisation, advanced statistical techniques, such as cross-lagged analysis and elastic net regression, will be employed.
For the purposes of power analysis, if we simplify this model to a multiple linear regression with 10 predictor variables (i.e. Time 1 and change in response for each of the measures), we would have 80% power (when 𝞪 = .05) to detect a small effect of f2 =.02 with N=822 participants (Linear multiple regression: Fixed model, R2 increase; G*Power version As participant drop-out can be a factor in longitudinal studies, recruitment targets will be based on “complete” datasets – assuming a drop-out rate of 10%, we anticipate that approximately 987 participants will need to be recruited into the study.


Participants who take part in the online experiment will be paid £7/hour for their participation in the study + a £1 bonus payment for completion.

We anticipate paying 822 participants in full (including bonus payment) but also expect that approximately 83 participants (10%) will partially complete the study, thus receiving part payment for their time.

We have budgeted for £9,585.64 in total:

822 participants @ £7/h + £1 bonus = £6,576
83 participants @ £7/h (50% completion) = £290.50
33% Prolific service charge = £2,265.95
VAT (20% of service charge) = £453.19


This study and analysis have been preregistered: https://aspredicted.org/fd3sv.pdf


We plan to make all anonymised data and analysis scripts available on our OSF page and all experimental materials publicly available on Gorilla (links will be added to the OSF). We will publish any manuscripts resulting from the study in open-access journals and make pre-prints available for public comment prior to publication.