Introduction
Sludge describes unnecessary frictions that prevent people from achieving their goals (Thaler, Reference Thaler2018; Sunstein, Reference Sunstein2021; Mills, Reference Mills2023; Newall, Reference Newall2023). Examples of sludge include overly complex paperwork requirements, confusing websites, contractual terms that are hidden or challenging to understand, difficulties in cancelling membership fees, and excessive or unexplained waiting times (NSW Behavioural Insights Unit, 2024). These frictions can reduce adoption rates of policies and programmes (Bearson and Sunstein, Reference Bearson and Sunstein2023) and can reduce the quality of service delivery and user experience. Sludge can also have detrimental effects on people’s experiences, including a sense of frustration and indignity (Sunstein, Reference Sunstein2021). Sludge and administrative burdens are pervasive in public and private settings (Herd and Moynihan, Reference Herd and Moynihan2019) and increasingly receive worldwide interest as governments and international organisations aim to understand how to identify and reduce sludge to improve processes (NSW Behavioural Insights Unit, 2024; OECD, 2024; OIRA, 2024).
Despite that interest, we are currently missing a comprehensive understanding of what makes people more or less vulnerable to different types of sludge. Research on administrative burdens shows that some groups (usually vulnerable populations such as those in poor health or with low incomes) are more affected than others by behavioural frictions when seeking access to government benefits (Heinrich, Reference Heinrich2016; Christensen et al., Reference Christensen, Aarøe, Baekgaard, Herd and Moynihan2020; Sunstein, Reference Sunstein2020; Herd et al., Reference Herd, Hoynes, Michener and Moynihan2023; Halling and Baekgaard, Reference Halling and Baekgaard2024; Martin, Reference Martin2024). But behavioural and psychological factors that may compound this inequality are less well-understood. For example, time or cognitive scarcity (Mullainathan and Shafir, Reference Mullainathan and Shafir2013; Christensen et al., Reference Christensen, Aarøe, Baekgaard, Herd and Moynihan2020), present bias (Laibson, Reference Laibson1997), or an absence of skills like administrative literacy (Döring, Reference Döring2021; Masood and Azfar Nisar, Reference Masood and Azfar Nisar2021) may all increase vulnerability to sludge.
We also have a very limited understanding of how acceptable people view different types of sludge to be. While there is a large and growing literature on the acceptability of nudges (e.g. Sunstein, Reference Sunstein2016), little is known about when and why sludge is seen as acceptable or not. To some extent we can rely here on the emerging literature on tolerance of administrative burdens, which shows that people differ in their attitudes towards frictions in government policies such as social welfare benefits, and that both political beliefs and personal experiences with benefits affect tolerance (Halling et al., Reference Halling, Herd and Moynihan2023; Baekgaard et al., Reference Baekgaard, Halling and Moynihan2025). But this research remains preliminary, and because sludge and administrative burden are not quite the same (see below), more research on sludge acceptability is needed.
Most existing work on vulnerability to and acceptability of sludge and administrative burden comes from policy case studies in specific contexts (Herd and Moynihan, Reference Herd and Moynihan2019; Halling and Baekgaard, Reference Halling and Baekgaard2024). Hence, there is little opportunity to compare sludge vulnerability and acceptability across contexts. It may be the case that people are more vulnerable to and less accepting of some types of sludge than others, but we have limited data about that. The data we have comes from comparisons of administrative experiences across domains such as tax, retirement, government benefits, bills, etc. (Martin et al., Reference Martin, Delaney and Doyle2024) and comparisons of burden tolerance across public sector domains such as income supports, health insurance, passport renewals, and small business licensing (Baekgaard et al., Reference Baekgaard, Halling and Moynihan2025). However, these studies do not compare different types of sludge (e.g. complex paperwork versus confusing websites), and there are no comparisons between public and private sector sludge. For example, sludge encountered in the public sector when dealing with the government might be seen as less acceptable than sludge encountered when dealing with a private company in the private sector. Or some people might be more vulnerable to sludge related to complex paperwork than to long waiting times and having to deal with unpleasant staff members. We know that sludge is often harmful and assume that people usually dislike sludge, but we do not know what types of sludge are more or less harmful and disliked, if this depends on the context, and for whom.
Our study is the first to measure and compare sludge vulnerability and acceptability across different types and contexts of sludge. To do so, we rely on a nationally representative survey from Ireland. Using 20 experimentally varied sludge scenarios, we test what factors predict people’s self-reported vulnerability to different types and contexts of sludge, and how acceptable they think this sludge is, based on a range of potential factors (e.g. socio-economic and behavioural).
Analysing vulnerability to and acceptability of different types of sludge is a novel contribution, yet this initial study has certain limitations that suggest important directions for future research. Rather than relying solely on surveys that provide correlational insights, future studies should employ experimental designs to measure causal effects. Instead of relying on self-reported evaluations of sludge, future research should investigate how people actually interact with and navigate real-world sludge. Moreover, rather than focusing exclusively on questions asking people how they would react to an abstract hypothetical task, future research should also examine reactions to concrete sludge that participants have directly experienced. It would also be useful to obtain more clarity about how, when, and what magnitudes matter. For example, waiting time of a specified and moderate amount might be tolerable, more tolerable that exceptionally difficult cancellation processes, but waiting time of a specified and large amount might be intolerable, more intolerable than moderately difficult cancellation processes.
The paper is structured as follows. Section 2 reviews the nascent literature on measuring sludge and the more mature literature on administrative burden. Section 3 presents the methodology including procedure, participants, measures, and analysis strategy. Section 4 presents the results of the analyses of self-reported vulnerability to sludge. Section 5 presents the results of the analyses of acceptability of sludge. Section 6 discusses the findings and concludes with suggestions for future research.
Motivation and literature review
Our study builds on both sludge research from behavioural science and administrative burden research from public administration. The two areas of research are related but also conceptually distinct (Madsen et al., Reference Madsen, Mikkelsen and Moynihan2022): sludge often refers to excessive or unnecessary behavioural frictions (Sunstein, Reference Sunstein2021) and may occur in any context where such frictions make it harder for people to act in their own best interest, including in consumer and business processes such as cancelling online services (Mills et al., Reference Mills, Whittle, Ahmed, Walsh and Wessel2023). By contrast, administrative burdens refer to the costs that people experience when they interact with government policy implementation, including due to the presence of frictions, in contexts such as obtaining access to social welfare benefits or public healthcare (Moynihan et al., Reference Moynihan, Herd and Harvey2015).
The typologies and language used to describe sludge and administrative burdens reflect their overlap and differences: administrative burdens involve learning costs (e.g. the effort needed to find out if one is eligible for a welfare programme), compliance costs (e.g. the time spent filling out complex forms to prove eligibility) and psychological costs (e.g. the associated stress, stigma or loss of autonomy) (Moynihan et al., Reference Moynihan, Herd and Harvey2015), while sludge, allowing for both public and private sector processes, has been said to involve search costs (e.g. frictions experienced when looking for options), evaluation costs (e.g. frictions experienced when comparing options), implementation costs (e.g. frictions experienced when pursuing chosen options) and again psychological costs (e.g. from pressure selling or scarcity messaging) (OECD, 2024; Shahab and Lades, Reference Shahab and Lades2024). It is evident that the two sets of costs overlap.
This study uses the term ‘sludge’ for internal consistency and because the study covers both private and public sector sludge, but we draw on both literatures to inform the study on potential predictors of sludge vulnerability and acceptability, and on the types of sludge to study.
What makes people vulnerable to sludge?
Socio-economic disadvantage
A large body of research on administrative burdens and sludge shows that frictions have distributive effects and create higher barriers for groups in society who suffer from poverty, a lack of education, or health problems. Members of such groups are often exposed to processes involving sludge, such as social safety nets, and their lives would also be severely affected if sludge stops them from completing such processes (Heinrich, Reference Heinrich2016; Christensen et al., Reference Christensen, Aarøe, Baekgaard, Herd and Moynihan2020; Sunstein, Reference Sunstein2020; Chudnovsky and Peeters, Reference Chudnovsky and Peeters2021; Döring and Madsen, Reference Döring and Madsen2022; Halling and Baekgaard, Reference Halling and Baekgaard2024; Martin, Reference Martin2024; Martin et al., Reference Martin, Delaney and Doyle2024). Of course, it is also true that entrepreneurs, large and small, might be affected by sludge when they are seeking to obtain permits or licenses; sludge can operate as a severe barrier to economic activity and have special and immediate adverse effects on people who are well off, potentially with adverse effects for large numbers of other people.
Groups who face high levels of sludge when interacting with public policy processes include not only entrepreneurs but also those with low incomes or financial well-being (Bhargava and Manoli, Reference Bhargava and Manoli2015; Holt and Vinopal, Reference Holt and Vinopal2023; Martin et al., Reference Martin, Delaney and Doyle2024), those in poor physical or mental health (Deshpande and Li, Reference Deshpande and Li2019; Arulsamy and Delaney, Reference Arulsamy and Delaney2022; Bell et al., Reference Bell, Christensen, Herd and Moynihan2023; Martin et al., Reference Martin, Delaney and Doyle2024), older people (Moynihan et al., Reference Moynihan, Herd and Harvey2015), ethnic or racial minorities (Ray et al., Reference Ray, Herd and Moynihan2023; Bell and Jilke, Reference Bell and Jilke2024), and women and gender minorities (Martin, Reference Martin2024; Herd and Moynihan, Reference Herd and Moynihan2025).
There are fewer studies on the distributive effects of private sludge, but evidence suggests that older people, those with health issues and those on low incomes may pay a higher ‘loyalty penalty’ in essential markets (Citizens Advice, 2018), which could be due to higher frictions in switching deals. Evidence also suggests that vulnerable consumers choose worse deals and face higher frictions in markets such as mobile phones and health insurance (Jilke, Reference Jilke2015; Domurat et al., Reference Domurat, Menashe and Yin2021).
Cognitive resources and scarcity
A potential mechanism contributing to the distributive effects of sludge is cognitive resources. Christensen et al. (Reference Christensen, Aarøe, Baekgaard, Herd and Moynihan2020) argue that those in poor health, experiencing age-related cognitive decline, or living in financial scarcity have constrained cognitive resources as the challenges they face take up most of these resources, which negatively affects these groups’ executive functions and therefore their ability to overcome sludge and complete important processes such as obtaining access government benefits (in the same vein, Mullainathan and Shafir, Reference Mullainathan and Shafir2013). In sum, people who most need government help may find it disproportionately hard to receive it, because of having fewer cognitive resources available.
Looking beyond distributive aspects and at the broader population, scarcity, understood as not having enough in the way of ‘bandwidth’, can influence decision-making (Mullainathan and Shafir, Reference Mullainathan and Shafir2013) and may therefore increase sludge vulnerability. Building on Christensen et al.’s (Reference Christensen, Aarøe, Baekgaard, Herd and Moynihan2020) argument that constrained cognitive resources impede ability to navigate sludge, it may be that vulnerability also increases when people do not have enough mental energy or time to navigate processes involving sludge, processes that by definition are energy- or time-consuming (e.g. administrative burdens have been called a ‘time tax’, Lowrey, Reference Lowrey2021; OIRA, 2024). It follows that scarcity may increase sludge vulnerability, but little evidence is available.
Biases and heuristics
Biased decision-making may increase sludge vulnerability across the population. When faced with complex or effortful decisions, people often rely on heuristics (mental ‘rules of thumb’) to simplify these decisions, opening the door to systematic biases (e.g. Kahneman et al., Reference Kahneman, Knetsch and Thaler1991). This may especially be the case in contexts involving sludge, because their features tend to make decisions more complex: sludge often involves unfamiliar material (e.g. an online software only used in that process), uncertainty (e.g. the rest of the process or its outcome is unknown) and a lack of frequent use (e.g. a once-every-few-years voter registration process) (Shahab and Lades, Reference Shahab and Lades2024).
An example of a relevant bias is the status quo bias (Madrian and Shea, Reference Madrian and Shea2001), which may discourage people from taking on effortful tasks that require them to deviate from the default such as cancelling a subscription. Another example is present bias, which can lead to a focus on the short-term and procrastination. Present-biased people might be discouraged from taking costly actions today (such as researching and implementing a process) although these actions would benefit them in the long run (O’Donoghue and Rabin, Reference O’Donoghue and Rabin1999; Lades et al., Reference Lades, Clinch and Kelly2021). People’s limited attention (Gabaix, Reference Gabaix2019) and their over-optimism that they can overcome frictions (Tasoff and Letzler, Reference Tasoff and Letzler2014) are likely to make sludge even more harmful. At the same time, there is a lack of research testing specifically whether behavioural biases affect vulnerability to sludge.
Administrative literacy
A small, emerging body of evidence finds (unsurprisingly) that ‘administrative literacy’, i.e. the ability to ‘obtain, process and understand basic information and services from public organisations needed to make appropriate decisions’ (Döring, Reference Döring2021), makes people less vulnerable to sludge. For example, Döring and Madsen (Reference Döring and Madsen2022) find that having high levels of administrative literacy helps reduce psychological costs when encountering administrative burdens, while Masood and Nisar (Masood and Azfar Nisar, Reference Masood and Azfar Nisar2021) find that people’s ‘administrative capital’, i.e. their ‘understanding of bureaucratic rules, processes and behaviours’, helps them navigate administrative burdens. Administrative literacy is a relatively new concept (although it is linked to the broader concept of human capital, discussed in Christensen et al., Reference Christensen, Aarøe, Baekgaard, Herd and Moynihan2020, and to other concepts such as health literacy, as discussed in Döring, Reference Döring2021). As a result, relatively little evidence is yet available, in particular for private sludge contexts (most of the evidence is on administrative burdens).
What makes sludge acceptable?
Administrative burden tolerance
There is little evidence on what features of sludge make it more or less acceptable to people. The closest evidence comes from emerging research on ‘burden tolerance’ (Baekgaard et al., Reference Baekgaard, Halling and Moynihan2025), which measures how people judge the acceptability of administrative burdens. A high burden tolerance means a strong belief that administrative burdens are justified. Halling and Baekgaard’s (Reference Halling and Baekgaard2024) systematic review of the administrative burden literature finds that factors shaping people’s burden tolerance include political ideological beliefs, as conservatism is associated with higher burden tolerance than liberalism (Baekgaard et al., Reference Baekgaard, Moynihan and Thomsen2021; Bell et al., Reference Bell, Ter-Mkrtchyan, Wehde and Smith2021; Haeder et al., Reference Haeder, Sylvester and Callaghan2021; Halling et al., Reference Halling, Herd and Moynihan2023; Haeder and Moynihan, Reference Haeder and Moynihan2025). In addition, burden tolerance, on the part of people who are observers, is affected by how such people perceive the group facing the administrative burden, for example, whether that group is a minority facing racism, racial resentment, or seen as ‘deserving’ (Baekgaard et al., Reference Baekgaard, Moynihan and Thomsen2021; Haeder et al., Reference Haeder, Sylvester and Callaghan2021; Barnes, Reference Barnes2023; Haeder and Moynihan, Reference Haeder and Moynihan2025). Providing people with a justification for burdens may also increase tolerance (Halling and Van de Walle, Reference Halling and Van de Walle2025).
In addition to beliefs, people’s personal characteristics also influence their burden tolerance. Baekgaard, Halling, & Moynihan’s (Halling and Van de Walle, Reference Halling and Van de Walle2025) survey of seven countries finds higher tolerance on the part of those who are males, young adults, less well educated and with good health. Low administrative capital makes people less tolerant of burdens (Haeder and Moynihan, Reference Haeder and Moynihan2025), while higher income is associated with higher burden tolerance (Halling et al., Reference Halling, Herd and Moynihan2023). People with personal experience using benefits have lower tolerance (Baekgaard et al., Reference Baekgaard, Moynihan and Thomsen2021; Halling et al., Reference Halling, Herd and Moynihan2023). There is also evidence that personality traits (such as conscientiousness and openness) correlate with tolerance (Aarøe et al., Reference Aarøe, Baekgaard, Christensen and Moynihan2021).
Learning from nudge acceptability research
Research on the acceptability of nudges may also help illuminate the acceptability of sludge. Nudges can be defined in various ways, but they are often taken to be changes in the choice architecture that promote people’s welfare while maintaining freedom of choice and without meaningfully changing incentives (Thaler and Sunstein, Reference Thaler and Sunstein2021). Nudges have sometimes been compared to sludge (Thaler, Reference Thaler2018; Mills, Reference Mills2023); many nudges simplify people’s ability to receive benefits or to avoid burdens, whereas sludge has the opposite effect. Factors that influence the acceptability of a nudge include its alignment or nonalignment with people’s values and interests, its policy domain, its perceived intrusiveness in people’s decision-making, its perceived effectiveness, whether it encourages a desirable behaviour or discourages a harmful one, whether people are already motivated to engage in the target behaviour, and (with mixed results depending on the study) whether people are already engaging in the target behaviour (Reynolds et al., Reference Reynolds, Archer, Pilling, Kenny, Hollands and Marteau2019; Yan and Yates, Reference Yan and Yates2019; Gestel et al., Reference Gestel, Adriaanse and Ridder2021; Lemken et al., Reference Lemken, Wahnschafft and Eggers2023; Vugts et al., Reference Vugts, van den Heuvel and Havermans2024; Rafaï et al., Reference Rafaï, Ribaillier and Jullien2025).
Findings from research on the acceptability of nudges should be used with caution in the context of sludge. They are not the same thing. Our study includes a range of public and private sector sludge scenarios to test what types of sludge are seen as acceptable.
Does the type of sludge affect vulnerability and acceptability?
There is a lack of literature exploring individuals’ vulnerability to different types of sludge. So far, we do not know whether people are more vulnerable to, and more accepting of, certain types of sludge, despite the fact that behavioural frictions can take many different forms, from complex paperwork to delays or even unpleasant interactions. This is a central contribution of our study. Moreover, the effects of different types of sludge are likely to differ across the population. While there is strong evidence showing the distributive effects of sludge, less is known about whether different groups might be more or less vulnerable to specific types of sludge. Some people might dread the thought of having to call a public service by phone whereas others might be particularly negatively affected by excessive paperwork requirements and complex language (Sunstein, Reference Sunstein2020). Likewise, the acceptability of sludge may depend on the specific features of sludge, and not only on people’s own characteristics or general beliefs.
More specifically, our study builds on existing typologies of sludge and administrative burden to test how vulnerability and acceptability vary across different types of sludge. We use the search, evaluation, implementation and psychological costs framework (Shahab and Lades, Reference Shahab and Lades2024) to classify different types of sludge in our study. This framework roughly maps onto the learning, compliance and psychological costs framework used in administrative burden research (Moynihan et al., Reference Moynihan, Herd and Harvey2015; Herd and Moynihan, Reference Herd and Moynihan2019) and accommodates both private and public sludge (see beginning of section 2 for detailed descriptions of frameworks). Classifying administrative burdens (in our case, sludge) into a cost typology is especially helpful because it makes it easier to identify and assess different types of administrative burdens (Madsen et al., Reference Madsen, Mikkelsen and Moynihan2022). However, we do not currently know how these different types compare in terms of vulnerability and acceptability. Our study helps answer this question.
Another categorisation of sludge potentially relevant for vulnerability and acceptability is whether the sludge is occurring in a public sector context (e.g. filling out a form online to obtain to a government service) or in a private sector context (e.g. filling out a form online to obtain access to a service from a private provider). Previous research suggests that there is an anti-public sector bias against public providers providing similar services to private actors (Marvel, Reference Marvel2016). The difference between the public and the private sector matters because sludge vulnerability and acceptance may depend on context. For example, sludge in a means test to access a taxpayer-funded benefit may be seen as more acceptable, or on the contrary to create higher costs and therefore vulnerability (e.g. due to added stigma), compared to sludge in a private context where paying consumers are attempting to cancel a service. In other words, (some) public services may carry higher stakes than (some) private ones.
Moreover, public services frequently operate as monopolies, whereas private services usually face competitive pressures. It follows that users of public services lack straightforward exit options, which can lead to more frequent and more negative experiences due to selection effects. Awareness of this lack of choice may further amplify the perceived burden of sludge in public-sector settings.
Most empirical research focuses on either public contexts (e.g. administrative burden research, by definition) or private contexts (e.g. most research explicitly studying sludge, see Mills et al., Reference Mills, Whittle, Ahmed, Walsh and Wessel2023; Xiao et al., Reference Xiao, Henderson, Yang and Newall2024; there are also public sludge studies, see Grieder et al., Reference Grieder, Kistler and Schmitz2024). At the same time, there is very little research that directly compares these two contexts. Martin et al. (Reference Martin, Delaney and Doyle2023) test the effect of public sector sludge (in a benefit scenario) and private sector sludge (in a bill scenario) on decision-making and find that sludge affects decision-making in both contexts, but that the effect is higher for public sector sludge. While this result may well be partly due to the design of the scenarios, it is consistent with the view that people experience sludge differently in the public and private contexts, with people being more vulnerable to, and less accepting of, public sector sludge, perhaps because of their dependence on basic services offered by the public sector.
Methodology
Procedure
We developed an online survey to understand self-reported vulnerability to, and acceptability of, different types of sludge. The survey presented participants with 10 different types of administrative procedures required to complete ‘a task’. We did not specify what this hypothetical task was but asked participants to think of ‘the task’ as something they would like to complete (e.g. obtaining a product, a service, information needed or another benefit) but something that is not absolutely essential. We clarified that our interest was in participants’ assessment of the different administrative procedures, rather than their assessment of different tasks.
This approach allowed us to measure self-reported vulnerability to typical sludge that people encounter. However, it did not capture whether some people are inherently more likely to encounter sludge than others. Indeed, it is plausible (even likely) that some people experience sludge more frequently than others and therefore suffer more seriously from it. Our survey, by design, focused only on how vulnerable people believe they would feel when confronted with different types of sludge, without requiring that people had experienced this type of sludge in the real-world.
We assigned participants randomly to one of two conditions: for half of the participants, the administrative procedures were framed as originating from a government agency, and for the other half they were framed as originating from a private company. For each administrative procedure, we asked participants four questions to measure how vulnerable they believed themselves to be to this procedure and two questions measuring how acceptable they found the procedure. Explaining the variation of the answers to these questions across individuals and types of sludge is the main aim of this study.
Participants then answered questions regarding various personal characteristics, such as their administrative literacy, their tendency to procrastinate, whether they felt that they did not have enough time, mental energy and money to deal with important things in their lives, their physical and mental health, whether they had specific disabilities, trust in the Irish Government, age, gender identity, ethnic background, education, marital status, employment status, where they live, how many people live in their household and income. Finally, we asked participants to report how many of 27 administrative tasks they had attempted (successfully or unsuccessfully) in the last three months. Most questions required participants to provide an answer to reduce missing values. Given that this is data from a survey, all information is self-reported.
The analysis plan for this study was pre-registered on the website of the Open Science Framework (https://osf.io/zpvs2) and received ethical approval from the Human Research Ethics Committee – Sciences at the University of Stirling (Research Ethics Reference Number: LS-LR-24-258-Lades). The full survey instrument is available in the Supplementary Information.
Participants
We recruited 1,785 participants through the panel provider Dynata in October 2024, with the aim of obtaining a sample representative of the Irish population in terms of age, gender, and geographic distribution. As in many other countries, there is much anecdotal evidence about people in Ireland complaining about administrative procedures (Lades et al., Reference Lades, Johnson and O’Connor2022). There is also evidence for sludge in contexts such as grants for landowners to plant trees (Lades et al., Reference Lades, Lentz and Augustenborg2025). In addition, evidence about the sludge experienced by low-income households within the cost-of-living crisis (Ceallaigh et al., Reference Ceallaigh, Martin, Timmons, Robertson and Lunn2025) is emerging. While there is no systematic evidence on problems related to sludge and administrative burden in Ireland and comparing those problems to those to other countries, various types of sludge indeed are prevalent in both the public and private sector in Ireland, making Ireland a relevant case for comparing vulnerability and acceptability of different kinds of sludge in both the public and private sector.
All participants received baseline monetary compensation through the panel provider. We dropped 8 observations because they had duplicate study IDs and 186 observations because they failed an attention check in which we asked participants to select ‘Agree’ as an answer. The final sample for our analysis thus consists of 1,591 participants of whom 57.1% are females. The mean age is 44.3 and 48.7% have an annual net household income of €50,000 or less. Further summary statistics of the sample are presented in Table 1. Table SI.1 in the Supplementary Information shows that there are no statistically significant differences in observables across the two experimental treatments (public versus private sludge). Table SI.2 in the Supplementary Information suggests that our dataset is not fully representative of the Irish population. However, the comparison is imperfect because our survey included only participants aged 18 and older, whereas population census data for Ireland are available only for individuals 15 and older (see Table SI.2 for more details). Nevertheless, the comparison suggests that our survey includes a higher proportion of females, participants with higher educational attainment and those in employment. A potential explanation is that the survey provider sampled additional participants to meet certain quotas, and we did not see a reason to use only a subset of the data.
Table 1. Descriptive statistics
Measures
Ten types of sludge
We presented participants with the 10 types of sludge in a random order using the wording described in Table 2. To enable us to compare sludge in the public sector with sludge in the private sector, half of the participants (randomized) saw the descriptions in the table with sludge emerging from interactions with ‘a private company’, while the other half saw the sludge emerging from interactions with ‘a government agency’. This randomisation allowed us to test whether vulnerability to and acceptability of sludge depend on the sector in which the sludge originates. To select the 10 types of sludge, we started from the list of sludge presented in the New South Wales Government’s (2024) Sludge Audit Method Guide. We then simplified the list and reduced it to ten types of sludge to avoid survey fatigue and included types of sludge we anticipated would vary in how people view them. We made sure that each type of sludge could originate from both the private and the public sector to allow clean randomisation.
Table 2. Sludge types in the private sector. Half of the participants saw these types of sludge. The other half saw sludge from the public sector. The tasks are sorted according to vulnerability (private and public sector combined)
Self-reported vulnerability to sludge
To measure how vulnerable participants believed themselves to be to sludge, we asked participants whether they disagreed or agreed with four statements: ‘It would be difficult for me to carry out this task’, ‘This task would be a hassle to me’, ‘I would get frustrated when carrying out this task’ and ‘There is a high chance that I would not complete this task’. Participants had to respond on a 5-point Likert scale from ‘Strongly disagree 1’ to ‘Strongly agree 5’. The reliability of this measure was high, with Cronbach’s alpha values for the ten sludge types ranging from 0.89 to 0.94. The mean vulnerability score for each type of sludge is shown in Table 2.
Acceptability of sludge
To measure how acceptable participants find each case of sludge, we asked participants whether they disagreed or agreed with two statements: ‘The procedure to complete this task is acceptable’ and ‘The effort needed to complete this task is fair’. Participants had to respond on a 5-point Likert scale from ‘Strongly disagree 1’ to ‘Strongly agree 5’. The reliability of this measure was high, with Cronbach’s alpha values for the ten sludge types ranging from 0.87 to 0.90. The mean acceptability score for each type of sludge is shown in Table 2.
Administrative literacy
To assess participants’ self-perceived administrative literacy, we adapted an existing measure for administrative literacy (Döring, Reference Döring2021; Döring and Madsen, Reference Döring and Madsen2022) and asked participants whether they disagreed or agreed with the following three statements: ‘I usually understand the content of administrative letters, documents, and emails I receive from the government or from private companies’, ‘I usually understand the rules, requirements, or contract details that apply to me as a citizen or as a customer’ and ‘I keep myself informed on any changes in regulation or contract details’. Participants had to respond on a 5-point Likert scale from ‘Strongly disagree 1’ to ‘Strongly agree 5’. Reliability was acceptable and Cronbach’s alpha was 0.74.
Tendency to procrastinate
To measure participants’ self-assessed tendency to procrastinate, we adapted the measure from Lillemo (Reference Lillemo2014) as also used by Mogensen and Thøgersen (Reference Mogensen and Thøgersen2024). We asked participants whether they disagreed or agreed with the following four statements: ‘I often put things off for so long that it has negative consequences’, ‘When I must start a task, I often end up doing something else’, ‘I often postpone things to do until tomorrow’ and ‘I often hurry to finish things at the last minute’. Participants had to respond on a 5-point Likert scale from ‘Strongly disagree 1’ to ‘Strongly agree 5’. The measure demonstrated good reliability, with Cronbach’s alpha = 0.87.
Scarcity
To assess perceptions of scarcity, participants responded to three statements to measure perceived time, mental energy and financial constraints. We asked participants how often the following statements applied to them or their situation: ‘I cannot deal with important things properly due to a lack of time’ to measure time scarcity, ‘I cannot deal with important things properly due to a lack of mental energy’ to measure mental energy scarcity and ‘I cannot deal with important things properly because I am busy trying to keep on top of my financial situation’ to measure financial scarcity. Participants had to respond on a 5-point Likert scale from ‘Very rarely’ to ‘Very often’.
Physical and mental health
We asked participants about their physical and mental health using ‘How is your physical health in general?’ and ‘How is your mental health in general?’ with a 5-point Likert scale from ‘Very bad’ to ‘Very good’ as the answer options. We also asked participants whether they identified as having any disabilities offering a list of five disabilities and an ‘Other’ option.
Correlations of main variables
To get a first impression of the data, Table 3 presents the correlations across the main, non-demographic variables we analyse. For this table, we calculated the average self-reported vulnerability score and the average acceptability score for each individual across all 10 types of sludge. We see a relatively high and significantly negative correlation between our two dependent variables self-reported vulnerability and acceptability (−0.693*).
Table 3. Pairwise correlations of the main variables
* Note:p < 0.05.
Analytic procedure
As pre-registered, we used multilevel regression models to predict self-reported vulnerability to sludge type s of individual i by a set of independent variables as described below. To account for repeated measures, where the same participant assessed their vulnerability to sludge ten times, we clustered the standard errors at the individual level. We pre-registered the following model:
\begin{align}
Vu{l_{i,s}} & = {\beta _0} + {\beta _1}\,SludgeType + {\beta _2}\,AdminLi{t_i} + {\beta _3}\,Procras{t_i} + {\beta _4}\,ScarcityTim{e_i}\nonumber\\
&\quad + {\beta _5}\,ScarcityEnerg{y_i} + {\beta _6}\,ScarcityMone{y_i} + {{\boldsymbol{X}}_{\boldsymbol{i}}}{\boldsymbol{\beta}} + {v_s} + {\in _{i,s}},
\end{align} where 
$Vu{l_{i,s}}$ is the vulnerability score of participant i for sludge type s.Footnote 1 The other variables indicate the measures as indicated by their names. The vector 
${{\boldsymbol{X}}_{\boldsymbol{i}}}$ represents the demographic control variables: physical health, mental health, age, female dummy, ethnicity, education, civil status, employment status, location (city, suburb, rural or other), number of people living in the household and income. The 
$\beta $ coefficients represent the associations between the independent variables and the self-assessed vulnerability to sludge and should not be interpreted causally. The error terms 
${v_s}$ and 
${\in _{i,s}}$ represent the structure of the data with error terms on the sludge type level and the individual-level, respectively.
Since the causal relationships between the psychological predictors are uncertain (e.g. people who procrastinate might have less time or vice versa), the model in equation (1) might potentially confound estimates. Hence, we also present the results when adding one psychological variable at a time without controlling for the other psychological variables in an exploratory analysis. To further explore whether the associations between independent variables and vulnerability differed across sludge types, we included interaction terms between 
$SludgeType$ and the respective independent variable based on the regressions that do not control for the other psychological variables. All other components of the model remained identical.
We then repeat the same analysis steps with a different dependent variable: acceptability of sludge. The only difference in this analysis is that we also include a dummy indicating whether the sludge originates in the public or the private sector. Other than that, the analysis is as described in the previous paragraph.
Results 1: self-reported vulnerability to sludge
Types of sludge predicting self-reported vulnerability
To test whether people believe themselves to be more vulnerable to some sludge than to others, we regressed reported vulnerability to sludge on the type of sludge (either originating in the private or public sector) without controlling for any other variables. The results of this regression align with the mean vulnerability scores presented in Table 2 and are illustrated in Figure 1.
Figure 1. Vulnerability to sludge predicted by type of sludge only and no other control variables based on multi-linear regressions with 1,591 individuals and 15,910 observations, depending on whether the sludge emerges from either the private or the public sector.
Among the 10 types of sludge, outdated websites with partly broken links were reported as causing the highest levels of vulnerability (M = 3.809, SE = 0.039 for private sector sludge and M = 3.833, SE = 0.038 for public sector sludge). Participants were least negatively affected by a requirement to report one’s religion and sexual orientation (M = 2.493, SE = 0.046 for private sector sludge and M = 2.319, SE = 0.042 for public sector sludge). While this is not a preregistered analysis, we disaggregate the data by whether the sludge emerges from interactions with either the private or the public sector as this is of particular interest to public administration researchers. Regressing self-reported vulnerability on the origin of the sludge shows no significant differences in the overall mean vulnerability score between public and private sector sludge (b = 0.037, p = 0.354).
Additional regressions for each type of sludge individually show some significant but small differences. People reportbeing more vulnerable to public than to private sector sludge for unfriendly contacts (b = 0.118, p = 0.043), websites with too much information (p = 0.117, p = 0.036), when having to attend a meeting at a specific time (b = 0.157, p = 0.011), and when having to send a form in the post (b = 0.158, p = 0.010). When having to state one’s religion and sexual orientation, people report being more vulnerable when asked by the private sector (b = −0.174, p = 0.005). In the following, and consistent with our pre-registration, we do not include the public/private variable in the regressions.
Socio-economic disadvantage predicting self-reported vulnerability
We then predicted self-reported vulnerability to sludge by the variables representing potential socio-economic disadvantages. As illustrated in Figure 2, mental health (b = −0.111, p < 0.001) and physical health (b = −0.061, p = 0.026) are negatively associated with self-reported vulnerability to sludge. To illustrate the strength of this coefficient, based on this regression somebody who reports having ‘1 = Very bad’ mental health reports an estimated value of vulnerability to sludge of 3.42 (SE = 0.062) and somebody who reports having ‘5 = Very good’ mental health reports an estimated value of vulnerability to sludge of 2.98 (SE = 0.039).
Figure 2. Variables representing socio-economic disadvantage as predictors of vulnerability to sludge based on multi-linear regressions with 1,591 individuals and 15,910 observations.
We also find a positive association between age (b = 0.049, p = 0.018) and self-reported vulnerability to sludge. Moreover, participants with lower secondary school education or less report being less vulnerable to sludge compared to those with higher education and especially undergraduate and postgraduate degrees. Participants living in a rural village or remote area reportbeing more vulnerable to sludge compared to participants living in a city. The Supplementary Information includes the regression Table SI.3, presenting these results with all coefficients and standard errors.
Behavioural predictors of vulnerability
In the next step, we added the five ‘behavioural’ independent variables to the regression. Figure 3 presents the coefficients of these behavioural variables. The figure presents the associations between the variables and self-reported vulnerability to sludge when adding all five variables to the same regression with the blue circles. This is the pre-registered analyses. The red squares in Figure 3 represent the same associations when including only one of the five variables at a time (i.e. when not controlling for the other four behavioural variables). This analysis shows that all effects are stronger when not controlling for the other variables and even financial scarcity becomes significant, suggesting that these five variables are linked to each other. Since it is a priori not known how the behavioural variables are causally linked to each other, the results illustrated by the red squares come from our preferred (albeit not preregistered) model.
Figure 3. Behavioural predictors of self-reported vulnerability to sludge from multi-level linear regressions with 1,591 individuals and 15,910 observations controlling for demographics and mental and physical health. The blue circles illustrate the results of one single regression with all five behavioural parameters included. The red squares illustrate the results of five regressions that include only the respective behavioural predictor (i.e. without controlling for the other four behavioural parameters).
We find several significant predictors of self-reported vulnerability to sludge. The next paragraphs rely on the preregistered models that include all controls (the blue circles in Figure 3). Administrative literacy is negatively associated with self-reported vulnerability to sludge (b = −0.192, p < 0.001) and the tendency to procrastinate is positively associated with self-reported vulnerability to sludge (b = 0.057, p < 0.001). Moreover, time scarcity (b = 0.065, p = 0.004) and mental energy scarcity (b = 0.101, p < 0.001) are positively associated with self-reported vulnerability to sludge.
In contrast, the association between financial scarcity and self-reported vulnerability to sludge is not significant (b = 0.030, p = 0.186). We also note that in these regressions, neither physical health (b = −0.033, p = 0.198) nor mental health (b = 0.004, p = 0.850) is a significant predictor of self-reported vulnerability to sludge, suggesting that the added psychological parameters might be a pathway that explain the previously significant associations between the physical and mental health variables and self-reported vulnerability to sludge. Table SI.3 in the Supplementary Information includes these results with all details.
Interactions between behavioural variables and type of sludge predicting self-reported vulnerability
We also tested whether the associations between the five behavioural variables (administrative literacy, procrastination, time scarcity, mental energy scarcity and financial scarcity) and self-reported sludge vulnerability differed across different types of sludge. In this section, we highlight one interaction that we consider most interesting and the other four interaction diagrams are presented in the Supplementary Information (in Figure SI.1). Given that the previous section suggests that there are some relationships between the five behavioural variables and we do not know how they are causally linked, we test for interaction effects without controlling for the other four behavioural variables.
As shown in Figure 4, administrative literacy plays a significant role across most types of sludge. However, its importance is particularly pronounced when dealing with documents that are written in complex language and use a lot of jargon. In contrast, administrative literacy does not matter for dealing with requests to provide personal information such as religious beliefs and sexual orientation.
Figure 4. Interaction effects predicting self-reported vulnerability to sludge by the interaction between administrative literacy and sludge types based on multi-linear regressions with 1,591 individuals and 15,910 observations.
Results 2: acceptability of sludge
What types of sludge are less accepted than others?
To test whether people accept some sludge more than others, we regressed acceptability of sludge on the type of sludge without controlling for any other variables. The results of this regression align with the mean acceptability scores presented in Table 2 and are illustrated in Figure 5. For example, among the 10 types of sludge, outdated websites with partly broken links are least acceptable (M = 2.415, SE = 0.429 for private sector sludge and M = 2.364, SE = 0.430 for public sector sludge) and having to print and send a form via post is most acceptable (M = 3.437, SE = 0.0438 for private sector sludge and M = 3.261, SE = 0.0415 for public sector sludge). Figure 5 shows sludge acceptability for private and public sludge separately. This analysis shows a small difference in the overall mean acceptability of sludge between public and private sector sludge (b = −0.074, p = 0.070), and only minor variations for individual types of sludge. Regressions for each type of sludge individually show that sludge in the public sector is less acceptable when a form needs to be printed and sent in the post (b = −0.177, p = 0.003), when a process is uncertain (b = −0.157, p = 0.009), and especially when and contacts are unfriendly (b = −0.260, p < 0.001).
Figure 5. Acceptability of sludge predicted by type of sludge only and no other control variables based on multi-linear regressions with 1,591 individuals and 15,910 observations.
Socio-economic disadvantage predicting acceptability of sludge
Again, we predicted acceptability of sludge by the variables representing potential socio-economic disadvantages. As illustrated in Figure 6, mental health (b = −0.091, p < 0.001) and physical health (b = −0.081, p = 0.005) are negatively associated with sludge acceptability. Somebody who reports having ‘1 = Very bad’ mental health has an estimated value of sludge acceptability of 2.66 (SE = 0.067) and somebody who reports having ‘5 = Very good’ mental health has an estimated value of sludge acceptability of 3.02 (SE = 0.040).
Figure 6. Socio-economic disadvantage predictors of acceptability of sludge based on multi-linear regressions with 1,591 individuals and 15,910 observations.
We also find a negative association between age in decades (b = −0.063, p = 0.003) and sludge acceptability. Moreover, more educated participants report substantially lower sludge acceptability, as do Irish compared to non-Irish participants. Participants living in a rural village or remote area or in a suburb report lower sludge acceptability compared to participants living in a city. The Supplementary Information includes the regression Table SI.4, presenting these results with all coefficients and standard errors. We also see a negative association between income and acceptability of sludge where those who are richer are less accepting.
Behavioural predictors of sludge acceptability
We tested whether the five behavioural variables predicted acceptability. Figure 7 presents the coefficients of these psychological variables again when adding all five behavioural variables in the same regression and when running five separate regressions (i.e. when not controlling for the other four behavioural variables). Again, all effects are stronger when not controlling for the other behavioural variables.
Figure 7. Behavioural predictors of sludge acceptability from multi-level linear regressions with 1,591 individuals and 15,910 observations controlling for demographics and mental and physical health. The blue circles illustrate the results of one single regression with all five behavioural parameters included. The red squares illustrate the results of five regressions that include only the respective behavioural predictor (i.e. without controlling for the other four behavioural parameters).
We find two significant predictors of sludge acceptability in the more conservative model that is illustrated by the blue circles. Administrative literacy is positively associated with sludge acceptability (b = 0.267, p < 0.001) and mental energy scarcity is negatively associated with acceptability of sludge (b = −0.073, p = 0.003). When not controlling for the other four behavioural variables, the effects become stronger and the tendency to procrastinate and time scarcity also become statistically significant.
Interactions between psychological predictors and sludge acceptability
Finally, we tested whether the associations between sludge acceptability and the five behavioural variables differed across different types of sludge. Again, we did not control for the other four behavioural variables in these models. For consistency with the previous analysis, Figure 8 presents the associations between administrative literacy and sludge acceptability by sludge type. The figure shows that sludge acceptability is generally higher for people with higher administrative literacy, and that is the case for all 10 types of sludge. We find the strongest effect of administrative literacy for documents that are written in complex language and jargon; that is particularly unacceptable for participants with lower administrative literacy.
Figure 8. Interaction effects predicting sludge acceptability by the interaction between administrative literacy and sludge types based on multi-linear regressions with 1,591 individuals and 15,910 observations.
Discussion
Key results
We summarise four main results. First, some types of sludge are more problematic than others. Participants indicated that they are particularly vulnerable (1) to websites that are outdated and include broken links, (2) when employees are unfriendly, (3) when documents are written in complex language and use a lot of jargon and (4) when websites are full of information and hard to navigate. These types of sludge were also judged to be among the least acceptable. Using an existing typology (Shahab and Lades, Reference Shahab and Lades2024), these four types of sludge can be classified as leading to search costs, evaluation costs, or psychological costs (see Table 1). We note that the six least problematic types of sludge are best classified as leading to implementation costs.
Still, it would be premature to claim that sludge that leads to implementation costs is less problematic than sludge leading to the other types of costs. Magnitude matters; very high implementation costs might well be more troublesome than low costs of other kinds. In any case, we evaluated only a small subset of 10 sludge types, and there are many other types of sludge that can lead to implementation costs that might be highly problematic. In addition, implementation might appear straightforward when thinking about it while doing a survey; in real life, implementation costs might be far more problematic than anticipated. One reason for the difficulty of anticipating how vulnerable one might be to implementation costs is the empathy gap that can make it hard for people to see the real-world effects of sludge (Soman et al., Reference Soman, Cowen, Kannan and Feng2019).
Second, we tested whether self-reported sludge vulnerability and acceptability differ depending on whether the sludge is emerging from a government agency (the public sector) or a company (the private sector). Since we randomised participants into a private or a public sector treatment group, these results can be interpreted causally. The patterns in the data suggest that there are only minor differences in sludge vulnerability and acceptability depending on whether the sludge emerges from the private or the public sector. We had predicted this for self-reported sludge vulnerability, but we were surprised to see such small effects also for sludge acceptability. Given the nature of our survey (self-reported, hypothetical tasks that people have not actually experienced and not incentivised), these results should be viewed with caution; the comparison of private versus public sector sludge is an interesting area for future work (see below).
Third, we found that some variables describing facets of socio-economic disadvantage do predict sludge vulnerability and acceptability. Significantly, participants with relatively low mental health indicated being more vulnerable to and less accepting of sludge. We also note that mental health was not a significant predictor of sludge vulnerability and acceptability when including the five behavioural variables in the same regression. In our interpretation, which is in line with the results presented in the regression tables in the Supplementary Information (Tables SI.3 and SI.4), it is likely that low mental health leads to mental energy scarcity, which in turn leads to more sludge vulnerability and less sludge acceptability. However, we are hesitant to conduct formal mediation analyses as these might give the impression that our data can be interpreted causally. Instead, we believe that future experimental work should test these effects using study designs that are able to show causal relationships.
We had expected more and stronger associations between demographic variables and sludge vulnerability. One reason for these relatively weak associations could be our focus on sludge vulnerability and acceptability conditional on experiencing some sludge. It is likely that vulnerable groups do experience more sludge and hence still are harmed by sludge more than less vulnerable groups. Our results merely suggest that vulnerable groups, at large, do not consider themselves as being more vulnerable to sludge than other groups once they must engage with the sludge. It might also be the case that repeatedly experiencing administrative frictions creates a learning effect and thus reduces the vulnerability to sludge. These are open questions for future longitudinal research. Given the ambiguity of the possible interpretations, we prefer not interpreting the correlations between demographics and sludge assessments at face value.
Fourth, and finally, the behavioural variables (administrative literacy, procrastination, time scarcity, mental energy scarcity and financial scarcity) are quite strongly associated with sludge vulnerability and acceptability. This highlights the importance of analysing these behavioural factors when dealing with the inequality that sludge can bring about. For example, the strong negative associations between administrative literacy and vulnerability to sludge suggests that administrative literacy can act as a buffer from sludge. This is in line with previous research on administrative literacy and administrative burden, which suggests that people with high levels of administrative literacy cope better with the imposed frictions (Döring, Reference Döring2021; Döring and Madsen, Reference Döring and Madsen2022).
Limitations and future work
Our study has limitations that suggest important directions for future research. First, this is a cross-sectional data set, and we can make only limited claims about causality (in particular for the comparison between private versus public sector sludge). There might be unobserved confounding factors that explain some of the other associations we have found. Future work should rely more on experimental designs that manipulate important determinants of vulnerability to and acceptability of sludge to measure causal effects. For example, future research should experimentally vary time scarcity or the level and type of sludge in a task and measure how well different individuals experience and complete the task.
Second, our measures of vulnerability to sludge and acceptability of sludge are based on self-reported, non-incentivised survey responses. This methodological choice might explain some of our findings. For example, our two dependent variables (vulnerability and acceptability) are strongly (negatively) correlated, which is theoretically plausible but could also reflect common-method bias or a desire for consistency. Future research should therefore complement self-reports with behavioural measures, such as performance in real-effort tasks, to capture non-hypothetical vulnerability to sludge.
A third limitation regards external validity. In the real world, participants are never asked to evaluate a process independent of context. We asked participants to assume that they would like to complete a ‘task’ without further specifying this task, and participants might view this as a highly hypothetical question. For that reason, our results are indicative of people’s ‘beliefs’ about their vulnerability, and these beliefs might not necessarily align with objective measures of vulnerability. Incorrect assessments of how vulnerable people report themselves to be to sludge might explain why sludge related to implementation costs is seen as least problematic in our data. To highlight this limitation, throughout the manuscript we refer to self-reported, rather than actual, vulnerability.
The gap between subjective beliefs about vulnerability to sludge and actual, objective vulnerability also has important implications for behavioural public policy. Some forms of sludge that impose significant real-world costs may not be perceived as problematic by people when filling out a survey and therefore may not be reflected in survey responses. Participants may simply fail to recognise that they would be affected by such sludge or may not consciously register its impact. Moreover, the survey made sludge salient to participants. In the real world, people may not notice sludge in the first place. These issues are particularly relevant in the private sector, where firms may deliberately design ‘strategic’ sludge, such as barriers to service cancellation, that effectively steer consumer behaviour while appearing unobjectionable when described abstractly or when evaluated without direct experience. Such forms of sludge are unlikely to emerge as problematic in survey-based research, yet they may be among the most concerning from a behavioural public policy perspective. It may also be the case that participants report sludge to be unacceptable in a survey which makes sludge salient, whereas people would not notice the same sludge in the real world. An important future research direction is therefore to measure the gap between subjective and objective vulnerability to sludge.
A related reason for the limited external validity of our study is that we asked participants to assess sludge that they did not necessarily experience themselves. This design choice allowed us to compare many different types of sludge without limiting our analysis to types of sludge or contexts that people had experienced, allowing for greater comparability while still using relatively common types of sludge. The lack of first-hand experience, however, is a limitation because it can be difficult for people to anticipate how onerous a task is without experiencing it first-hand due to an ‘empathy gap’ (Soman et al., Reference Soman, Cowen, Kannan and Feng2019). This might explain why public and private sector sludge is evaluated so similarly, and why sludge related to implementation costs is generally seen as least problematic in our study. Future studies might make tasks more concrete by referencing specific government agencies and private companies, focusing on tasks participants have previously encountered, or using real-effort task experiments in which participants must navigate actual sludge within the study context.
A final limitation, and a focus for future work, involves the issue of magnitude. In the abstract, people might think that one kind of sludge is more acceptable than another kind of sludge. But if the apparently more acceptable form of sludge turns to take far more time than the apparently less acceptable form, people’s judgments might flip.
To conclude, we see several avenues for future research in this area. First, it would be important to extend the types of sludge that are analysed. For example, having to cancel a membership is one of the most often mentioned types of sludge, but we did not include it as it was difficult to come up with this type of sludge in the private and the public sector as well. Second, additional (combinations of) predictors of vulnerability to and acceptability of sludge should be tested. For example, data on overoptimism, present bias, status quo bias, limited attention, thinking styles, and cultural and financial capital might help us better understand who is most vulnerable to sludge. Moreover, our data suggest that poor mental health contributes to mental energy scarcity, which in turn increases vulnerability to sludge and decreases acceptability of sludge. This pathway should be investigated further.
Third, repeated exposure to sludge may increase administrative literacy, potentially reducing vulnerability to sludge over time, which could be tested using longitudinal study designs. Fourth, cross-national studies should test whether the patterns found in Ireland are similar in other countries as well. Fifth, stating and varying magnitudes (measured in relevant terms, including time) would be quite valuable. Finally, we see great potential for real-effort task experiments that systematically vary the level and type of sludge and identify individual-level moderators. These experiments would allow the investigation of the non-hypothetical, concrete and causal impact of sludge on experiences and behavioural outcomes.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/bpp.2025.10020.
Funding statement
This work was funded under the EPA Research Programme 2021-2030, a Government of Ireland initiative funded by the Department of the Environment, Climate and Communications, as part of EPA Research Funding project 2022-CE-1149.
Competing interest
The authors declare none.
Data availability statement
All data and analysis code are vailable on the website of the open science framework: https://osf.io/cpdfv/.
Open access
