Working together : a study on co-produced parenting support services in a Swedish municipality

This article describes an evaluation of co-producing parenting support services in a rural municipality in Sweden. The evaluation looked at the implementation process of a series of positive activities for children and their parents, data from the participating parents and the economy of the project. Methods included interviews, questionnaires, observations and a review of protocols, e-mails, and web pages. Both qualitative and quantitative analytical methods were used. We identified two essential components for successful co-production: 1) The balance between the two actors (public and voluntary sector) in terms of interests, commitment and experienced positive impacts and 2) That both actors remain within their core competencies. Parents were satisfied with the number and quality of the activities and reported improvement in their relationship with their child. Costs for the six sessions provided during the project were 1897 SEK/participant for the orienteering club and 2635 SEK/participant for the riding club.


Introduction
The Swedish public sector has changed radically during the last few decades.Similarly to many other (Western) countries, a wave of global transition and neoliberalism in the economy has swept over the country.This so-called New Public Management (NPM) aims towards reducing public sector expenses, privatising or outsourcing public sector responsibilities and accepting market rules (Sundin and Rapp 2006).Within such a paradigm, where both the public and the private sector have become competitors in delivering public service contracts (Gonäs et al. 1997), Sweden is still a country with an impressive public sector with high government ambitions to regulate the ongoing privatisations.In addition to affecting freedom of choice, the fundamental aim of this reform is to provide more control for those who demand the services, as well as a more effective and decentralised way of utilising public resources (Montin 2002).Two models in this new governing method are the "Clientperformer" and "Co-production" models.
Sweden is recognised throughout the world for the involvement of its citizens' in non-profit associations (voluntary sector or third sector).There are approximately 200,000 active associations in Sweden (www.voluntarius.com, 2012) and 75% of all adults are members of at least one associ-ation (Olsson 2007).Like many other countries, the voluntary sector in Sweden cooperates with the governmental authorities to produce or deliver public services.Relatively few empirical studies however, have investigated the circumstances under which successful co-production might occur, the possible benefits for users of such co-production, and the costs incurred.
This article investigates a series of "positive activities for parents and children together" that were co-produced by the public and voluntary sectors in a sparsely populated municipality in Sweden.The central premise of the project was that the creation of circumstances where parents and children could spend time together in fun activities would improve their relationship.In addition, the municipality was keen to introduce activities that would accommodate families with diverse backgrounds in order to avoid the development of segregation in the community.
The aim of our study is to identify and analyse the circumstances under which successful co-production of parenting support services may occur, to describe the possible benefits of the activities co-produced for its users, and to calculate the exact costs of this co-production based on the different activities.In this article we use the words local associations and clubs synonymously and, along with NGOs, view them each as part of the voluntary sector.Most associations have member fees and some have fees for activities when these involve additional costs, such as for venues or specific equipment.Some activities are more costly than others, such as riding, potentially creating a social gap between those who can and cannot afford the activity.

What is co-production?
Co-production means that two or more parties produce something together.The term was introduced in the 1970s by Professor Ostrom at the University of Indiana (Stephens et al. 2008).Co-production is either sporadic or temporary or, as Joshi and Moore (2004) define, can be based on a regular long-term relationship between state agencies and organised groups of citizens.During recent decades the idea of co-production has emanated from the currently dominant neo-liberal perception that the state organisations are inefficient (Robinson and White 1997).In poor countries with scarce state organisations, co-production can be a particularly cost-effective means of mobilising resources (Joshi and Moore 2002).
Co-production involves citizens through the voluntary sector in the distribution of public services, promotes cooperation and collaboration between individuals and organisations and consequently, also promotes democracy through collective responsibility and decision-making.Through co-production, the government also has better control over the citizen's needs, wishes and problems as well as better control over the budget, especially in a recession (Mottiar andWhite 2003, Pestoff 2008).Low production costs, as a result of free and non-bureaucratic jobs, performed by volunteers and in some cases even their 130 Barn 30(4)

Sweden is still a country with an impressive public sector with high government ambitions to regulate the ongoing privatisations.
financial in-kind donations can be other major reasons for supporting the co-production concept (Hoggett 1991).Bennett (2008), with support from several international studies, argues that the voluntary organisations allegedly: command public trust; have extensive professional expertise in specialist areas; are much closer to service users than government bodies; provide jobs that lead to innovation, flexibility and superior productivity; and care more about values and quality than profit.

Arguments against co-production
Critical voices in Sweden claim that through co-production the full responsibility is put on the voluntary sector rather than being shared and can therefore be a way for the government to cut social expenses, as well as threatening salaried employment (Svedberg et al. 2010, Olsson 2007).Salamon (1987) criticises inadequacies in performance associated with co-production such as lack of training, experience, and professional skills.The Swedish National Board of Health and Welfare (Socialstyrelsen 2004) questions the low competence level in e.g.volunteers performing social work and the lack of rules for individual attorney's activities in sparsely populated areas (Socialstyrelsen 2005).Service workers and public administrators may perceive the voluntary workers as untrained and so resist both them and the whole idea of co-production itself (Pestoff 2008).The principal concerns about problems occurring in co-production is that public accountability may be diluted, and boundaries blurred between public, private and voluntary sectors through the division of power, authority and control over resources amongst the three sectors (Bovaired 2007).Brett (1993) looks at prob-lems associated with the measurement of efficiency and the enforcement of accountability in organisational life and at the need for incentives and sanctions that effectively provide a stable basis for maintaining commitment.

The case project: Positive activities for parents and children
This study evaluates a parenting support project where the participating associations would start a series of positive activities for children and parents together, designed to integrate and support socially isolated families (Figure 1).A welfare officer in the municipality put forward the idea of cooperation with local associations to provide parenting-support activities and thus relieve families from the social stigma of being helped by the welfare services and provide them with "real" and not only professional networks.
The method was co-production as opposed to the client-performer model, meaning that the exact forms of the intervention would be developed together with the participating associations.In other words, the municipality would work together with the local voluntary sector to develop the interventions as opposed to the municipality ordering a predefined intervention from associations and paying according to some predefined performance measure, as would be done in the client-performer model.The target groups were Swedishborn children and their parents along with established immigrant families on the one hand, and socially or geographically isolated parents and children, with both Swedish and immigrant backgrounds, on the other.
Local associations in the municipality were invited to an initial meeting by the project leader who was a representative of the municipality's department of Social Services.Representatives from 11 local associations participated and were informed about the project.They were asked to return within three weeks with their application, and more detailed information on how they would like to deliver positive activities for children and parents, should they wish to participate.Three associations were subsequently selected: an orienteering club, a riding club and an adult educational club.These clubs had well described ideas and schedules and conformed to the defined target group of 6 to 12-year-olds with a minimum of six activities.

Aims and research questions
The aim of our study is to identify and analyse the circumstances under which successful co-production of parenting support services may occur, to describe the possible benefits of the co-produced activities for service users, and to calculate the exact costs of this co-production based on the different activities.a whole) and 23% have pre-secondary education (14% for the country) (SCB, Education and Research 2012).
The design of the study is a case study where data sources were triangulated to include participant observation, questionnaires from different stakeholders, and written documents (Appendix 1).
We conducted e-mail questionnaires with the project leader and two heads of involved associations.They received 7-9 questions concerning how they managed the project, what motivated them to participate, what their expectations were prior to starting, what they perceived to be the strengths and weaknesses of the project, and what could have been done better.The participants were asked to answer questions on two different occasions: at the outset and after the activities were performed.The head of the Adult Education association preferred to be interviewed by telephone.In addition, all e-mail correspondence with local project workers that referred to the project was saved and included in the analysis.
We also asked the participating parents to fill out a questionnaire (one per family) immediately after the first occasion -baseline, and after the activities were finishedfollow-up.The questionnaire included 27 questions relating to demographic variables, motives for joining the project, general health (Sullivan et al. 1995), the parent-child relationship, and participants' social networks, using the Appraisal and Belonging subscales from Cohen's Interpersonal Support Evaluation List, ISEL (Cohen et al. 1985).
We also used other materials, such as official documents, our own minutes from different meetings, e-mails and phone contacts with the project leader and heads of the clubs, and observations at the activities, registered through field notes, by one of the researchers (S.A).Participatory observation made it possible to speak with both participants and representatives of the clubs and provided a sense of how the activities were being carried out.
Information on costs involved and hours spent were collected from the project leader and complemented with official database information.Data collection occurred between February 2010 and February 2011.

Data analysis
To describe the process of co-production, we used minutes and observational data.
For the e-mail texts we employed content analysis in accordance with Graneheim and Lundman (2004).Content analysis involves two parts: analysis at both the manifest and latent levels.The first one deals with the objective, systematic and semi-quantitative description of the manifest content of communication, whilst the latter includes interpretations of latent content.In our analysis of the participants' perceptions of the process, we printed out all available materials and read through them several times to identify the manifest content categories.These categories were agreed upon after a co-reading procedure that included both authors.After having identified the manifest categories, we searched for latent content asking ourselves the question: what might have been most crucial in all this for successful coproduction to occur?
The questionnaires were entered into SPSS and descriptive statistics were used.The statistical data was complemented by free comments given by the participants.The ISEL subscales could only be properly computed in 14 cases as a result of internal dropout in the rest.Due to the small sample size, no further statistical analyses were conducted.

Marketing the co-produced activities
The co-production in this project involved the local Social Services, with a project leader, and the voluntary sector in the form of local associations, whilst the activities and the collaboration were initiated and financed by the local municipality.The project was named "Positive activities for parents and children" to reflect its content and broad target group.It was promoted through a printed, detailed information brochure that was developed with input from the participating associations.
The municipality provided assistance with the layout and printing of the information brochure, whilst the content was created together with the associations.The municipality then utilised their channels for effective marketing through schools, SFI (Swedish for Immigrants), and the local press.The local project leader personally delivered the brochures to schools and Swedish classes for immigrants and used his knowledge of these institutional systems to advertise for the project.
In addition, adverts were published in the local newspaper.The clubs, on their own, could not usually afford this, but they enjoyed the exposure and did their part by using their own web page proactively.
There was a story in the local paper that ought to give a positive picture of the Club's activities.We are also going to do an own story for our website.
(Follow-up e-mail questionnaire from the head of the Orienteering Club) It was found however, that it is not only the reach of the advert, but also the content, which is of importance in attracting participants.The project leader described the problem that one of the associations had with the announcement: The information about the activities was fuzzy.There were too many options to choose from.They should have concentrated on one or two activities.In the brochure they offer: "Music, dance, drama, theatre and creative play".The participants get no clear picture of the nature of the activities.

Initiating the co-produced activities
At the first activity of the Orienteering Club 21 adults and 24 children attended, both Swedish-born and foreign-born in approximately equal numbers.This was an explicit strategy in the project: to blend families with different cultural and social backgrounds.The first activities consisted of learning to use a map while taking a walk in the inner city.For both the first and the subsequent excursion to the forest, transportation was organised and participants were informed about how they should dress and what they might need.The club provided snacks and drinks.The local project leader made sure to be present and personally engaged in conversations with those participants who risked being left out.The club and its activities received great attention in the local paper and a positive reputation amongst families in the community while it was on-going.This could be one of the reasons why the number of attendants increased to over 50 after the second week, and to over 65 parents and children after the third week.
The Adult Education (Sensus) that offered parents and children activities like music, dance, and theatre was only attended by seven parents and eight children at the start-up meeting.The next meeting had even less parents and by the third meeting only one parent and child came along.Consequently, the club and the project leader together decided to adjourn the activities until the following semester.
The Riding Club received approximately 50 applications prior to start.Due to the limitations of the activity (for security reasons), they were forced to reduce the number of participants to nine parents and nine children.The parents and children who did not get access to the activity in the first round, were, however, invited to participate once the first series was over.

Perceptions about co-producing activities -manifest categories
Local anchorage and shared interests Both the municipality and representatives of the voluntary sector had a clear sense of local anchorage.The municipality representative (the project leader) had good general knowledge about local inhabitants, especially about vulnerable families.He had also been active in associations himself.
We have had this idea of doing something with local clubs and associations for a while and have wanted to try it out.That's why I wrote the project proposal.(E-mail questionnaire from the project leader) The associations, on their part, had their facilities, activities, materials and mem-bers locally.They also had access to their own local networks as well as to various social functions.They welcomed the idea of using their networks for social inclusion.
It's a good idea to use already established associations to create new networks.(E-mail questionnaire from the head of the Riding Club) In addition, they believed in what they did and had a sense of purpose about their activities.
We want to spread knowledge about nature and its possibilities to promote a sound and healthy life.(E-mail questionnaire from the Orienteering Club) Since the municipality had identified a significant target group, for whom they had previously not been able to offer sufficient social assistance, the Social Services therefore had an interest in finding a solution to support vulnerable families.The project leader identified a need for a nonprofessional network and activities for the target families.This seemed a key point in relation to the initiation of co-production.
Many marginalised families experience loneliness and alienation.Here club activities can give a more long-term social network than for example an effort from the Social Services can offer.To participate in a club activity with other parents is not an effort but a positive, normal activity "for me and my kid together with many others".(E-mail from project leader) The clubs themselves had struggled with too few members and economical prob-lems, and therefore had an interest in recruiting new members and in finding new ways to strengthen their economy: Perhaps we get many new active members in our club until next year.(Orienteering Club's website) The economic stimulation increased our motivation to apply.(E-mail questionnaire from the Orienteering Club) In dealing with these challenges, the clubs wanted to profile themselves in something that was close to their ideal.
The target group of parents and children together suits us very well.That's how orienteering works, at its best.(…) (E-mail questionnaire from the Orienteering Club) At the same time, knowing that the project would cease, the head of the Riding Club was concerned that financial reasons would keep otherwise willing immigrants, or those with less resources, from becoming members in such clubs: It would be great if those who have no means but want to continue got help with the economy.It is not so easy to be a refugee/immigrant without income and get the finances to work.(Followup e-mail questionnaire from the head of the Riding Club)

Commitment and a flexible approach
Commitment has been a further common characteristic shared between the actors; there seems to be a need for a "driving spirit" on both ends to get co-production going.
We really feel it is important to take good care of all the people at the associations who put a lot of effort into the project and we would like to invite them for an inspiration day.(E-mail from project leader assistant) The municipality representative attended start-up meetings and encouraged families he knew, to participate.He made sure that socially isolated families who might have felt a bit awkward to begin with would join the conversations.He was committed to be there for those who needed his assistance and support: We have learnt that the municipality has to have close contact with associations and parents.Keep it together where it can fall apart.Support parents to participate, support the club to go on.Assist practically with transportation and, in the beginning, attend the activities.That creates a good cooperation climate and trust between the municipality and the association at hand.(E-mail questionnaire from the project leader) On the other hand, the club leaders were passionate about their clubs and wanted to reach out with their message.They were all obviously committed to what they did: We felt that we are good at parent-child activities and that everyone can exercise our interest for riding and horses at any age, whether they are a girl or a guy, with or without a handicap.(E-mail questionnaire from the head of the Riding Club) On the contrary, when the activities of the Sensus Adult Education Association did not work out the project leader partly blamed himself and felt that his lack of presence affected the activities negatively: I could not attend the activities.It is good to attend and support parents who are insecure, parents who know me since before.(E-mail from project leader) The municipality also helped the process by providing transportation, interpreter services and financing, which made the activities free for the parents.The fact that the project leader himself participated in the activities and helped with e.g.transportation, was perceived as very important and symbolic: We have had invaluable assistance of the project leader and his private minibus.(E-mail questionnaire from the head of the Orienteering Club) Many of the participants live in [a segregated suburb] and came with the minibus and got a ride from the station from the project leader and some other participants.(Orienteering Club's website 15/11 2010) The first two clubs adjusted their activities and the times; they arranged for backup, coffee breaks, etc.Some problems, such as an unexpected number of participants, late arrivals, lack of language ability, etc, were all solved through flexibility and adjustment by the clubs: I am happy with our efforts.Sometimes we had a little too few assistants, but most of us worked on a voluntary basis and all activities were on Saturdays.So it is not so easy to get people to help out, but it worked out well in my view.(Follow-up e-mail questionnaire from the head of the Riding Club) As a contrast, when flexibility and adjustment was not shown by an association and its active members, the activities suffered: Occasion two was cancelled because the session leader became sick and nobody could substitute her.(Phone interview with Adult Education)

Experienced impacts of co-production
Both the project leader and the heads of the two successful clubs experienced that socially isolated families came in contact with each other, with other people, and also socialised outside the activities.The project leader certainly considered the project to be a success: Very successful!All three groups have started and two of them had more applications than we could manage.(Follow-up e-mail questionnaire from the project leader) Breaking the social isolation and creating new networks is something the municipality would never have been able to do through their professional social workers, the project leader concluded.In addition, the activities were perceived to have positive impact on participating immigrant parents' integration into the new society.
I think the new immigrants in the group easily melted in the school class and their language learning has gone faster plus that we recognize each other when we meet in the stores or at other places.(Follow-up e-mail questionnaire from the head of the Riding Club) The guy who had just arrived from Syria three days before last time now spoke several words of Swedish and absolutely shone from the horseback.Great atmosphere!We talked about the importance of common interests parent/child and how positive it is during the teenage years.(E-mail from head of the Riding Club, mid-activity) The clubs were happy to take this kind of social responsibility, performed it successfully and wished to continue with their efforts.New members and positive media attention along with a new social role encouraged the clubs: The greatest part was the large attendance and interest.For the two closing days we had a total of 54 participants.(…) Many were Somali who got to experience the Swedish nature, so different from that in their home country.(Orienteering Club's website 15/11 2010)

The prerequisites of successful co-production -latent categories
In the latent analysis we aimed to find answers to the question: What aspects of the process seem most essential for successful co-production to occur?We found that the balance between the two main actors was crucial: No actor was stronger or more powerful or even more "needy" than the other.For all manifest categories we found a balance: local anchorage in the community, shared interests, commitment, a flexible approach, and positively experi-enced impacts.When this balance was lost for some reason, it resulted in insufficient performance or disrupted co-production altogether.
Another crucial factor seemed to be that the parties were both appreciated for their core competencies.The municipality let the clubs make decisions about the specifics of the activities and respected the voluntary actors' expert knowledge about their fields.The clubs, on the other hand, let the project leader frame the activities, recruit participants, and welcomed the municipality's presence at the activities as a supporting force.

Results from the questionnaires
The demographic characteristics of participants who answered the questionnaire at baseline are shown in Table 1.Of the 35 parents attending the first activities at the orienteering club and the riding club n=21, (60%) answered the baseline questionnaire.Although interpreters were made available and the project leader himself tried his best to help out, filling in the questionnaires proved to be difficult for some of the parents.Others were wary of any kind of paperwork collected from them, as they were generally afraid of authorities.In terms of demographics, it is notable that 24% of participants were single mothers compared to the official statistics of the municipality where the number is 18% (Statistics Sweden, Family statistics 2012).Also, more than half of the participants were born outside of Sweden, a number way above the official statistics of 12,6% (Statistics Sweden, Population statistics 2012).Thus, the project managed to reach especially vulnerable groups.In terms of education, parents with more than 3 years of university education were overrepresented.This could either be due to more educated parents being more prone to answer the questionnaire, or to the large number of immigrants in the project, who tend to have higher educational levels than the general population (Statistics Sweden, Education and Research 2012).The 14% unemployment rate in the sample, along with the 50% who had rather high family income, indicates the mix of families, which was one of the specific aims of the project leadership.In addition, the majority of the sample indicated that they had not attended similar activities before (Table 1).
Reasons given for attending in the first place were both social (finding new friends, 33%) and health-related (exercise, 47%), although most participants simply expected to have fun together with their children (Table 1).In terms of the parent-child relationship, three parents (15%) seemed to have struggled, whilst the rest reported good relationships with their child.
At follow-up, n=17 parents (80% of baseline) answered the questionnaires (Table 2).Almost all respondents liked and were satisfied with the activities and said their expectations had been met.Parents were asked what they thought was most positive: To do things together with my child instead of just standing at the side.To meet many of the asylum seeker families, there is no natural place to meet otherwise.
We learned something new both me and my child.Nice to have had that experience together.Some participants had wanted further activities and the majority said they would continue activities at the club and/or become members (Table 2).All but one respondent felt that the activities had improved their relationship with their child.Some parents commented: I guess we have found aspects of each other that we didn't know about before.We do more things, just the two of us.
Although specifically solicited, no negative comments were made.A participant concluded: A fun and great initiative, good that it's free of charge so everyone can afford to participate.

Costs
Cost is always an important factor in providing a service and this co-production was financed by the municipality through a parenting support grant from the National Institute of Public Health.At a later date the municipality decided to make the coproduction model permanent at a lower budget.The general impression was that the funding received from the municipality was much appreciated by the associations involved: The economic stimulation increased our motivation to apply.(Email questionnaire from the Orienteering Club) Whilst the activities were provided free of charge to the participants, there is no such thing as a "free lunch".Costs included the pre-determined amount paid to the clubs (20 000 SEK per series of activities), the salary of the project leader and his assistant, costs for transportation, advertisements, and interpreters (Table 3).In addition parents' and volunteers' invested time was priced.The total cost for the activity was then divided by the mean number of participants for the six meetings to yield the cost per participant for the full activity.To illustrate the costs involved we present the example of the orienteering club in further detail, based on 2010 figures (Table 3).After having run three groups within the project, the municipality decided to continue co-production as part of its strategies and decreased the amount paid to the club to 10 000 SEK (running costs).We also calculated the costs per participant based on this sum for both clubs.
Assuming all other costs were the same, the riding club's six sessions per

Discussion
We have described two successful, and one less successful, attempts at co-production between the public and voluntary sectors.We have identified two essential components for successful co-production to occur: the balance between the two actors (public and voluntary sectors) concerning interests, commitment and experienced positive impacts and for both actors to remain within their core competencies.The project attracted vulnerable local families and managed to mix this group well with other groups of "average" Swedish families.Almost all parents who answered the questionnaires were satisfied with the number and quality of the activities and reported an improvement in their relationship with their child.participant cost 2635 SEK at the 20 000 SEK per series of activities level and 2079 SEK at the 10 000 SEK per series of activi-ties level (running costs), as the activity could only accommodate 18 parents and children.

Methodological considerations
We followed the project closely for more than a year.The materials for our analysis consisted of e-mail questionnaires, usual e-mail correspondence, minutes from meetings, official documents and on-site observations.The advantage is that the combination of different methods gave insight into certain procedures as well as to individual stakeholders' perspectives.Short e-mail answers, of course, did not provide the in-depth information that an individual interview would have provided.
To analyse data of the kind we had, content analysis (Graneheim and Lundman 2004) was suitable it kept the level of analysis close to the data (manifest level), but still allowed for the examination of possible latent content.A special strength of the study is the presence of both successful and non-successful cases of co-production.
We only used descriptive statistics and restrained from further statistical (e.g.pre-and post) analyses due to the low number of questionnaires.National statistics were used, where available, for comparison in order to give the reader an idea of the reach of this co-production effort.
To assess cost-effectiveness, one needs to have access to evaluative data relating to the effects of the interventions and thus link costs to effects.Ideally, there would also be a comparison group of some kind.Questionnaires were collected from participants, but due to language difficulties, and low comfort levels in answering questions posed by any authority, we were unable to assess changes in social support, using measures of Appraisal and Belonging on the ISEL, and subjective wellbeing in participants, as had been our original plan.The costs therefore, are com-pared to usual care to give a point of reference, but without assessing effectiveness (see next section).

The cost of parent-child activities through co-production
The cost per participant was 1497-1897 SEK (Table 1) for the six 2-hour sessions held at the orienteering club and 2079-2635 SEK for the riding club.To put these numbers into context it is worthwhile to consider that one hour of a social worker's time spent helping a vulnerable parent with social activities (finding, financing, going along, etc) costs 350 SEK without an interpreter, and 547 SEK with an interpreter.For twelve hours spent with a social worker, the cost is 5100 SEK without an interpreter, and 7464 SEK with an interpreter, per person, including parents' time.Thus, the cost of 12 hours participation in a non-stigmatising social activity for parents and children together is 20% to 35% of the cost of 12 hours spent with a social worker assisted by an interpreter.Even if we look at the most expensive activity at the highest amount paid to the club and compare it to 12 hours with a social worker without an interpreter we end up with the co-production activity costing 51% of the social worker intervention.
Thus, this model of co-production has low production costs, especially in terms of the reduced funding level provided to the clubs by the municipality after the project.One of the main reasons for this low cost is that the associations use volunteers.Also, the more participants an activity can accommodate, the lower the cost per person.For example, when 54 people participated in the orienteering club's activity, the cost per person for that session was 146 SEK.The quality of the activities was high due to the commitment of the clubs.This means that partners for co-production must be carefully selected, as has been the case in this project.One way to achieve this is to have the local associations prepare applications with their ideas based on the municipality's call for co-production of parenting support services.The local football club for example, did not apply for funds through the project: they had no economical motivation, had all the members they needed and were not going to adjust their activities or commit to special considerations for the target group.

A practical example of proportionate universality
There are compelling arguments within modern public health that to even out social inequities in a society, the key is universal access to services (The Marmot Review 2010).The problem is often that even if services are offered universally, they are not actually universally accessible to people from different social backgrounds due to barriers.These barriers can be of a geographical or financial nature, due to stigma or lack of language skills.Eliminating these specific barriers to service access is referred to as proportionate universality, as proportionately more effort is needed the greater the barriers that exist.
The activities of the associations became part of the municipality's universal offer of services to parents and children in the project.When the project leader took his minivan to drive participants to the activities, when their participation was paid for, when interpreters were made available, and a mix of families eliminated the stigma of participation, proportionate universality was exercised.
Since the idea of proportionate universality is relatively new (The Marmot Review 2010) it is of international interest to describe practical examples of how the concept might be implemented in a variety of settings.

What makes co-production work?
Sometimes it is argued that the voluntary sector lacks the sufficient skills and experience to deliver public services (Salamon 1987).This was not the case in our study.We believe that the reason for this was that the associations delivered services that matched their core competencies and were strongly supported, both financially and practically, by the local municipality.The two successful clubs showed professionalism and full control of the activities and had no problem in seeing where their responsibilities began and where they ended, something that has otherwise been described as problematic (Bovaired 2007).The voluntary sector's engagement and knowledge was thus used in optimal ways.Had they been required to engage the families from the beginning, or advertise their services effectively on their own, they probably would not have succeeded.This, on the other hand, was performed efficiently by the municipality workers.
Similar to Mottiar and White (2003), we found that the voluntary sector in our studied municipality involved citizens in the construction of a service that could not have been delivered by the municipality's social workers.This makes voluntary asso-144 Barn 30(4) The voluntary sector's engagement and knowledge was thus used in optimal ways.ciations potential public service providers, and a means of renewing service delivery in the welfare state (Pestoff 1992(Pestoff , 2008)).Instead of hiring enterprises from the private sector, the voluntary sector had the chance to deliver the service in a nonbureaucratic and, in the view of the target group, interesting way.Moreover, this kind of project creates a wider social network that interlinks different actors and organisations around a common idea, that of coproducing parenting support services, and has the potential to create mobilisation (Joshi and Moore 2002).In fact, at our last contact the project leader said he had been appointed as coordinator of future co-production efforts in the community, using different actors, but serving the same purpose: to co-produce positive activities for parents and children.
We have identified two essential components for success in co-producing parenting support services.One is the balance between the involved actors concerning interests, commitment and experienced positive impact.This balance means that the ideological framework of co-operation is not a client-performer model, but a model of mutual contribution to, and influence over, the process.The other essential factor is for all actors to remain within their specific, core competencies.The municipality provided a clear framework of coproduction including: type and goal of the activities, number of activities, target group, ages of children, and level of financial compensation.For the association at hand the activities constituted their core business.When these criteria are not met, co-production will most likely fail, as has been observed in the case of one of the associations in this study.
Local anchorage was something that seemed to contribute to the commitment of both the municipality and the associations.This kind of "sense of community" seemed to contribute to the associations' willingness to participate in a project promoting social inclusion.A question that remains therefore, is whether this kind of co-production is feasible in larger communities or in urban areas.

Figure 1 .
Figure 1.The building stones of the co-production project.

Table 1 .
Participant characteristics at baseline (completed at the end of the first session) N = 21

Table 2 .
Participant responses at follow-up (parents who answered both questionnaires)

Table 3 .
Overview of costs in the project