The contribution of deliberative forums to studying welfare state attitudes: A United Kingdom study

This article introduces democratic forums as a method of studying attitudes towards the welfare state, and sets out briefly its strengths and weaknesses in comparison with existing methods. This is done by reporting the findings of a 2015 two ‐ day forum based in the United Kingdom, in which the future of the welfare state was discussed by a largely representative sample of participants. The results show that participants linked both moral and economic arguments to come to two major framings that could encompass the debates surrounding the future of welfare states. One focuses on the inefficiencies of the welfare state, which found that welfare resources were largely misdirected and unsustainable. The other focuses on the possibilities for improving it via social investment, for example providing individuals with better training and education opportunities. The democratic forum method is helpful in allowing researchers to investigate the conceptual framings people use when thinking about the welfare state, and to see how people link different concepts and justifications together to argue their position. We argue that such framing can be distinct from that used and understood by policymakers and academics, and those applied in the more commonly used large scale surveys.

attitudes towards redistribution or the welfare state's role in it. This is usually done using a theoretical framework of self-interest, ideology, and the existing welfare state institutions (Blekesaune & Quadagno, 2003;Chung & Meuleman, 2017). Most of these studies aim to understand the drivers of attitude formation. In this article we introduce an innovative method for attitudes research: democratic forums. Although commonly used in other fields such as political science (see e.g., Flinders et al., 2016;Warren & Pearse, 2008), this approach is yet to be used on a large scale in the field of social policy research. This article shows how this method can help contribute to understanding welfare state attitudes of the population in a different light. We especially draw attention to the role democratic forums can have in understanding the "framing" of ideas-or the conceptual framework the general public has which links together attitudes and assumptions in generating an understanding of the welfare state. We argue in this article that democratic forums can be especially useful in their ability to uncover the way the general public frames and links ideas.
Large scale surveys, one of the most commonly used methods, have many advantages especially in terms of generalizability and comparability of the data. However, they are restricted by the predesign and precoding of questions which may fail to recognize some issues that may be paramount to the public, whilst not significant for the policymakers and/or academics designing the survey. Although focus groups, the most frequently used qualitative method in welfare attitudes research, can help us to understand meanings and framings in a more bottom-up manner, they tend to be small scale, usually in homogenous groups of the population, which may be limited in their ability to represent the larger population.
By including a larger number of participants, largely representing different sub-population groups, and by gathering data from over a much longer span of time, democratic forums can be useful in order to identify the framework of ideas that ordinary people use regarding attitudes towards the welfare state. We demonstrate this by providing the results of research focusing on public attitudes to the future of the welfare state in the United Kingdom (UK) conducted in 2015.
Our findings show that the most commonly raised issues about the welfare state were framed around people's confidence in the sustainability of the welfare state, and their trust in government.
Section 2 reviews the strengths and limitations of democratic forums in relation to other commonly used methods of welfare state attitudes research. Section 3 presents the method used for the data, which comes from the UK component of a cross-national, NORFACE-funded research project entitled "Welfare State Futures: Our Children's Europe" (WelfSOC 1 ). Section 4 presents the findings, and we conclude with a discussion outlining the contribution democratic forums make to welfare state research. Discussion of methods is necessarily brief in order to make available sufficient space to present the empirical material.

| ATTITUDES RESEARCH: STRUCTURED SURVEYS, FOCUS GROUPS, AND DEMOCRATIC FORUMS
A large and growing number of articles deal with attitudes towards the welfare state, and examines public attitudes to different aspects of welfare policies. Initially, attitudes were approached through a "file-drawer model" in which the individual was understood to possess a sheaf of attitudes on relevant topics and to, as it were, look them up and report them to the researcher in responses to particular questions. Over time, approaches developed that saw attitudes as social constructs and as influenced by information, perceptions, and social communication and framing.
The quantitative structured survey method produces data which can easily be generalized to a population through the use of representative sampling, and is convenient for statistical analysis. The major surveys, such as the European Social Survey (ESS) and the International Social Survey Programme (ISSP), are now routinized so that findings on key points can be provided relatively quickly. Such surveys may sometimes tend to be driven by the interests of policymakers and researchers. From the point of view of government, the welfare state is very much about funding and taxation, what the public desires and will tolerate, but also about the impact of public services, in terms of both subjective (in relation to values and ultimately voting) and objective impacts (in relation to the goals of policy).
Academic concerns are more to do with social changes and how these relate to social values and also with the relationship between attitudes and interests among different social groups. These interests are reflected in the long-standing batteries of questions on government responsibility for various needs, the living standards of different groups, and the impact of state provision on values which figure prominently in the ISSP and ESS questionnaire sections on the welfare state. These surveys are valued cross-national sources for comparative data on welfare attitudes, and have received considerable scholarly attention, leading to more than 1,500 publications in the past two decades (Sundberg & Taylor-Gooby, 2013). We do not go into great detail outlining the findings of specific studies, but for more information and a summary see the introduction to this regional issue .
What we are interested in this article is the choice of topics examined in welfare attitudes research, as well as how it is framed.
There is a tendency for researchers using the structured survey method to predefine the problems to be explored in the topics covered, the phrasing of questions, and the categories within which they can be addressed in the precoded answers (Goerres & Prinzen, 2011). In practice, the interests of policymakers and the theories of researchers tend to guide the construction of the surveys. This makes it difficult to be sure that the framing employed is, in fact, that of the mass public, and that priorities are correctly identified. There is thus a risk that major surveys may ignore areas that the public thinks are important, but that do not emerge in elite thinking, and that the conceptual frameworks used by the public may not be adequately reflected in analysis. Equally, the attitudes people hold may be more complex and ambivalent than the model of the researcher acknowledges, or concepts may be understood differently.
Structured surveys may impose a framing or misunderstand the complexity of people's priorities: as a result, this method is unable to identify non-attitudes (Goerres & Prinzen, 2011).
Qualitative methods have strengths in addressing these issues because they give participants in a moderated discussion greater opportunity to explain, explore, and argue about the matter in question. The most commonly used qualitative method in welfare attitudes research is focus groups, and to our knowledge other qualitative methods such as participant observations are rarely used for welfare attitudes research. Focus group methods are normally limited in time (between one and two hours of discussion), in the range of population groups that can be included, and in the fact that the topics included in the moderator's schedule are typically predefined by the researcher.
The approach we propose in this article intended to overcome some of these limitations. Democratic forums come from a tradition that draws on the development of interest in more participative and discursive models of democracy (Mouffe, 2010), and provide five key strengths. Observation of discussion in forums allows us to: (i) explore how people understand issues and link together their ideas about different aspects of the welfare state; this includes: (ii) the identification of particular issues as priorities; and (iii) the justification of those priorities in relation to ideas about how different aspects of the welfare world impact on each other; forums also allow us to: (iv) see how people respond to information from different sources, and what information carries the most weight with different groups; and (v) to explore the way group interaction influences discussion and final conclusions over a more extended period of time than is possible in focus groups (Wakeford, 2007). Some of these advantages are similar to those when using focus groups. However, democratic forums overcome some of the aforementioned limitations of focus groups by allowing for extended discussions taking place over two or three days, in which the participants are able to play the leading role in deciding the issues to be discussed and how they are approached. Most forums allow opportunities for small breakout groups as well as plenary discussions, and for participants to consult information that they request to be provided by experts, and are only very lightly moderated.
There are also some limitations. Analysis is detailed and laborious even when using modern computer-assisted coding techniques, given the vast amount of data gathered compared to that of shorter focus groups or interviews.
There are limitations in generalizing the findings compared to surveys, despite the best efforts of researchers in including a diverse population covering a wider range of demographics and socio-economic groups. However, the method does not pre-define the issues nor pre-code responses. It is thus less likely to elicit responses to fit pre-coded response categories when those studied do not, in fact, feel strongly about an issue, a problem sometimes experienced when using survey data. 2 In this article we do not have sufficient space to consider all contributions and strengths of the democratic forum approaches, such as that pertaining to how people respond to different types of information, and the group dynamics and attitude formation that can be observed through the data (see Zimmerman & Heuer, 2016). We focus here on the first three strengths, which relate to allowing for a better understanding how the population "frame" their perceptions of the welfare state.

| FRAMING WELFARE ATTITUDES AND THE CONTRIBUTION OF DEMOCRATIC FORUMS
Framing refers to the point that the ideas relevant to understanding the welfare state do not emerge independently, but are linked to other ideas which may support or provide a context for them (Druckman, 2001). In relation to welfare attitudes, ideas about the relative costs of services for different needs as well as the impact on society of the perceived behaviors encouraged by particular welfare programs will influence priorities. Framings link together understandings of the world of welfare with ideas about priorities, and justify the positions that people take up. They may be provided through mass media, politicians' statements or other communications (e.g., Van der Pas, 2014), or be understood as part of people's mental mapping of the issues (Chong & Druckman, 2007). This article focuses on the latter approach.
In conventional welfare state attitudes research (e.g., Blekesaune & Quadagno, 2003;Chung & Meuleman, 2017;Goerres & Tepe, 2012;Knijn & Van Oorschot, 2008;Taylor-Gooby, 2010;Van Oorschot & Meuleman, 2012), welfare attitude preferences are understood to depend on self-interest, ideological beliefs, opportunity structures at both the individual level and the national level and the institutional framework in which they live (for more detail, see the introduction to this issue, . These results are usually derived from analysis of the covariate structures of variables measuring the various explanatory factors along with the dependent variable which it is intended to explain: support for a certain policy, and or preferences towards redistribution. A major limitation to this approach is that we are unable to demonstrate that the covariate structures correspond to factors that might or might not underlie people's attitudes. Democratic forums allow people to offer their own linking between ideas without reference to the conceptual structures of experts, to explain why they hold a particular attitude in detail over time, to say what they think justifies it and to shift or reinforce their position if they encounter other views. In many cases, the framings offered are not supported by academic argument or by the best empirical evidence. We suggest that democratic forums offer different and complementary insights to those gained from structured surveys and focus groups in examining people's attitudes towards the welfare state by allowing the observation of how people frame issues, priorities, and justifications. Given their aforementioned limitations, forums have most value when used alongside structured surveys and other methods, and improve insight into the finding of these methods.
We now illustrate these points by reference to a recent democratic forum exercise carried out in the UK, paying particular attention to the capacity of forums to identify people's key ideas on the topic and the way they link them to justify their positions: the framing that ties together priorities and justification.

| METHODOLOGY
Our forums were carried out across two full days in a large metropolitan area in the heart of the UK (Birmingham) in October 2015, i.e., five months after the 2015 General Election and six months before the (official) start of the Brexit referendum campaign . They included 34 participants in a sample that was roughly proportional to the make-up of the UK population in demography (gender, ethnicity, age grouping, and family composition) and socio-economic characteristics (employment status, class, and education). We must, however, note that there may be some self-selection of those interested in taking part which may influence the representativeness of the sample. The overarching question discussed over two full days was as follows, "What should the priorities of the UK government be for benefits and services in 2040?".
The process of the forum was as follows. At the beginning of day one, participants were asked to identify the key themes and priorities they consider relevant to the future of the UK welfare state, with no input from the research team. The five themes selected by participants were, in order of importance: immigration; lack of money (regarding welfare state financing); unemployment; overcrowding/aging population; and lack of/access to education. These themes were debated throughout the first day. The second forum took place two weeks later. Between the two days of discussion a stimulus package of relevant factual information was distributed to participants in response to questions raised by participants at the end of the first day. The information provided focused on key welfare state issues, such as immigration rates, the cost of various services and the number of claimers for the various benefit (for further details, see Taylor On day two, participants were asked to prepare and justify policy recommendations for the future of the welfare state. The objective of this activity was to give some direction to the discussions. In contrast to day one, the discussion was structured around five broad themes imposed by researchers: income inequality; immigration; gender; intergenerational issues; and labor market. This list reflected the main themes in academic debate, and was imposed in a departure from standard democratic forum procedure to ensure comparability between the various national forums carried out as part of the cross-national WelfSOC project. There were two major differences between day one and day two: we conceptualized the issues more broadly (referring to labor market issues, which includes education and training opportunities, rather than unemployment, and intergenerational issues as a demographic concept rather than simply an issue of pressures on the welfare state from an aging population); and we included gender inequality as an issue (in parallel to the class and income inequality which had emerged spontaneously on day one, and which we associated on day two with the labor market). Discussions took place in plenaries at the beginning and end of each day and in smaller breakout groups of 10-12 people, to facilitate interaction. In each breakout group, participants sharing one specific socio-economic characteristic (i.e., self-employed, unemployed, or ethnic minority status) were over-represented in order to determine how shared characteristics contributed to group dynamics. In total, 35 hours of discussion were audio and video-recorded. Analysis took place through a process of coding themes and ideas and tracing the development of arguments. Data was coded with NVivo, starting from the five issue areas identified by participants, and expanded the list of codes through iteration on the basis of topics emerging in the discourse to identify 21 topics in all. We then examined the structure of the various discussions, paying particular attention to the themes which dominated and the arguments that were used to support them, as the discussions moved to a conclusion in the final plenary.

| FINDINGS
Before providing the results of the forum, we discuss briefly the main characteristics of the UK welfare state. The UK is often seen as a liberal welfare regime, heavily reliant on market-focused approaches and individual responsibility (Esping-Andersen, 1990). Welfare policies are generally targeted, and there are high levels of inequality and poverty rates, especially after the series of tax cuts and cuts on social care and benefits carried out in recent years as a part of a larger set of austerity measures. Despite the low levels of benefits, stigma towards benefit recipients especially on unemployment benefit is prevalent (Baumberg, 2016). Over the past decade, the issue of immigration has been high on the political agenda, especially at the time of the fieldwork when the Conservative government committed to reduce net migration "to the tens of thousands" and with immigration being a decisive issue discussed during the Brexit referendum campaign. Other concerns in relation to the welfare state include the crisis of the National Health System (NHS), the lack of housing, the collapse of social care systems, and issues of low productivity growth within both state services and the broader economy (for more detailed analysis, see Taylor

| Framing the UK welfare state
The framing of the future of the welfare state used by the majority of participants linked together practical and normative issues: perceptions of the financial unsustainability of current provision were combined with a morality of individual responsibility through a view of the state as inherently wasteful, inefficient, and above all counter-productive in the way it allocated resources and failed to reinforce the work ethic (see Taylor-Gooby, 1985). The UK welfare state was seen to face pressures primarily due to immigration, since immigrants were believed to impose a burden on public services that far exceeded any contribution through work and tax. Other pressures were the cost of pensions and health care, generous unemployment benefits, and overseas aid. The weakness of the state was indicated by its failure to control immigration, cut benefits, tackle dependency, finance pensions and health care, and ensure that multinational companies pay taxes. At the same time, the main services-pensions and the NHS-were highly valued, and most people regretted what they saw as their unsustainability. The needs of vulnerable groups were discussed in relation to the same principle of deservingness, which explained why immigrants and work-shy unemployed people did not deserve help. Some refugees and people who could not work merited help, together with people who worked for inadequate wages and on precarious contracts such as zero-hours contracts. The policies desired for the future were built round the idea of deservingness: the state's main role was to provide opportunities for people to work effectively and contribute to the economy and to society. In many cases, this was thought to be possible through provision of training, education, and cheap available childcare. This logic links practical and moral ideas together, and leads to two overall framings: one stressing the unsustainability of the welfare state in view of current pressures and damage to the ethic of individual responsibility; and the other concerned with social investment, opportunity, and enhancing commitment to work. We trace these frameworks in more detail in the following sections.

| The first framing: The misdirected welfare state
The core idea is that the UK welfare state faces severe pressures and is incapable of managing them. Government spending was seen as poorly controlled and inherently wasteful: "there's a lot of waste in publicly funded things, because nobody feels that it's real money. It's kind of public money, so therefore accountability I think goes out of the window" (UK P-87) 3 Accordingly, the public sector was heavily criticized for a lack of accountability and transparency: "Somebody's given £10, £15, £20 million, whether it be council or a government department, they've got to say what they're spending the money on, where they're spending the money and what are going to be the benefits to us as, as part of the society, yes so there's got to be more accountability of where the money's going and what it's being spent on and more transparency" (UK P-84) UK contributions to the European Union (EU) are also seen as an aspect of the government's failure to control spending and direct it where it should be going: "[W]e really ought to make enquiries into the amount of money that's paid by this country into the EU, the astronomical amount of money that is paid into the EU every week, billions." (UK P-85) Many highlighted that fact that they do not know how taxes are used and profligacy in overseas aid was often cited in relation to this.
Immigration was perceived as by far the biggest challenge to the UK welfare state. It was framed by most participants as a major strain on state spending in areas such as health care, education, and housing through increased demand. Several participants believed that immigration damaged the sustainability of the NHS: "there's no way I would pay for an NHS service for people to come in and abuse it, I would pack my stuff and I would go and live in Australia" (UK-P80) These views were expressed across the parallel breakout groups and also in plenaries. However, participants made a clear distinction between "economic migrants", who came to the UK for a better standard of living, and "refugees", mostly from the Middle East and North Africa, who seek asylum, and this introduced the moral value of vulnerability: "the people who are trying to get away from war and then you've got the economic refugees who haven't really got a reason for leaving where they live except they want to get some more money" (UK P-45) Economic migration was mostly framed as having negative material consequences, "[w]e are a small island so we can't keep bringing people in" (UK P-86) was one of the key justifications from the participants, who overwhelmingly favored the introduction of an Australian-like points-based entry system to cut immigration and reduce what they saw as the resulting strain on public services. Nonetheless, the participants expressed considerable sympathy for refugees.
The view that in any case incomers were a burden than the UK could not sustain predominated: "Obviously, if we keep going on as we are now, there'll be a lack of housing and space to house people.
So hopefully, it'll reduce that. Again, it puts a strain on the NHS. And hopefully, it will put some people off coming because they won't be…we're not the Promised Land, you know" (UK P-45)

[…]
"I think you allow immigration more and more and more, the Government have to spend more money on benefits for them instead of putting the money into education and social care and other kind of stuff." (UK P-61) This discussion illustrates how groups would come to agreement reflecting the strongest view and contrary reviews tended to fade from discussion.
The participants tended to see immigrants as benefit recipients and a burden on public services, rather than as net contributors. Furthermore, participants perceived immigration to be a major cause of high unemployment, due to increased competition for what is seen as a limited pool of jobs: "It's an issue because it's probably restricting our British citizens from having access to you know, like education, to job opportunities. It's simply because, you know we have Europeans and other people from around the world coming in and it's restricting and taking all the opportunities that we could possibly have." (UK P-n.i.) However, a minority of participants also acknowledged the contribution immigrants make as workers: if you look at the NHS, it's a lot of the immigrants who sort of work in there, nurses who come over from Africa and things like that" (UK P-66) In spite of this, almost all participants agreed that current levels of net migration are too high and need to be addressed, to avoid further strain on the welfare state.
Perceptions of how government managed unemployment heavily influenced participants' framing of the welfare state as a burden. This was both a material and a moral issue, since benefit abuse wastes money and undermines an individual's sense of responsibility. This is linked to the previous two factors in that unemployment benefit was an example where the government made unjustifiable decisions and where spending is perceived to be out of control: "You see people in the paper, on a weekly basis, 'I'm fifty, I've never worked in my life, I live in this massive great house, I get all these benefits, why would I need to get a job?' And that's the government's fault for not keeping tighter control" (UK P-n.i.) Government benefits for unemployed people are also seen as a further aspect of state incapacity because they are believed to damage incentives, lead to irresponsibility, and damage moral responsibility that can be passed on to future generations: "I think we're too soft on unemployment. I think, you know if you, you know, if you're fortunate to have a job and you're being made redundant or anything like that, why should it be so easy still to live on benefits or state money and not feel the get up and go to have to go to work?" (UK P-68) "… the poor children being born into a society that encourages financially the mother to have another child because she gets paid more benefits, what sort of chance are they going to have with life and role models"  This problematic nature of benefit dependency was also linked to the low pay and working conditions for workers such as those on zero-hours contracts, who might be demoralized through work that was unpleasant and paid worse than benefits. This was also linked to the tendency for companies to hire foreign labor, especially in times of global competition and possibility to relocate elsewhere: "people on zero hour contracts, they're not entitled to the same things as permanent full time employees […] they don't qualify for sick pay, there's no guarantee of work, they're not paying the […] tax or National Insurance, […] they can't get loans, they can't get a mortgage, they don't know when they're working, they have really no say" (UK P-45) "How can someone on benefits be earning, be getting £40,000 a year and the average person is probably getting around £21,000 to £22,000 a year, working hard, and they're getting double the amount? It's not fair" (UK P-43) "[T]here's a belief that it's easy to remain unemployed because you are given the hand-outs and that people are unwilling to do lower paid jobs. They don't want to sign a zero hour contract, […] but somebody from abroad will come and they'll happily sign it, because a job's a job, and they'll work as many hours as they can but we've also said companies outsource and because small businesses have to pay such taxes then it might be cheaper to make a product abroad and ship it in, or car parts, a factory will close and it will go to another country." (UK P-n.i.) The core idea here is vulnerability: the contracts make life worse for a particularly vulnerable group, and this was widely seen as unfair. Moral commitments associated with the needs of vulnerable people also conflicted with the assumption that costs were unsustainable. This led to discussions regarding cuts in three other areas: overseas aid, highly-valued health care, and pensions.
The widely-held view that overseas aid was an area in which perceived British over-generosity indicated poor judgement by government was mentioned earlier. Some participants defended aid, and this created divisions within the group. These members argued that helping other countries is a key British value, much like welcoming refugees, drawing on the notion of helping the vulnerable: "[…W]e need to care for those people. In other countries whereby we consider, you know, us as human beings we have some sort of moral…" (UK P-67) Others, however, believed that the government is sending too much money abroad and should care about domestic issues first. There was an overall consensus to cut back on foreign aid, and most believed that aid was an extravagance: "Priorities should be here at home, and our needs, the public's needs, the country's needs. The future, we're talking here now" (UK P-49) The NHS is the most highly-valued service in the UK, and this was reflected in the discussion. There was agreement that cutting health tourism and imposing greater responsibility on people for managing their own health through lifestyle choices would contribute to savings, but the overall conclusion was that the service was unsustainable: "If we keep doing what we're doing the [NHS] is going to be bankrupt, … it's going to run out and it cannot sustain itself" (UK P-84) […] "It will have to change, it's got to change, but I mean, it's the biggest gift that…I mean, it's a thing of beauty, isn't it?" (UK P-84) "Absolutely, the best thing England's ever introduced" (UK P-87) "It is, yes, but it is being abused already" (UK P-84) Similarly, pensions provoked debate about the value and the difficulty of meeting the needs of a vulnerable group.
When one participant argued that pensions would simply be unsustainably expensive in the future, another responded, to general approval, that pensions provide "a quality of life, not work yourself to death until you're 75" (UK P-44). When one breakout group proposed increasing the retirement age to 70 or 75 and stopping funding the basic state pension ("we just didn't think the funds would last unless we did something like that",U KP -85), participants from the two other groups raised objections: "[this] policy is, like, to create more unemployed people if they raise the retiring age to 70 or 72. In 2040, it's going to cause more unemployed people than employment. That's not going to help the economy" (UK P-43) However, the consensus was that universal pensions at current levels were also unsustainable although a meanstested pension might remain: "It all came round to the fact that, in 25 years' time […] because of increases in the NHS and people living longer, and more people at retirement age, it's just going to be impossible to do the maths […] it's going to be a totally different system in 25 years' time than it is today […] … 12 million now […] That's going to go up to 20/25 million and it's just going to be impossible" (UK P-84) These commitments to fairness and meeting the needs of vulnerable groups, for example by scrapping zero-hours contracts, and maintaining provision in health care and pensions, conflicted with the framing of ideas that linked unsustainable spending and misguided government and set against it the value of individual responsibility.
However, no one offered a counter-framing that brought these ideas together. A thoughtful comment by the spokesperson for one of the groups in the final plenary drew the arguments about responsibility, benefit-dependency, immigration, and unemployment together, and is quoted at some length: "it's easy to remain unemployed because you are given the hand-outs and that people are unwilling to do lower paid jobs.
[…] Reducing […] immigration or making sure that only a certain set of people come into the country will mean that there are more jobs to go around the people that are here…" (UK P-46) In general, stronger views tended to emerge as a consensus and weaker views to dwindle. The claims about state wastefulness reinforced each other, as did the ideas about dependency, allowing more immigrants who were a burden on public services to enter the country and the way benefit generosity and failure to control exploitative employers eroded work ethic morality. There was no corresponding integration of the minority ideas about helping vulnerable people, refugees, and the unemployed, and consequently no redistributive way forward. The upshot was that discussion in general moved towards the view that immigration, and spending on pensions, health care, overseas aid, and unemployment benefits were key problems. Government profligacy should be brought under control, and most importantly immigration cut back in the national interest. This would help to contain unemployment and health care spending by making more jobs available. It would also save spending. The profligacy of government also damaged worker morale through poorly regulated unemployment benefits, and this fed back into damage to the economy.

| The second framing: Social investment as intervention
Alongside the first framing of the misdirected welfare state, a second emerged related to intervention and social investment. This linked employability, education, training, opportunity, and individual responsibility. The core idea is that it is low skills, poor labor market opportunities, and working conditions that generate unemployment and poverty, and that these issues can be tackled in the longer term through better training and a restructuring of education (see also Zimmerman, Chung, & Heuer, 2018). Better opportunities will instill a stronger sense of responsibility into individuals, and reduce welfare dependency, linking together a moral and a practical discourse. Participants were particularly keen to see better training opportunities for less academically successful young people, often summed up in the idea of apprenticeships. The following summary received a lot of support as a key priority in the final session: "We need to think of some more long-term solutions…Education, investing the money in the education to try and combat unemployment. We discussed a lot around [how]…to retrain the youth…with apprenticeships to keep employment on a rolling scheme…Cracking down on tax avoidance that we see in a lot of large companies…Limit money to other countries." (UK P-64) Apprenticeships and training opportunities are also seen as directing behavior in virtuous directions: "There would be no layabouts on the street, causing trouble." (UK P-47) More generally, training and jobs would help instill responsibility: "Until we get to the state where people are taking responsibility for their actions then it's always going to be the same. So that's something that's got to change before 2040…"  This links together the two framings. On the one hand, participants stated that it is simply wrong that government benefits allow unemployed people to sustain a living standard close to that of the low-paid. On the other hand, they also voiced a need to support individuals in taking responsibility for themselves and contributing to the society by providing better labor market opportunities and training.
These arguments were drawn together by one participant, who linked the different points as follows, to general approval: "People who aren't working and they're receiving benefits, we discussed the fact that, […] they should have to contribute in some way, be it charity, voluntary work or whatever. We said the current system effectively encourages people to stay out of employment […] there should be equal opportunity, equal access to education, so perhaps more money invested in trained skills and apprenticeships. There should be tax credits or tax incentives for companies to provide more apprenticeships, not just for the young people but also people that want to retrain, and that would obviously in the future bring tax back round, more people being employed. We said that another issue with apprenticeships is they're not potentially giving people a future. Some companies are opting into apprenticeships just to gain government funding […] zero-hour contracts is another issue". (UK P-67) Interestingly, the issue of gender was not raised unprompted by participants on the first day, and when introduced on the second day, many found it of little relevance and did not frame it in terms of inequality. Many academics argue that weak provision combined with assumptions about the gender division of caring responsibilities systematically discriminates against women (Lister, 2003;Orloff, 2010), but most participants seemed to identify "gender" as a narrow theme focusing on equal pay: "You don't hear people talking about that" (P-88) "In 2040, right, I think the gender won't be a problem, men and women earning, be used to be same level" (UK P-48) However, there was considerable support for affordable and accessible childcare (see also Chung, Filipovič Hrast, & Rakar, 2018). It was clear that many were aware of the limited availability of places and very high costs: "It's very expensive, very expensive. If you've got a job and you're not earning loads and loads of money, and you have a baby and then you go back to work, half of your money, well even more… would go on childcare" (UK P-41) Participants argued that the state should invest more in childcare facilities and reduce costs for working parents.
The rationale was not expressed in terms of gender rights, but in terms of expanding labor market opportunities, to benefit national economic interests. Accordingly, participants voiced a need to enforce the involvement of parents in productive activity when receiving free childcare: "With more women back in work and earning fair wages then they're going to be paying more tax." (UK P-43) "[parents who receive free childcare should] do something, even if it's an educational course or […] training" (UK P-84) Some participants saw childcare as also a part of a larger social investment strategy focused on future economic development. This linked with the first framing through the belief that benefits and welfare state inefficiencies hindered active participation in paid work; the state should promote people's active involvement in economically productive behavior.

| DISCUSSION AND CONCLUSIONS
In this article, we examined data from a UK-based democratic forum study to examine the usefulness of this method in understanding the welfare attitudes of individuals. We argue that forum data allows us to better examine the priorities and justifications of attitudes and to see how information is linked together in the way issues are framed.
We found that participants perceive the welfare state and its future through two framings which differ from those common in academic discussion. The first framing saw government provision to be fundamentally inefficient and misdirected. It mismanages immigration, pursues wasteful NHS and pension practices so that these services, though highly-valued, are unsustainable, and operates an unemployment benefit system that both wastes resources and damages people's sense of responsibility. These ideas also linked to a perception of immigration as a major problem: incomers were seen as competing with British workers for a limited number of jobs. Participants largely ignored the possible benefits of immigration.
This framing is bound up with a second framing that sees the welfare state interventions as essential to provide better education and training and affordable childcare so that individuals have better opportunities in employment. This different, more active, approach believes that improving work opportunities will both help economic growth and reinforce the work-ethic morality. These two framings are linked together by an overarching morality of self-responsibility expressed most importantly through the valuing of work and the individual's contribution towards society.
These framings of welfare-the burden of immigration (identified as the first priority by participants), the moral damage inflicted by the unemployment benefits system and the value of work opportunities-offer an approach to the problems facing the welfare state that is in many ways an alternative to that dominant in academic literature.
For example, many academics stress population aging, or deindustrialization as the main issue (e.g., Pierson, 2001).
The issues that academics see as linked to population aging (rising costs in pensions and health care) also figure in popular understanding, but are seen as of secondary importance against the more dominant perception of government as wasteful, and support the view that current levels of provision are unsustainable. However, this is of secondary importance compared with the moral implications of the profligacy of government.
Everyday arguments stress a morality of work and responsibility for oneself rather than collective provision. This conflicts with the original conception of the welfare state as providing care "from cradle to grave" (Beveridge, 1942).
Elements of an alternative view are present in the discussions. Some people saw some immigrants as assets, defended overseas aid, and felt a moral obligation to help refugees and other vulnerable groups essentially through redistribution. These views were held by a minority, and tended to decline in significance as discussion progressed.
Most importantly, they never came together in a unified defense of a redistributive interventionist welfare state as justified, practically because it is efficient, and morally because it meets needs and reinforces the public good. The dominant viewpoint grew stronger and more integrated during the forums. There was no mention of trade union politics or of collective workplace rights.
Because participants express framings differently from those that form part of conventional academic wisdom, they allow researchers access to ordinary people's understanding of how welfare state attitudes work. These framings link together moral (individual responsibility; obligations to the vulnerable) and practical concerns (sustainability; growth). Probably, the most exciting aspect is the strength of support across the participants for what many expert commentators would term "social investment" as the heart of an effective welfare state in 2040. Perhaps the most sombre is the misunderstanding of the real benefits and costs of immigration, and the assumptions that government is irretrievably inefficient and misguided, nurturing a determination to cut immigration and benefit spending, and the belief that misguided state spending rather than demography and tax cuts are responsible for current spending pressures. All these issues emerge in analyses based on pre-coded quantitative material. However, the forum approach shows how they link together in a larger conceptual framing encompassing a multifaceted understanding of the welfare state that reinforces them and makes them into a coherent view of the world with policy implications: cut back the provider welfare state and build up social investment.
As this article has shown, democratic forums generate new knowledge in relation to framing, priorities, justification, and the use of information. They do not provide the kind of precise information on the attitudes of members of different population groups towards a defined and well-understood attitudinal object that competently-executed structured surveys can provide. They can also be expensive and difficult to conduct and analyze. They have two kinds of value as an addition to the armory of social research methods. First, they can help to explore the framings that ordinary people employ when thinking about a particular topic, and thus contribute to the design of surveys, ensuring that issues are considered which may not immediately be included because they do not figure in expert theories about the issue. Second, they contribute to understanding the strength with which positions are held, how ideas develop, and the areas on which people will compromise in discussion. Democratic forum researchers may wish to include some kind of goal to the discussion, in our case agreeing bullet points for a report, to motivate participants to arrive at such an agreement.