Apr 18, 2011 - mobile phones, and low-cost video camcorders. ...... on Google's Android 2.0 OS, itself based on Linux),
Comparative Laboratory Study of 12 Devices for Agriculture Extension Phase 1 Report of Devices for Agriculture Extension: A Comparative Landscape Study Kentaro Toyama (D-Rev and UC Berkeley) April 18, 2011 This report is available at http://www.d-rev.org/projects/accessforagriculture.html. Please direct inquiries to kentaro_toyama(a)hotmail.com.
Table of Contents Executive Summary....................................................................................................................................... 2 Background ................................................................................................................................................... 3 Agriculture Extension ................................................................................................................................ 3 Information and Communication Technologies for Development ........................................................... 5 Electronic Technologies in Extension ........................................................................................................ 8 Other Domains of Development ............................................................................................................. 10 Caveats .................................................................................................................................................... 10 Study of Devices .......................................................................................................................................... 12 Methodology........................................................................................................................................... 12 Anticipated Users .................................................................................................................................... 12 Extension Scenarios ................................................................................................................................ 13 Choice of Devices .................................................................................................................................... 16 Conclusions ............................................................................................................................................. 17 Device Survey .......................................................................................................................................... 20 Other Devices .............................................................................................................................................. 47 Report Contributors .................................................................................................................................... 48 Glossary ................................................................................................................................................... 49 References .................................................................................................................................................. 49
Executive Summary The commoditization of consumer audio-video electronics offers a tantalizingly low-cost channel for disseminating information to under-educated, possibly illiterate, populations such as those living in remote rural areas in developing countries. Examples include devices such as Dictaphones, mid-tier mobile phones, and low-cost video camcorders. Agriculture extension is the focus of this study, but analogous applications are seen for instructional scenarios in hygiene and healthcare, literacy instruction, and microfinance. We surveyed 12 available devices in terms of cost and features, through laboratory analysis that did not involve investigations in the field (an ongoing follow-up study tests 4 devices in low-income farming contexts). Devices were analyzed with respect to their cost, ruggedness, usability, ease of content authoring, and other factors which are critical for large-scale usage. This report overviews the conclusion from the study, preceded by several caveats regarding the use of electronic technologies for agriculture extension. Among the caveats: 1) Agriculture extension is best thought of as a process of deep education, which requires not only the thin dissemination of information and knowledge, but fundamental changes in behavior and habit which are not readily transmitted outside of a strong institutional framework. 2) Technology is an ongoing cost, of which cost of hardware is a relatively small part. Costs of maintenance, repair, upgrade, training, and so forth rapidly add up. 3) Automated processes are not necessarily cheaper or better in international development contexts, as the human touch can provide important elements of trust and hand-holding that technology alone cannot. Also, with low costs of labor, technology is not always cheaper. 4) There is no single “perfect” technology even for the purposes of agriculture extension. Different scenarios call for markedly different features. The study’s conclusions include the following: a) With respect to electronic technology use, agriculture extension can be broadly classified into several usage contexts, including one-to-one or small-group instruction, large-group presentations, individual or household self-study situations, peer-to-peer content production, and formal content production. b) Small-group presentations require rugged, portable devices, but demand less in the way of feature richness or ease of use, as the assumption is that an instructor or mentor figure is present. Large-group presentations require high-volume, high-quality audio and/or large-screen video. Self-study situations require what small-group presentations require, but additionally must be very low-cost and have a simple user interface. c) There is a very consistent tradeoff between device cost and features in the obvious way, with higher cost devices providing richer features. d) Among devices that make the optimal feature-cost tradeoffs are netbooks, smartphones, lowcost video recorders, and low-cost audio recorders.
Background The study reported here occurs in the context of several streams of effort in international development:
Agriculture extension: Dissemination of expert agricultural knowledge and effective practices to actual farmers. This study focuses on agriculture extension for financially poor, smallholder farmers in the developing world. Information and communication technologies for development (ICT4D): Interventions that apply electronic technologies to various domains of international development, as well as studies thereof. Electronics in agriculture extension: Intersection of the above.
Below, we provide brief overviews of these areas with respect to their history and dominant current practice.
Agriculture Extension Of the 1.1 billion people who now survive on less than a dollar a day, 800 million earn their primary livelihood from small farms in developing countries. Information about up-to-date farming practices typically does not reach smallholder farmers who live in remote rural areas, and the farmers themselves have been deprived of a broad education that would allow them to find and absorb such information on their own. Yet, the right information absorbed and applied correctly can double or triple income in many of these households. Agriculture extension – the dissemination of expert agricultural knowledge and practice – is, thus, at once among the major challenges in rural development and a great opportunity. Classical extension encompasses a full supply chain of knowledge from research at agriculture universities, training of extension officers (also at universities), educational broadcasts via radio and TV, and direct training of farmers. In this study, we focus on “last mile” extension, which involves interaction with the farmer who actually implements practices. Last mile extension is often cited as the greatest challenge in extension (Feder et al., 2001). Extension is a challenge even in developed countries with literate farmers, organized agricultural groups, and well-developed communications infrastructure, so the problem is all the more greater in developing countries where a number of systemic issues are commonplace:
Farmers lack self-efficacy, the sense that they can control outcomes through changes in their own behavior. Farmers are often poorly educated; many are illiterate or semi-literate. Farmers have little disposable income or free time, and will not invest in expensive or timeconsuming forms of education. Farmers do not generally self-organize, and are socially and geographically dispersed. Formal programs for extension (e.g., government agriculture extension officers) are often unavailable, under-funded, and/or ineffective. Farmers receive conflicting or incorrect information from sources that may not have their interests in mind (e.g., fertilizer/pesticide salesmen).
The modern roots of agriculture extension begin in the 1970s, when the World Bank backed a methodology known as “training and visit” (T&V, or “classical extension”). In T&V, individual agriculture extension officers, generally hired by the state, visit agrarian villages and evangelize more productive farming practices on a one-to-one basis to farmers (Anderson et al., 2006). “Green Revolution” technologies involving chemical fertilizers and pesticides have been the dominant content of messaging, though extension can involve non-Green-Revolution techniques as well. In countries such as India, agriculture extension was considered widely impactful, and although the merits of the Green Revolution’s long-term net effects remain debated, it is credited with having turned India from a net importer of food to a net exporter. Classical extension remains the dominant form of extension, but funding has decreased over the decades, partly due to uncertain ongoing impact and high costs (Antholt 1998). India’s extension force, for example, has shrunk to 100,000 officers, which is small when considering the country’s 610,000 villages, with an average population of 1000 people. Impact is difficult to establish, due to confounding factors such as soil erosion, subdivision of land plots, and introduction of new technologies and techniques. Furthermore, with fewer resources, extension has become a more challenging activity. Too many households are assigned to a single extension officer, and individual officers have difficulty establishing rapport with their clients (Antholt 1998; Feder et al. 2001). China is among the rare exceptions in having an effective extension force, with one million officers who are reputed to have near command-and-control authority over what farmers do. More recently, Farmer Field Schools (FFS) have risen as an alternative to T&V (Tripp et al. 2005). An FFS is a semi-formal, weekly gathering in which a small group of farmers observe and evaluate possible agricultural interventions on one individual’s farm. The FFS model is credited with spreading Integrated Pest Management (IPM) practices in Asia by graduating more than four million farmers in 50 developing countries (Van den Berg & Jiggins 2007). The success of the FFS extension model has encouraged donors to put extension back on the development agenda, but impact at scale is still limited – over the past 15 years, FFS programs in Asian countries have only covered one to five percent of all farm households. Additionally, some economists argue that the model is not financially sustainable after foreign aid is withdrawn (Feder et al. 2004; Longley 2006). Apart from the dominant modes of extension, surveys reveal how farmers typically learn about agricultural practices. A 2005 national sample survey in India, for example, revealed that the most common way that farmers learned about new practices was through their neighbors, with ~17% citing it as a form of learning. Product salespeople and the radio were each cited by 13% of farmers. Television and newspaper followed with 9% and 7%, respectively, while extension officers were cited by only 6%. Finally, although this document focuses on agriculture extension, it should be noted that extension alone is rarely sufficient to transform farmer livelihoods. Almost always, there needs to be concomitant attention paid to credit, fair markets, formation of cooperatives, and so on.
Agriculture Extension as Education Though agriculture extension is generally considered a subfield of the agricultural sciences, it should be emphasized that the core activity is fundamentally a kind education. (The word “extension” itself comes from education programs offered by universities for non-students.) This cannot be emphasized enough in the context of agriculture, where it is all too easy to believe that extension is simply a matter of information or technology “dissemination,” as if the mere spreading of superficial information or relevant technology is enough to change agricultural practices. At heart, most agricultural extension requires not only new knowledge or technology, but also deeper behavioral changes among farmers. And, such behavior change typically requires a multi-front effort at the individual and societal levels, akin to the efforts against smoking or for seatbelt wearing in the United States. As a consequence, theories of pedagogy can shed light in understanding and taxonomizing the activities of extension, though these are rarely brought in by agriculture specialists. For example, it is illuminating to consider extension in terms of Bloom’s taxonomy for educational goals. Bloom posited six levels of educational goals for cognitive skills, from knowledge, to comprehension, application, analysis, evaluation, and synthesis (Anderson & Sosniak 1994). For last-mile agriculture extension, it seems clear that knowledge, comprehension, and application by the farmer are the critical goals. (Analysis, evaluation, and synthesis are the strengths of universities and research labs.) Bloom’s taxonomy highlights the fact that extension cannot end with the mere communication of information; for impact to be felt, agriculture extension must ensure true comprehension, and ultimately, application of knowledge. This point is essential as information and communication technologies are considered for extension. For the poorest, least educated farmers, extension must be recognized to have components of remedial education. These farmers can benefit not only from straightforward transmission of individual agricultural practices, but from an improvement in their self-efficacy and self-confidence. A lifetime of experience that whims of weather, pests, and local markets have greater sway on their income than anything they can do for themselves can lead to what psychologists call “learned helplessness.” Learned helplessness extinguishes any propensity to try to improve one’s situation, thus erecting an internal barrier to growth. Successful extension is able to turn this around by actively reaching out to farmers and demonstrating in gradual steps that self-initiated changes in practice can result in positive outcomes.
Information and Communication Technologies for Development Over the past fifteen years, there has been increasing interest in applying the power of recent technologies such as the PC, the mobile phone, and the Internet to international development. Dubbed “information and communication technologies for development” (ICT4D), the interventionist side of the field asks questions such as the following: Can PCs be meaningful used even in schools too poor to maintain toilets? How can mobile phones be used to decrease market inefficiencies? What kinds of devices are the effective for conveying agricultural concepts to farmers?
History and Trends1 Though precursors of ICT4D can be traced to the era of radio and television, the current strain of activity can be traced to the commoditification of digital electronic technologies. Compared to development, which as an international activity originated in the mid-1940s, the consumer digital electronics industry has its roots in the 1970s, and it encompassed mainstream consumers only in the 1980s and 1990s. This forty-year gap brings with it its own dynamic, as the young-but-successful technology industry brings a certain bluster and brashness to a community sometimes criticized for its failures and inefficiencies. As with many areas of development, ICT4D involves multiple sectors and multiple disciplines. Government, universities, foundations and non-profits, and multilateral organizations have all been active, and academics from disciplines such as anthropology, sociology, economics, political science, design, and engineering have engaged. But, while these traits are common to international development as a whole, in ICT4D, there is unique attention from the computing and communications industries. These industries bring with them technical expertise in cutting-edge technology, but also often immense financial resources and public visibility. They have thus rapidly become an audible voice in development. For its part, the international development community has become increasingly interested in digital technologies, with opinions ranging from a few who believe they are important for communication among development agencies, to others who advocate them as panaceas for poverty. Historical swings also occur as focus shifts from one technology to the next. As with many debates, a moderate view is perhaps most justified. For example, the well-known Millennium Development Goals (MDGs) put forth by the United Nation explicitly urges that we should “…ensure that the benefits of new technologies, especially information and communication technologies […] are available to all” (United Nations, 2000), but this directive is embedded in a sub-clause of one of its eight main targets, all of which are along traditional non-technology development concerns – poverty and hunger, equality between genders, baseline healthcare, primary education, and so forth. Implied is the thesis that technology is important, but that it should be subservient to larger, more basic goals. What constitutes “information and communication technologies” in ICT4D? A strictly literal interpretation might allow everything from the printing press to talking drums, but recent usage focuses “ICT” on modern electronic technologies, with PCs, mobile phones, and the Internet at the center. Of particular relevance to this study is the historical arc of ICT4D technologies. A very rough timeline of this activity is enumerated below. Dates should be taken as approximate and indicating intervals when corresponding technologies peaked or dominated ICT4D activity (and not that the activity was absent outside of the interval).
1
1960-1990 (Analog electronics): Though not called “ICT4D,” most work that applied electronic technology to development in this period focused on landline phones and broadcast technologies such as radio and television. 1990-2000 (IT for non-profits): Early ICT4D was focused on the use of IT systems in non-profit organizations and multilaterals. Most of this work concentrated on the introduction of PC-and
Some portions of this section have been adapted from an introduction to ICT4D that appeared in IEEE Computer magazine (Toyama & Dias, 2008).
server-based systems into development organizations, paralleling similar developments in corporations. This computerization of development organizations can be considered the “first wave” of ICT4D. 1995-2005 (Telecenters): Renewed excitement in ICT4D came with the advent of the telecenter, in which PC-based centers, operated either as small, local entreprises or community centers, were expected to deliver development outcomes through access to PCs and the Internet. 2005-present (Mobile phones): With the dramatic penetration of mobile phones in the developing world, the development community has become excited about using them for a wide variety of applications. The telecommunications industry has fed this excitement in a significant way, and the excitement continues as of this writing.
Several points emerge from this timeline. First, ICT4D efforts often echo the dominant trends in the electronics industry in developed countries, with a slight lag. This is not surprising, and is likely due to several factors. The electronics industry tends to move quickly from one generation of technology to the next, as old technologies and data formats become outdated. For example, anyone hoping to use audio technologies today will be all but restricted to digital technologies; it is increasingly difficult to find cassette-tape technology, and all but impossible to purchase eight-track devices. The nature of massmarket economics means that the most popular forms of electronics will also tend to be less expensive, compared with less popular technologies with similar features. Those involved in development activity will also only see the current set of technologies on the market, naturally limiting the range of creative solutions. Second, the trend in ICT4D, again following that of the larger electronics industry, is for devices to become increasingly digital, increasingly affordable, increasingly portable, and increasingly feature-rich. These trends have several consequences for development. Digitization has made content easier to produce, store, query, and disseminate, particularly on platforms such as the Internet. This tendency has brought with it the moot rhetoric of “democratization” that accompanies discourse about the Internet. Affordability has put more devices in the hands of poor populations as never before. For instance, secondhand mobile phones can be bought for US$5 or less in many markets, and in 2010, there were more than 5 billion mobile phone accounts according to the ITU (ITU 2010), more than the total population of the world over the age of 20. Affordability and portability together have contributed to greater mobility of the technology. TV sets and film projectors were heavy and difficult to transport, especially over unpaved roads; today, however, handheld video playback devices and pico projectors make it trivial to bring high-end multimedia to even the remotest village. Finally, portability and featurerichness have contributed to increasing personalization of technology. Some observers note the rich personal expression that occurs on mobile phones, for example, even in spite of extensive device sharing (Ling & Donner 2009), and this, in turn, contributes to a robust demand for some technologies among even the poorest populations. Third, ICT4D experiences phases and fashions, as others have noted for development overall (Easterly 2001). Initial excitement around a technological idea often leads to hyperbole about its potential. Donor agencies establish task forces and new groups dedicated to the new technology. But, as interest grows, reports of the technology’s limitations also start to come in, and the trend crests. Afterwards,
excitement tapers off, but diehard supporters as well as naïve newcomers keep the concept alive. This cyclical process in development is amplified by the crests of must-have gadgets and troughs of obsolescence in the technology industry. Beneath the dominant trends in ICT4D noted above, there has also been less prominent, but ongoing experimentation with a wide range of specialized devices. In the decades straddling the turn of the millennium, a range of digital technologies became commoditized and experienced dramatic drops in cost. Wireless chips, GPS capability, imaging and video technology, and so forth, are all now within a price range that makes their use in local development projects feasible. Long-distance WiFi to remote villages (Surana et al. 2008), monitoring based on digital cameras (Duflo 2006), and local video production as a component of agriculture extension (Gandhi 2009) have all seen successes in the last decade. Though these technologies are reaching asymptotic lower bounds in price, many have settled at levels that present broad potential for further evolution of ICT4D. The study reported in this paper looks at devices in this category that seem the most practical for the purposes of educating smallholder farmers.
Electronic Technologies in Extension Spurred by reinvigorated interest in agriculture in the global development community, and in part by recent ICT4D activity, stagnant issues in agriculture extension have returned to the fore. In particular, there has been increasing experimentation in the use of electronic technologies specifically for extension. Technology is hoped to help deliver knowledge customized either to the individual or a local community, or to mitigate the costs of transport (e.g., of agriculture experts to rural villages and households). Some examples are described below. AAQUA (“Almost All Questions Answered”) is an Internet system designed at the Indian Institute of Technology, Bombay. It allows anyone with access to the Internet to post agriculture-related questions which are then answered by expert paid staff. A database of past questions and answers are also retrievable through a straightforward search (Ramamritham et al. 2006). Typical farmers are assumed to gain access to the Internet through rural telecenters, which are a kind of development-focused Internet café described in the previous section. Although the system is versatile in permitting anyone with the ability to access and use the Internet to ask questions of experts, the reality among smallholder farmers is that most are neither aware of the service nor proficient in Internet use, and telecenters are themselves vulnerable in their long-term viability. Finally, the very fact that farmers must come to AAQUA, makes it less likely that this kind of system will directly have widespread impact. Lifelines is a system run by One World South Asia, where farmers call in their agriculture-related questions by phone, rather than on the Internet. Human operators answer these calls, log the questions, and then determine whether the answer already exists in their database, or whether an expert must be consulted. In the former case, the caller receives a notification by text-message that their question has a response that can be heard at a particular number. In the latter case, the expert records the response, which is then entered into the database, and again the caller receives a notification. Over time, the database is expected to answer more and more of the questions, and the need for expert time is
decreased in the long run. The system seems to appeal most to somewhat wealthier farmers with medium plots of land. In contrast to models where farmers must pull for information, eSagu, a project headquartered at the International Institute of Information Technology in Hyderabad, advocates a “query-less push-based” model in which field coordinators surreptitiously take digital photos of crops and farm plots and send them to a processing center manned by agriscientists to diagnose. Prescriptions are then printed out and distributed or read out to farmers by the coordinators. Although initial trust in the system is a hurdle, once farmers are satisfied that the advice is helpful, they are likely to become ongoing clients. An advantage of eSagu is that the system actively reaches out to farmers, which is all but necessary to reach the least capable farmers. Grameen Technology Center’s AppLab project in Uganda provides agricultural information over SMS text messages on mobile phones, mediated by Community Knowledge Workers (CKWs) who are an integral part of the system. The CKWs are conceptual descendants of “Village Phone” operators that earlier ran small enterprises as walking phone booths (they sold access to mobile phone services in villages). Now, their goal is to find a meaningful role in their villages as information brokers. The CKWs are not formal extension agents, but they fill some of the roles of an extension force, through active outreach to farmers, and by active as human portals for agricultural information. The four projects described so far perform extension in a one-to-one model, but it’s also possible to address farmers en masse or in groups. Farm Radio International, a non-profit organization based in Canada, supports a network of community-radio broadcasters for agriculture extension in Africa. Community radio is often called “Africa’s Internet,” and it remains a constant feature of the development landscape in Africa. The relative intimacy of community radio permits the inclusion of local farmers in Farm Radio broadcasts, which builds the trust necessary for farmers to accept the information broadcast (Amt 1986). Digital Green runs a model in which groups of farmers watch and discuss videos with extension content that has been produced in their local area (Gandhi et al. 2009). Video production occurs by following extension officers or NGO staff as they perform one-to-one extension, and then video-recording actual farmers as they learn and demonstrate agricultural practices. Videos are then shown to groups of farmers, called together by local mediators hired to screen videos and provoke discussion among the audience regarding the content of videos. In all of these projects, the technology tends to play two roles consistently. First, the ICT in these examples decreases the costs required to bring expert agricultural information to farmers in villages. In AAQUA and eSagu, this is done via the Internet or dial-up (though the need for such infrastructure raises the question of how it is maintained); in Farm Radio, community radio broadcasts fulfill this role; and in Digital Green, digital content is carried by extension staff. In all cases, the need for expertise, in the form of an agriculture extension officer, to travel is diminished. Second, technology permits the customization of knowledge for the individual or the local community. With AAQUA, eSagu, AppLab, and Lifelines, the questions and answers are uniquely tailored to specific
questions. Most of these systems take advantage of the redundancy in questions by maintaining a database of questions and answers, and shrinking the need for expert attention. In projects like Farm Radio and Digital Green, the information delivered is more coarsely customized at the community level, but due to consistency in local geography, weather, and agricultural practices, it is nevertheless far more appropriate than state broadcast programs. Commodity electronics thus permits small-scale targeted broadcast, which is more effective than large-scale broadcast television and radio. In these projects there appear to be two different paths to success. The first is to aim at wealthier, better educated farmers, who tend to have more self-motivation and capacity to undertake new practices. The second, which is necessary for less self-confident farmers, is to ensure that there is strong organizational support for the extension effort, regardless of the technology.
Other Domains of Development As noted earlier, extension is ultimately a kind of education. As such, some of the lessons learned in methods and tools for agriculture extension may also be applicable to other domains of development with some modification. A detailed discussion is beyond the scope of this document, but applications in formal education, hygiene and healthcare education, and vocational education could benefit from the appropriate use of technologies, provided the right combination of target audience and organizational support.
Caveats It’s worth issuing a number of caveats in interpreting the study described in this document. First and foremost, the only reliable test of a technology is in the context in which it is expected to perform. Just as it is impossible to predict whether a new technology product will do well in the developed-world consumer market, so, too, is it difficult to gauge how a technology will do in for a development scenario – the parameters for success are different, but the environment is at least as complex. More accurate predictions could be made with more testing, with the technology embedded in the expected scenario, over the long term, at the desired scale, and in the appropriate cultural context. This means, however, that there is a limit to how much can be known short of trying things in the full context. The descriptions made in this document are thus, at best, good-faith guesses based on the authors’ experience with similar or related technologies and with on-the-ground international development. Among the greatest pitfalls is to overgeneralize analysis, particularly across populations. External validity is never guaranteed. Although there are similarities among impoverished communities, especially when compared to wealthier populations, a range of geographic and cultural differences may cause very different reception to technology. Second, technologies must be considered with regards to their ongoing, total cost of ownership (TCO), not the hardware device cost. A common misperception with regards to information technology is that it is a capital investment that once bought, does not require additional expenditures. Nothing could be further from the truth. Hardware costs must be considered ongoing costs, as there is continuous need to
maintain, replace, repair, and upgrade them. Because hardware is tied to the nature of the technology industry, rapid obsolescence remains a challenge. Then, as ongoing costs, hardware costs are typically a mere fraction of the total costs of keeping them supported and integrated into an effective extension effort. Third, though developed-world intuitions suggest that technology is cheaper or better than human labor for certain tasks, this intuition must be questioned consistently in developing-world applications. In developing contexts, ICT use is often intermediated by people who, either formally or informally, serve as operators of the technology. These people have or gain trusted relationships with the end beneficiaries of the technology; that trust is not easily replaced by a purely technological system. In addition, because labor costs can be very low the cost to involve such mediators may be lower than the cost to fully automate a system with technology. Careful examination with regards to the need for technology is necessary. Fourth, there is no single technology that will serve every extension scenario. Current hype around mobile phones recommends them as the device of choice due to their ubiquity. But, while phones are useful for many purposes, they also have their drawbacks – their small screens make them inappropriate for large-group gatherings or interaction with, say, spreadsheets; makes and models have proliferated to the extent that a single software platform is increasingly out of reach; telecoms are unrelenting in the commercial services they ply through them, etc. Similarly, radios and TVs are affordable, but they don’t allow two-way communication. PCs are versatile, but often expensive to maintain. And so on. Each device has its pros and cons, and these must be evaluated with respect to the exact application. In particular, the desire to use a technology for the sake of it, or because of hype around it, should be checked. A better approach is to focus on solving specific problems, and in some cases, a natural solution may require technology. Finally, it should be noted that despite frequent claims that “information is the bottleneck” to agriculture extension, the reality is that a host of challenges make extension difficult. Many farmers do not have basic education and achievement-oriented mindsets; trust must be established between extension officers and farmers; knowledge must be translated from the jargon of agronomy labs to the vernacular in the field; and, so on. Technologies are best seen as magnifiers of human or institutional intent and capacity (Toyama 2010). As such, technology should be viewed as amplifiers of strong institutions already accomplishing extension, and not as an effortless means to scale without good institutions.
Study of Devices Methodology 12 devices were surveyed for their appropriateness for agriculture extension in the developing world. The objective was to gain a high-level understanding of the existing commodity electronics that are available, and to do a comparative analysis of their costs, capabilities, complexity, and ultimately, potential for developing-world agriculture extension. The study presented here is analytical in nature, with no primary research in the field. Actual farmers were not consulted specifically for this study, although the study does take into account the direct experience of the authors and other researchers who have investigated the use of electronic devices for agriculture extension. Several of the devices have, in fact, undergone preliminary trials for agriculture extension in a handful of contexts. Where appropriate, those studies are referenced. We note, however, that the ultimate test of any technology is in the field, and under conditions similar to those of its expected application. Although this report provides an initial guide for understanding how various multimedia technologies might play out in the field, we discourage its use as a definitive statement on how individual technologies might work in reality. An ongoing follow-on study conducts more in-depth studies with actual farmers, using three of the devices examined here.
Anticipated Users Agriculture extension involves a long chain of people, including agriculture scientists and extension coordinators, but in this study, we focus on “last mile” interaction. Last mile interaction involves two kinds of parties which are our primary concern: the farmer and “the trainer” (a generic term we will use for people who interact directly with farmers). Both of these categories can be diverse, even when the focus is on low-income, developing-country agriculture. Farmers There are approximately two billion farmers falling under the World Bank’s poverty line of less than $2.25 income per day, and of these, 800 million are “dollar a day” farmers earning less than $1.15. Apart from their financial poverty, these farmers have in common that they depend on small plots of land of roughly 3 acres or less, with 1-acre plots being common and ½-acre plots not at all rare. Statistical descriptions, however, can hide both commonalities and variations that exist among farmers. Most will have had less than a secondary-school education, and illiteracy as well as the complete absence of formal schooling is common. At the same time, there are communities in which the average farmer has a solid secondary school education, with fluency and literacy in two or more languages. Low levels of education can also be a cause of poor hygiene and healthcare, at least relative to the standards of modern medicine. Many low-income farmers also alternate agricultural activities with other work, especially if climate constrains when productive farm work can be accomplished. Those living in or near forests will often gather “forest goods” such as wood, mushrooms, honey, etc. Those living in countries with robust urban
economies may seasonally migrate to cities as daily wage workers. Temporary jobs may also come from government programs hiring construction workers or road builders. Perhaps most relevant to agriculture extension, marginal farmers often lack self-efficacy, or the confidence in their ability to change their life or their farming practices for the better. A “learned helplessness” can be the norm, where farmers appear uninterested in improving their own conditions.2 Thus, a range of characteristics beyond simply poverty or lack of knowledge contributes to farmers’ ability to absorb the benefits of extension. Last-mile extension must necessarily be sensitive to all of these issues – some global, some local – if it is to be effective. Trainers For the purposes of this document, we will use the label “trainers” to indicate anyone who interacts with farmers directly in extension. These may include government extension officers, NGO staff, community workers, community volunteers, agriculture input salesmen, or farmers who take on instructional or mentoring efforts. They may range from those who have university degrees in agriculture to those who have minimal education (though programs which involve local community workers will typically make efforts to identify people with a secondary education, or any case, more education than the farmers they interact with). Usually, though not always (especially in the case of private-sector salesmen), their incentives are aligned with an attempt to aid farmers. We explicitly exclude from consideration those who are more than one degree of separation removed from farmers, or who do not interact with farmers directly. Therefore, those who work primarily as trainers of trainers, university professors, extension managers, TV program hosts, etc., are beyond the scope of this study. Exact numbers are difficult to compile, but worldwide, perhaps several million trainers work with smallholder farmers. China alone has a state-run force of one million extension agents. India has ~100,000 government extension agents, and perhaps many more who work for state-level schemes or NGOs. Many other South Asian countries have government extension systems. A few African countries, such as Ghana, Uganda, and Ethiopia also have (or plan to instate) state-backed extension services, although many are not fully functional.
Extension Scenarios It cannot be overemphasized, that for an activity as diverse and complex as agriculture extension, a device can only be judged to be “appropriate” with respect to a particular scenario. Thus, a device that may be effective for presenting content in farmer field schools may be ill-suited as a study aid for individual farmers to use in their homes, and vice versa. And, a similar statement could be made for channels by which devices are distributed. A device used by NGO staff to support discussion of agricultural methods in training sessions may not be suitable as a publicly financed village bulletin board. 2
In some quarters, this is even interpreted as laziness, but any such judgment must also be seen in the context of an environment in which factors such as weather, pests, seed merchants, middlemen, and regional commodity prices often have much more impact on a farmer’s income than anything he or she does.
Because technologies in isolation almost never contribute to meaningful development objectives, evaluations of technologies must necessarily take place with specific application scenarios in mind. Electronic devices could support agriculture extension in a variety of usage scenarios. We first describe five ways to classify last-mile extension activities, and then identify what we believe to be the most common modes of usage. The analysis of devices that follows will make frequent reference to these scenarios. Technology-mediated instruction can be classified along five different dimensions, as below. (These categories can be applied to other instructional domains, but we discuss them here in the context of agriculture extension.)
Push versus pull: In “push” models of extension, an agency external to farmers and their communities intentionally push agriculture education onto farmers. In “pull” models, farmers actively seek out education. Broadcast radio shows, or classing T&V extension can be thought of as push models; telephone helplines, as well as Grameen Technology Centers’ Community Knowledge Workers, can be thought of as primarily pull models, where farmers must actively ask questions of a service, before the service is provided. Few models are exclusively one or the other, since education requires active participation of both the instructor and the student. Successful extension systems tend to start with a heavy push as a way to spread awareness and trust in a service, but gradually switch to pull models as farmers recognize the value of extension and increasingly see it out. Individual versus group: Extension can take place one individual at a time, or it can seek to educate farmers in groups. Individual extension can be highly customized to individual needs, but it can be costly. Classical T&V is typically conducted on an individual basis. Group extension allays cost issues, and it can additionally bring a number of social advantages to bear. Farmer field schools are an example of group extension. Trust issues can be mitigated when an entire community is involved; group peer pressure can accelerate technique adoption; and group members can support one another in learning. Groups, naturally, can be of varying size. Mediated versus direct: Instructional technology can be mediated during use, or directly manipulated by the learner. Mediation requires a capable person to be present during technology-based instructional sessions, and can therefore be costly (in terms of recruitment, training, and ongoing cost), but it often has much greater impact than direct usage of technology. Digital Green and Grameen Technology Center’s CKWs are both mediated models. One World South Asia’s LifeLines, where individuals make phone calls into a service, is an example of direct technology usage. Synchronous versus asynchronous: Synchronous systems provides interaction and responses in real time; asynchronous technology does not enforce real-time responses, and typically involves latency between each instance of communication. Telephone helplines in which live operators take calls are an example of a synchronous system, while aAQUA’s online bulletin board is an asynchronous system. Automated SMS-based Q&A systems have the flavor of a synchronous system, although long delays in SMS transmission would technically qualify as asynchronous communication.
Consumption versus production: Some systems of extension may seek active participation of farmers, through feedback or content production that ultimately impacts last-mile extension. Because many electronic devices can capture content, in addition to playing it back, this opens the door for devices to support content production activities, through recording of farmer questions, comments, or activities. LifeLines, for example, captures farmer questions through a helpline, and the questions (as well as the answers) are stored in a database for future use. Digital Green produces video content that involves farmers as actors. Consumption and production are not necessarily exclusive, in extension or with devices – both may occur in a complete extension system, though rarely at the same time.
Common Scenarios Below, the scenario categories are identified from the point of view of the content consumer, and not in terms of the technology or perception of content suppliers. Again, extension scenarios that are outside of the scope of this study – such as extension via broadcast radio or TV programs, or the training of agriculture extension officers through tertiary education – are not included. Scenarios are listed in approximate decreasing order of their frequency in worldwide extension activities, with the most common forms at the top. (1) One-to-One Interaction and Small-Group Presentation: Classical agriculture extension (also called “training and visit” or T&V extension) typically involves a single agriculture expert interacting on a one-to-one basis with farmers, and this remains the most common form of extension. This kind of interaction involves a farmer and an acknowledged agriculture expert or salesman. The intimacy of the relationship permits significant interaction with minimum social pressure. Typically, any devices in this scenario would be managed and operated by the agent. In some cases, a family, or a very small group of farmers (say, five or fewer) may be involved. More than the group size, the key characteristics of this category are that (1) the group members are sufficiently comfortable with each other that they feel little social pressure when asking questions, etc., and (2) the extension agent can pay attention individually to group members. Challenges in this extension scenario which could be mitigated by technology include the following: lack of trust; poor or incomplete knowledge on the part of the extension agent; difficulty of explaining agricultural techniques through spoken dialogue only; (2) Large-Group Presentation: More recent extension techniques, such as Farmer Field Schools, have begun to incorporate larger-group interactions, with one or more agriculture experts presenting to a larger audience of farmers (more than five, and typically more than ten). Typically, any devices in this scenario would be managed and operated by the presenters. The key traits of this kind of interaction are that (1) audience members are not necessarily familiar with each other, and social pressures may keep individuals from interacting; and (2) extension agents have less ability to interact with individuals on a personal basis. Group size, however, can be an advantage, if, for example, peer pressure or the prospect of community recognition can be applied in a positive direction.
Challenges which could be mitigated by technology include the following: reduced ability to project clearly to a large audience; poor or incomplete knowledge on the part of the extension agent; difficulty of explaining agricultural techniques through spoken dialogue only. (3) Personal Study Aid: Instructional materials may be consumed in private by individual farmers or their families. Though not uncommon for printed materials (e.g., brochures and pamphlets), this kind of device usage is rare, possibly because of the relatively high cost of electronic devices compared to paper, as well as the very real likelihood that materials end up unused and gathering dust. This kind of interaction allows private consumption of materials in the home, and intimate one-to-one usage of the device. It is further assumed that devices will primarily be owned (or rented), and operated by individuals and households, rather than by NGO staff or extension agents. Devices may be shared between households – a frequent phenomenon with electronic devices otherwise – but usage will tend to be by individuals or by very small groups. Challenges which could be mitigated by technology include the following: delivery of upto-date content; limited formal education of potential users; difficulty of explaining agricultural techniques through paper only. (4) Content Production by Farmers: There are cases when farmers are themselves the source of agriculture practices, and so incorporating their knowledge into the extension process can be valuable. With respect to this study, this scenario imagines farmers proactively recording practices they know of or which they have originated. This scenario is not expected to occur with great frequency.3 (5) Content Production by Professional Staff: Most agriculture extension in universities and formal institutions results in some kind of content being produced. The vast majority of this content is highly technical and not of immediate value to smallholder farmers. However, many institutions are aware of their target audience and increasingly produce print, audio, and video materials for frontline farmers. We do not spend a lot of time with this scenario, as these institutions typically have reasonable production capacity. Note that a particular system of agriculture extension may involve combinations of these scenarios.
Choice of Devices We chose 12 devices for the initial study. The criteria for choosing these devices were as follows: Potential for impact – Devices were considered only if they have some hope of being effective in agriculture extension. 3
There is a tendency to romanticize the notion of good agriculture practices originating from farmers themselves in the developing world, but this is no more common than it is for the average family physician to identify new cures for cancer. The poorest farmers are often the least educated and the least capable even among their peers (the more capable farmers often migrate to the cities for higher-income livelihoods). Farmers can introduce superstitious practices at least as often as they identify scientifically sound practices. And, at least within local communities, a considerable amount of information dissemination already occurs through word of mouth. For all of these reasons, we do not strongly advocate an approach that glibly promotes farmer-sourced knowledge dissemination. Nevertheless, we include this scenario here, for organizations that are careful in their screening of such knowledge prior to dissemination.
Portability – All of the devices are relatively small, portable devices that can be easily carried by a single person. They range in size from items that can be carried easily in a pants pocket, to items which are the size of a large book. Affordability – Cost is one of the key constraints, but since different implementation models can admit different costs for technology, we looked at a range of costs. Devices range from $10 to $500 retail cost (corresponding to bill-of-material costs from $0.50 to $200), with most devices in the $20-80 range. Multimedia functionality – All devices will permit some kind of audio and/or visual display that goes beyond text, in anticipation of users with low literacy. Functional diversity – As a set, the 12 devices were selected to span a wide variety of costs, form factors, and functionality. For similar categories of devices, we chose a single exemplar within that category. Recency: Devices are relatively recent either in their invention or in their commoditification. Notably, radios and television sets were not included, because their usage and impact in extension is well understood.
Conclusions We summarize the conclusions from this study before providing details of each device. Scenario-Device Match Different scenarios require different feature sets. Small-group presentations assume that an instructor figure is present. Thus, feature requirements are low, as the instructor can be relied upon to fill gaps in knowledge of usability. Video/imagery alone or audio alone are often useful as instructional aids, and video (with audio) is a useful point of departure for discussion. Small-group scenarios also often assume that the instructor will carry the device with them from location to location, so the device must be portable, rugged, and battery powered for best effect. Among the devices surveyed, those best suited for small-group presentation include…
Custom voice recorder Voice recorder with hand crank Portable A/V player Mobile smartphone Netbook
Large-group presentations have similar requirements as for small-group presentations, but the dimensions of the audio and video must be greater. Whereas a small screen, such as on a mobile phone might suffice for small-group presentations, large groups require a full display, at least as large as a traditional television set (e.g., 17” diagonal), and audio volume and quality are critical needs. Power and portability may be less of a problem in some circumstances, depending on whether the device is stored in a secure place. None of the devices surveyed is particularly good for large-group presentations, due to issues of display size or audio volume. The Saber device is perhaps the one exception, as it has considerable audio volume.
Self-study scenarios demand additional features beyond small-group presentations: they need to be very low cost, as replacement costs for breakage can add up. In addition, the user interface for the device must be exceedingly simple, so that people can manipulate the device on their own. In this regard, feature richness may be a disadvantage as rich features often mean more challenging UIs. Among the devices surveyed, self-study scenarios are best served by…
Audio greeting card Audio-enhanced book Educational toy (Power Touch and Whiz Kid – however both would require signficant cooperation with the manufacturers to create relevant content)
Feature-Cost Tradeoffs Plotting functionality versus cost on a graph, we find, not surprisingly, that there is a clear feature-cost tradeoff, and that greater functionality comes at greater cost. (Note that the cost axis is on a log scale, and that the feature axis is cardinal, not ordinal.) Less immediately obvious is that there are a few devices that pop out as delivering more functionality for cost. These include: the netbook, the mobile smartphone, the compact video camcorder, the portable A/V player, and the custom voice recorder. For the three devices used in the follow-up study, we will choose as follows: Between the netbook and the mobile smartphone, which comparable in their feature-cost tradeoff, we select the mobile smartphone, due to currently high interest in mobile-phone applications. Both the portable A/V player and the custom voice recorder appear interesting in that they are the two lowest-cost alternatives.
Figure 1. Functionality versus cost of the 12 devices surveyed, plotted on a log scale with respect to cost. Two points to note: (1) There is a clear feature-cost tradeoff, as greater functionality comes at a greater cost. (2) Those devices that tend towards the top-left are those that have relatively high value for cost.
Device Survey [1] Device: Digital Voice Recorder (Dictaphone) Model: Olympus Digital Voice Recorder VN-5200PC Overview:
“Dictaphone” with audio recording and playback Internal microphone and speakers, with option for external microphone and speakers Upload/download to PC via USB port Low-to-moderate UI with LCD display (no backlighting) 35.5 hours of audio 30+ hours playback on one pair of AAA batteries $60 retail; ~$30 bill of materials Optimal for one-to-one interactions and small-group presentations
Description: The Olympus Digital Voice Recorder’s (DVR) primary function is to permit casual, monoaural recording and playback of audio, and for these tasks, it is exceptionally well-designed. Recordings are best made with the device held next to the sound source, as the in-built microphone has no special accommodations for directionality, noise cancelation, or high fidelity. Speech at a distance of ten feet can be captured under conditions of low noise (e.g., presentations indoors, or loud speech outdoors). Playback volume is moderate without external speakers and best suited for intimate playback for one person or for a small group. Jacks for an external microphone and speakers or earphones, however, significantly extend the potential of the Olympus DVR as both higher-fidelity recording device as well as a large-group playback device. Of course, these capabilities would require additional hardware microphones or speakers, but they make the device more versatile. The UI is extremely simple, with one-touch recording and playback. Recordings can be organized into a two-tier hierarchy of folders and audio files which are also easy to navigate. A small, but effective LCD display allows users to navigate folders and audio files by number, as well as to monitor parameters such as battery life, length of audio file, microphone quality, etc. The display is easy to read in lit conditions, even with bright sunlight, but it lacks backlighting and will be difficult to read in the dark. The build is moderately rugged, with hard, light, plastic construction. The electronics are all solid state, suggesting that it will endure any movement apart from a hard drop. Drop tests from up to 60 inches onto carpeted floor did not harm the device. The device is not designed to be dustproof or waterproof, and dust or water in the audio jacks, microphone, or speakers could pose problems. The device uses two AAA batteries for power, permitting over 35 hours of playback time, which is excellent. The use of rechargeable batteries will reduce long-term costs for batteries and minimize waste.
The Olympus DVR is typical of off-the-shelf “Dictaphone” devices in terms of its cost, functionality, portability, and performance. Functionally, it is similar to Literacy Bridge’s “Talking Book” and Global Recording’s “Saber” devices, both audio-recording devices that are a part of this study. It differs from the latter devices, however, in a number of ways. The Olympus DVR is extremely portable, fitting easily in an adult hand or in a pocket, whereas both the Talking Book and Saber devices are quite a bit bulkier. Conversely, the Olympus DVR may be less rugged (both of the other devices are designed for ruggedness) and because of its smaller buttons, slightly more difficult to operate for users not comfortable with electronic devices (the buttons are comparable to those for a low-end mobile phone). Suitability for extension scenarios: The Olympus DVR seems best suited for one-to-one or small-group interaction as a playback device. Trainers could use it to playback audio, possibly in conjunction with an illustrated brochure, to present content to individuals and small groups. Its extreme portability makes it easy for trainers to carry it with them, and long battery life permits at least several days of usage without new batteries or recharging. Content can be easily generated and swapped by USB interface to a PC. With external speakers the device could also serve well in playing back audio to larger groups. It seems less suited as a personal study aid, however, due to issues of cost and potential to be used as a personal entertainment device. The user interface is simple enough to be learned by children and adults with minimal formal education (e.g., 8 or more years; ability to read numbers), but it may present challenges for extremely poorly educated adults. (Note: Studies undertaken by Literacy Bridge on their Talking Book devices suggest that this kind of usage may be feasible under good supervision, however.) The device also seems suitable for audio content production, particularly, if the device can be held right up to the speaker or sound source. If not, an external microphone will be needed for good audio quality. Future versions: We expect that future Dictaphone products will not change dramatically in terms of their functionality or form factor, because the market for digital Dictaphones is mature, having tracked the analog-digital revolution from miniature cassette-based devices to current digital products. In the absence of a large new market, they will likely hold steady at similar retail prices, with minor quantitative improvements (e.g., greater capacity or better audio fidelity), and sideways transitions to new interfaces and formats. Interfaces may change as PC and mobile-phone interfaces evolve. The core bill of materials also seems difficult to reduce: The greatest possibility would seem to be in reducing the cost per byte of memory, but mass-market trends in memory tend to maintain a lower bound on cost while increasing capacity as a way to handle manufacturing overhead and to maintain profit margins. Retail prices, however, are currently double or triple the bill of materials, and there is room for low-cost manufacturers to enter with enough new demand. If this occurs, costs may sink to very close to the BOM price of ~$30. Product website: http://www.olympusamerica.com/cpg_section/product.asp?product=1389
[2] Device: Custom Voice Recorder Model: Literacy Bridge “Talking Book” Overview:
Ruggedized “Dictaphone” with audio recording and playback Internal microphone and speakers, with option for external microphone and speakers Upload/download to PC via USB ports No display 35.5 hours of audio 10-15 hours playback on one pair of D batteries $35 bill of materials (projected
