Institutional Review of Educational Radio Dramas: Synthesis of Findings

Commonalities

Most programs are about the same in dramatic terms, regardless of budget size. Almost all have three or four storylines per episode and eight to ten core characters, plus about 10 to 20 supporting roles. Most run between three and five storylines at any one time, although more than one 'message may be woven into one storyline to increase realism.

The average length of an episode is about 20 minutes; the shortest being 10 minutes and longest 30 minutes. Most projects have found that audiences ask for longer episodes, and have tried to adjust accordingly. In terms of numbers of episodes per week, the largest in this study is St. Lucia, with an amazing five episodes per week; the average is one or two (with at least one repeat per week).

An analysis of the fourteen case studies indicates that successful entertainment-education radio dramas share many of the following features:

Capital for start-up and sufficient funding for sustainability. Soap operas and serial dramas involve high production costs over the long-term. If they are to attract mass audiences, they need to be as good or better than programs that merely entertain. They must reflect real life so that listeners can personally identify with the characters. This entails using many diverse characters, realistic sound effects, and credible and dramatic storylines. It means investing in the best actors, musicians, editors, producers, writers, directors, researchers, and advertisers. Wide and good quality audibility normally means buying airtime on the most popular stations. Sponsorship is needed throughout the life of a project, sometimes ten years or more. Most projects in this study rely on international or UN donor funds; a few have found local commercial sponsorship. None would exist without exterior funding of some kind.

Vision. A serial drama for behavioral or social change implies some kind of vision. Sometimes, this vision is codified into a written Moral Framework[3] to which everyone - producers, funders, local government - signs on. In other cases the vision is simply understood by a team sharing common values, backgrounds, and purpose. Normally, the whole team must be fully behind the vision, since production often involves long hours, tiring field trips, and responsibilities that fall outside of regular job descriptions.

A very hard-working team. Time and again project leaders in the cases studied speak of an 'intrinsic quality' in their directors and the 'commitment' and 'dedication' of their staff. In almost every case, staff seem to be 'going the extra mile' simply for love of the job. For example, in Burundi, the writer of Our Neighbours, Ourselves has written over 400 episodes at a stretch (and risked burn-out as a result). In the case of Tinka Tinka Sukh in India, the executive producer, not only produced a 104-episode drama, but also single-handedly carried out the audience research and wrote the copy for all the advertising.

Champions who provide leadership. To inspire this kind of dedication there needs to be a strong and charismatic person who articulates and drives the vision. Such a 'champion' seems to be a sine qua non in almost all the projects studied. Many successful entertainment-education dramas might not have happened but for special individuals such as Pamela Brooke (Malawi), Jo Dorras and Peter Walker (Vanuatu) or Alma Rivera (Cambodia).[4] Singhal and Rogers both talk about the need for several 'champions' with the ability to 'leverage status power and resources' (Singhal and Rogers, 1999). In the case studies, these individuals are often the producers or directors themselves, with close involvement with the audience and in the creative process. They also need high-level contacts, and the ability to circumvent local politics and conflicts. Leaders of many entertainment-education projects are expatriates, particularly during start-up, although the presence of expatriates is not a necessary condition for success.

High-degree of professionalism and expertise. The case studies show that without good research, production and follow-up one cannot expect positive behavior change. Expertise and a high quality product generate prestige, which closes a virtuous circle where the target audience remains hooked and trusts the pro-social messages being conveyed. Not only does the audience remain loyal, but advertisers are attracted, particularly if impact research is viewed as professional and independent.

Many projects have had to build up expertise from scratch. A solid skills base is not always available in developing countries, so training often has to start with absolute basics - like typing, accounting, and English-language training. For example, Health Unlimited (Urunana, Rwanda) had to train staff in computing and report writing before they could get on to editing and story-lining.

Local on-the-ground experience. Artistry and flair do not necessarily emanate from training courses. Local experience is often the key. For example, the Wan Smolbag theater company in Vanuatu shows how 13 years of live theatre and intimate knowledge of the rural problems in their islands translate into a highly professional adaptation for radio - with very few staff having received formal training. Soul City (S. Africa), CHEMS (Cambodia) and Mediae (Kenya) are all now using their experience to train others in information, education, and communication (IEC) initiatives.

Collaboration among the many partners and stakeholders. Collaboration is a hallmark of all the entertainment-education media initiatives studied. Organizing a successful social communication project designed to bring about developmental or behavioral change among large populations requires partnerships among many and diverse individuals and organizations. These partnerships must be sustained over an extended period of time, often in a country setting very different from the originator's own. Diverse interests and individuals, from government ministers to sound engineers, from large UN organizations to tiny local clinics, must constantly be encouraged to work together if the media initiative is to succeed.

As Backer and Rogers point out, the coming-together of many stakeholders for health communication campaigns makes them unique and complex as organizations (Backer and Rogers, eds. 1993). It is on the strength or the weakness of inter-organizational links that campaigns stand or fall. Both synergies and conflicts invariably arise, especially when the originating organization is based in the Western world and the implementing partner is an indigenous organization rooted in the local setting. A high degree of trust and goodwill among all parties is essential, often depending on personal relationships and contacts. This is where a charismatic, and well-connected, individual 'champion' is often so important.

For example, Soul City (South Africa) has conceptually divided its partners into two groups:

'On the one hand the people who produce the drama (researchers, scriptwriters, funders, marketers and actors - normally the paid people) and on the other hand, people whose buy-in is needed to ensure that the drama is accessed by the audience - including gatekeepers for the topics being tackled and the owners of the media being used...The complete buy-in of all partners is crucial...these partnerships work best when both parties have something to gain....It is important to include partners as early in the process as possible. The more they understand the process and its goals, the more they will be motivated to see the project achieve its aims' (Soul City site)

Enabling Environment. An enabling environment refers to the physical, political, cultural, financial and broadcast conditions in which the media project must function. An enabling environment is as much about moral and political support from government, audience, the local press, spiritual, and community leaders, particularly on sensitive subjects such as HIV/AIDS and sexual/reproductive health, as it is about sound financial backing, good baseline data and reliable broadcasters.

One of the most crucial elements of an enabling environment is the project's relationship with national government. For example, a media initiative promoting positive health behavior should have a strong relationship with the local Ministry of Health and other non-governmental health providers. It should also understand and work around the reality of health care delivery on the ground. It is useless to recommend that listeners be tested for HIV if testing where listeners live is unavailable, too expensive or of inadequate quality.Another crucial element is audience support. This can be engendered to some extent, but certain prerequisites, such as radio access, batteries, scheduling, reception, and audio-quality, are obvious considerations. Soap operas are a 'foreign' format in many countries (Leicester P., undated), as are feedback letters, phone-ins and audience panels. Critical appraisal is something rural listeners in developing countries are generally not used to, and for which they will need much coaching and encouragement. Finally, there is the danger of possibly shocking or alienating the audience with sensitive material.

Research, Monitoring, and Evaluation. The environment can be made more receptive and enabling with a strong emphasis on research and monitoring, and audience liaison and feedback. Good formative research is needed to assess to what extent the local environment is enabling, and what barriers need to be overcome.

Other supporting media-based or interpersonal communications activities. Almost without exception, the most successful projects seem to have strong community links and other non-radio educational offshoots. The case studies provide details of these efforts and the extra stakeholders involved.

Organizational Variables

Every serial drama project is characterized by a unique set of variables, related to funding, technical support, dramatic styles, impact and so on. No two projects have exactly the same set of variables and no one set of variables guarantees success. There are both strengths and weakness for any one variable.

Variable 1: 'Sustainable' vs. 'Dependent'

All of the projects in this report are dependent on external financing to some degree, at least at the beginning. The difference lies in the source and style of that finance. Some have become self-sustaining, usually with local donors or in-country commercial sponsors. Many projects congratulate themselves when they reach this stage, (e.g., St. Lucia, Albania, and Kenya), but this is not always desirable. For example, in Malawi the producers of Zimachitika have deliberately avoided commercial sponsorship because they feel it would compromise their educational and pro-poor messages; instead they have chosen to rely on international donors.

The strength of a project that is funded by one single international donor (such as Nepal's Cut Your Coat funded by USAID and Tanzania's Twende na Wakati funded by UNFPA) is that it can spend more time on the creative process and day-to-day implementation. It has more freedom to develop storylines, without worrying about different sponsors for separate topics; it only has to nurture one relationship; and it has far fewer headaches in terms of reporting schedules and accounting minutiae. The weakness of a one-donor project is that it is vulnerable if, for some reason, it falls out of favor with the donor, or if the donor discontinues funding for political reasons. Another weakness is that there is less scope for personal involvement with a remote donor, and consequently fewer possibilities for replication or scale-up of a successful project.

The strength of a project sponsored by a variety of smaller donors is that it has not "put all its eggs in one basket," and therefore has more flexibility and security. However, it may have to find other ways to support administrative, research and publicity costs because smaller funders or sponsors often only cover direct project expenditures. If individual storylines are separately sponsored, the project can find itself constantly adjusting the storyline to fit the donor, and spending a great deal of time and energy on simply servicing the funders. Kenya's Tembea na Majira and Malawi's Zimachitika are cases in point.

An advantage of local commercial sponsorship is that it is more likely to continue over the long-term and is not usually subject to onerous reporting requirements. Any 'profits' can be used freely by the project to start other supporting activities (such as a magazine program), or simply to raise staff salaries. Most projects in this study exist in countries that simply lack the strong economic base that makes commercial sponsorship possible.

Some disadvantages of commercial sponsorship include possible pressure to advertise specific products, potentially compromising realism and storylines, and creating ethical problems and conflicts of interest. The time and energy required to attract commercial donors is another negative factor (e.g., St. Lucia's Apwé Plezi, which had to fold for lack of available staff to pursue potential commercial sponsors). Finally, the danger with social marketing-style sponsorship (e.g., advertising condoms in Bolivia's Wila/Yawa Kasta), though it has some advantages, is that it may limit the drama to do anything but urge its listeners to adopt condoms, rather than opening a discussion and offering choices.

Sustainability is not just about money; it is also about self-sufficiency in terms of organizational maturity, staff experience, and expertise. Many projects that started under the aegis of a parent organization have now become or are about to become independent, for example, Albania, Rwanda, Cambodia, and Nepal. The BBC's Tim Williams aptly described this detachment process as 'adolescent parenting,' meaning the gradual nurturing of independence by the originating organization, leading to the project ultimately being entirely run by locals and supported by local donors and/or businesses (Williams, personal communication, 2001). In cases where the project originator (an individual or an organization) is based in country and remains there as the project continues (e.g., Vanuatu, South Africa, Burundi, Malawi, Kenya, etc.), attaining self-sufficiency may be more a matter of the careful withdrawal of a key expatriate presence.

Variable 2: Local or global

Some serial dramas and soap operas have been started by individuals or small organizations as one-time efforts (e.g., Burundi's Our Neighbors Ourselves, and Vanuatu's Sarah's Family), while others are part of a string of productions, depending on a head office in the U.S. or Europe for technical, creative and monitoring input (e.g., Nepal's Cut Your Coat, a project of Johns Hopkins University (JHU), or St. Lucia's Apwé Plezi, a project of Population Communications International (PCI)).

The strength of the one-time project is freedom and flexibility. It is not tied to a pre-determined methodology, can experiment with different topics, writers and styles, and is not necessarily bound to produce 'results.' There is something fresh and creative about such dramas, which is somehow lacking in those that stick to a 'house-style.' The difficulties for one-time projects are that they must start research, staff training and writing from scratch, without the backup, training facilities, research findings, contacts and prior experience that a large entertainment-education organization can offer.

Small projects relying on the talents of one or two individuals often under-estimate the huge workload that a soap opera entails and risk "burn-out." Being part of a parent organization like PCI or JHU also has the added advantage of better, if not assured, funding possibilities.

It is interesting to note that some medium-size outfits like Search for Common Ground, Mediae, and Story Workshop, which began with one-time projects, are now expanding their work on serial dramas to more countries, building on the expertise gained during the 1990's. It is unlikely that they will grow as large as JHU, but one wonders whether expansion will compromise the freshness of their approach.

Variable 3: Short-term or open-ended

Some serial dramas are finite and exist only for a relatively short campaign period (e.g., India's Tinka Tinka Sukh and West Africa's Yamba-Songo), while others keep going for years and are more open-ended in terms of goals and structure (e.g., Albania's Rruga Me Pisha).The strength of a short-term, finite drama, from an organizational point of view, is that budgets, staffing, research, and evaluations can be planned with precision. This may make it more attractive to donors, who may not want a project with no foreseeable end and who may want to see results quickly. It may also be easier to attract interest and funding from academic institutions, as impact evaluation is much easier with a finite project.

The weakness of a finite drama is that there might not be time to develop characters able to model desired behaviors realistically, or to deal with all the facets of a complicated subject. If too much information and dramatic content is crammed into a short time period, it may leave the audience overloaded in terms of messages, and confused about characters and plot. This happened to some extent with Wila/Yawa Kasta in Bolivia, where a character became unrealistic and therefore untrustworthy in the eyes of the audience, partly because too many things happened to him too quickly.

Variable 4: Highly planned or ad hoc

Some serial dramas adhere strictly to a set of pre-determined values and messages, whereas others change emphasis as time passes, and as funding and other circumstances dictate. Cases that adhere to a highly planned format or blueprint include Nepal's Cut Your Coat, Tanzania's Twende na Wakati and St. Lucia's Apwé Plezi. Opposite examples include the relaxed approach of Burundi's Our Neighbours, Ourselves. Between these two extremes lies a middle way, in which messages; themes and storylines are somewhat planned and somewhat flexible and improvised, depending on donor demands, audience feedback, and practicalities such as staff shortages. Most of the cases studied fall into this middle way.

The advantage of the highly planned approach is that all stakeholders know exactly where they stand. If using a Values Grid[5] or a Moral Framework (the PCI approach), it is clear to everyone which messages should be conveyed and how. Sabido insisted that all stakeholders - particularly government and sponsors - should sign-off on a Values Grid, as he believed that "such agreements mitigate dependency relationships ...between producers and funders. If such agreements are not in place, producers...can be subject to direct censorship by commercial sponsors, or may exercise self-censorship" (Singhal and Rogers, 1999).

With a Design Document approach[6] (e.g., JHU's Cut Your Coat) considerable writing and production time is saved because content and messages have been determined ahead of time. Furthermore, government liaison is simplified and straightforward, and evaluation easier because it is clear exactly what was broadcast and when.

The dangers of a highly planned approach include lack of spontaneity, loss of entertainment value, and consequently loss of audience. There may be no time or capacity to incorporate audience feedback. Paradoxically, the same dangers may equally apply to an ad hoc approach, in which one writer, without guidance and direction, gets bogged down in just one or two storylines, out of touch with the issues, and relying on stereotypes - thus also running the risk of losing the audience.

Creating a serial drama involves a mixture of artistry, intuition and pragmatism. Strict adherence to plans and messages could make a drama boring. On the other hand, sometimes it may be necessary to forget about artistic integrity and incorporate a new storyline simply because a new sponsor wants it, or because an actor is taken ill. Such compromises are part of the daily reality of running a serial drama.

Variable 5: Message-based vs. issue-based

Some radio dramas are designed to impart specific messages relating to behavior change, sometimes using a Values Grid. They either consciously model desirable and undesirable behaviors through positive and negative characters, or directly state their messages through storyline, characters, announcements, songs or 'summary couplets.' They tend to prefer quantitative impact assessments. In this category are Bolivia, India, Nepal, South Africa, St. Lucia, Tanzania, and West Africa.

Within the message-based approach, a further distinction could be made between those that favor directly stated messages as opposed to modeling desired behaviors. The PCI/Sabido method has sometimes been characterized as modeling-based, whereas the JHU method is more message-based. However, in these case studies, the distinction is blurred. For example, JHU explicitly uses role-modeling in its serial drama Service Brings Reward for health workers in Nepal. Conversely, PCI talks unambiguously about "promoting locally identified messages and values" (PCI, 2000). Its projects in St. Lucia, India, and Tanzania communicate a set of messages encased in Values Grids/Moral Frameworks. Thus, this study does not distinguish between and compare a messaging versus a modeling approach, because there were no cases that strictly conformed to one or other. Instead, the two are linked and called the message/modeling approach.

Projects must constantly balance the number of messages to carry overall or at any one time. Too many messages confuse and overload audiences; however, if there are too few, the audience will feel 'hammered over the head' with the constant repetition of the same messages. Leicester argues that any 'message' must be examined from as many different angles as possible in order to be realistic and to avoid stereotypes.[7] This means spending a reasonable amount of time on each topic. CHEMS in Cambodia runs about nine messages over a six-month period and Mediae in Kenya tackles about seven separate topics over a year, ranging from marketing milk for smallholders to preventing child sex abuse.[8]

Other media projects see their task as raising issues, and are not worried about whether audiences are receiving 'right' or 'wrong' messages. The focus is on promoting discussion and informing audiences of their choices. In this category are the following case studies: Albania, Burundi, Cambodia, Kenya, Malawi, Rwanda, and Vanuatu.

Producers of the issue-based dramas are not uninterested in impact. They tend to be more concerned with generating discussion and qualitative rather than quantitative impact. They talk much less about 'behavior change'. Some have commissioned very convincing independent studies that show behavior change (e.g., Kenya and Malawi), however they do not make claims such as "contraception prevalence increased by x percentage points, due to exposure to this series." Such meticulous evaluations are more likely to be found in message-based projects that have linkages with universities, (e.g., JHU/CCP, University of Dar es Salaam, New Mexico, Ohio).[9] They have usually been designed to show impact from the very beginning, with carefully planned baselines and control areas.

As to the advantages and drawbacks of each approach, one might say that the message/modeling dramas are less realistic because they have to create morally consistent characters who will model those messages. This may be true in some cases, but not all, as countless faithful listeners of Soul City or Twende na Wakati would no doubt testify. If storylines and characters were unrealistic, radio serials would lose audiences - but the evidence from audience surveys and letters shows that they are not. Furthermore, there is firm data that listeners to these series are modeling themselves on characters, and increasing in self- and community-efficacy (e.g., see Vaughan and Rogers 2000 for evaluative evidence on Twende Na Wakati).

While some projects do not theorize about the approach they choose to take (e.g., Burundi and Vanuatu) and others stick closely to one particular approach (Nepal or Tanzania), most projects work on instinct and borrow from here and there. What matters seems to be what has worked in the past in their country context, what the audience seems to expect and want, what fits best with each individual project's leanings and preferences, and, realistically, what the donors will fund.

Variable 6: Government-linked or independent

Some media projects are an integral part of a wider campaign/national effort, typically involving national health ministries and service providers (e.g., Nepal's Cut your Coat, or St. Lucia's Apwé Plezi), while others are separate from or only loosely linked to national governmental initiatives (e.g., the West African Yamba-Songo, or Albania's Rruga Me Pisha).

Getting on the wrong side of government may be problematic, but it may also mean that what is being broadcast really matters to the audience. Programs that have occasionally been pulled off the air (e.g., Burundi and Malawi) have been able to ride it out because of close knowledge of the political situation. Political astuteness is vital, even for seemingly neutral health projects.

Variable 7: Drama as the main vehicle or as one component

Some projects have a soap opera as the centerpiece or the sole element of their education effort (e.g., Tanzania's Twende Na Wakati, and West Africa's Yamba-Songo), whereas other dramas are inserted into a larger magazine radio program or twinned with a magazine or advice program (e.g., Cambodia's Lotus on Muddy Lake or Kenya's Tembea Na Majira). Vanuatu added radio drama to a community live theater project, while South Africa's Soul City combined print-based education with related radio and TV dramas.

An important question to ask is can an entertainment-education drama be effective on its own, or does it need factual back-up or supporting activities to have an impact? Most researchers suggest that radio alone will not bring about behavior change (e.g., Adam and Harford, 1998, Norrish et al, 2001). However, the case of Twende Na Wakati may contradict this idea, since it is a stand-alone project, which showed significant behavior-change data (Rogers et al, 1999, Vaughan et al, 2000, etc.).

The case studies suggest that what matters most is responsiveness to audience wants and needs. If, for example, audiences demonstrate a desire for more factual information, an obvious step is to provide a back-up magazine program. If the technology is available, phone-ins or hotlines may be possible (e.g., Cambodia and St. Lucia inter alia). In some cases, supporting activities happen spontaneously, as in Albania, where Radio Tirana has taken the initiative to start and fund a discussion program based on Rruga Me Pisha, quite independently of the project. In India, listeners inspired by Tinka Tinka Sukh started their own self-help groups, and even schools. This kind of spontaneous follow-up is perhaps more indicative of success than any quantitative evaluation, as it shows the project has become truly indigenous.

Variable 8: Large or small

All serial dramas are relatively large undertakings, which explains why there are very few, if any, long-running serial dramas run by community or local radio stations. The small end of the range in these case studies starts in Burundi with the soap opera, Our Neighbours Ourselves, which has one full-time writer and two part-time producer/technicians, and functions on a budget of a few tens of thousands of dollars per year. At the other end of the scale are projects such as Nepal, Tanzania and Rwanda, which employ up to twelve full-time staff, and have budgets three times larger than that of Burundi. (In most of these cases, the larger projects also run complementary magazine programmes, trainings, and audience outreach - so their annual budgets are not directly comparable to Burundi.)

The strength of a larger project is that it can do more research, evaluation and audience outreach. It runs less danger of going off-air periodically because it has the resources to cover staff absences. It also has more resources to promote itself to the audience and to potential donors. However, the larger a project becomes in terms of numbers of funders and other stakeholders, the more complicated it is logistically. When increasing size requires sub-contracting different sections of the production process, the project could become merely a managing agency and lose contact with the creative process.

Efficacy, evaluation, and cost-effectiveness

In attempting to assess the efficacy of different radio entertainment-education models, one must first consider how efficacy is measured. Is it based on how funny it is, how popular, how relevant, or by how many percentage points it changes the audience's behavior?

Ultimately, all pro-social projects aim for some kind of developmental change, but generally they can be divided between a) those that aim for information, discussion and choice around an issue, or b) those that aim for quantifiable behavior change. Quantifying behavior change can be problematic. For example, how does one measure changes in social responsibility? Sexual and reproductive health behaviors may be difficult to measure because of the delicacy required in asking the necessary questions. Furthermore, some claim that those individuals most likely to change behavior are also those most likely to listen to a media campaign.[10]

In order to attribute significant changes in health behavior or dramatic increases in knowledge to the effects of the radio drama, the methodology must be sound. Some suggested elements of a good evaluation methodology for behavioral and social change radio dramas include:

• A thorough, representative baseline study, before the intervention begins
• The ability to properly isolate those who have not been exposed to the drama
• The ability to control for other biases relating to educational, socio-economic status, etc.
• The ability to control for other educational and media influences that might show the same effect
• A rigorous monitoring of the same audience members' knowledge, attitudes, and practices (KAP) over time

This study has not attempted to dissect methodologies or to judge the reliability of each project's impact studies. This would have required detailed scrutiny of over twenty separate evaluations, cross-checked against other measurements of the reality on the ground, which neither money nor time allowed. Some claims made by impact studies in this report may not stand up to thorough scrutiny. Some projects use vague terms in their literature, such as 'regular listeners' or 'independent evaluation' or 'significant impact.' Each project has its own standards and it was not always possible to determine the exact definition of these terms in each context. The discrepancy between theory and reality is often difficult to pin down. For example, a project may present what looks like a very organized production cycle, or an extremely positive impact assessment, whereas the reality may well be different. A few, such as Twende Na Wakati, appear to have applied a well-designed impact methodology. Other projects may be equally effective, but do not have the tools, money or structure with which to measure their impact. Therefore, projects are taken at their word and no conclusions are reached as to which projects are more or less 'efficacious' than others. The emphasis is on the organizational and structural features of the programs.

Finally, any discussion about efficacy and impact inevitably leads to consideration of audience numbers and cost per listener. Although radio is more cost-effective than television and other media in developing countries, entertainment-education radio projects are under increasing pressure to show cost effectiveness.

This report has not attempted to assess the relative cost-effectiveness of projects for three reasons. First, budgets were difficult to analyze. It was often unclear what proportions of budgets covered expatriate salaries or research and evaluation. Often a budget covers much more than just writing and production; it may include publicity, accompanying spots and magazine programs, listening clubs and back-up print material. Furthermore, it was sometimes unclear how much of the official budget included overhead at the originating organization's head office and whether or not the cost of air-time and other in-kind contributions were factored in.

Second, it seems morally indefensible to compare projects simply on a cost-per-head basis. For example, was Tinka Tinka Sukh, which reached 40 million people, a more worthwhile project than Apwé Plezi because it only reached 40,000 people? Why should remote or small island states be deprived of development projects simply because they are more costly per listener? Third, it is invidious to attempt to quantify the benefits of an entertainment-education project simply in monetary terms. One cannot quantify pleasure gained or information imparted per person. Entertainment-education practitioners should resist being pulled into a cost-effectiveness argument.