Approach on Mobile Learning Design for the Underserved

This paper discusses an action research study focused on developing a mobile learning model of literacy development for underserved migrant indigenous children in Latin America. The research study incorporated a cyclical action model with four distinctive stages (Strategize, Apply, Evaluate, & Reflect) designed to guide constituencies involved in the study to design, test, and enhance a mobile learning model. The findings, to date, reveal some of the contextual phenomena that create both challenges and opportunities for a mobile learning model. From this, design strategies are evolving focused on sustained literacy exposure for extremely marginalized (economically, educationally, geographically, and technologically) migrant indigenous children who have no consistent access to a formal education system. Introduction This study shares lessons learned from ongoing action research conducted to design and test a mobile learning model that may help migrant indigenous children in Latin America. The study is part of a larger project established to develop entrepreneurial strategies that help the underserved with educational capacity building (i.e., initially through literacy development), ultimately leading to increased economic and social mobility. At the onset of this study, a prototype of a handheld mobile learning device was adopted as an integral part of the learning model. An action research approach was employed to investigate the possible effects of various affordances for the handheld mobile learning model. Action research has been reported to be an effective research method for technology implication studies involving economically and digitally marginalized populations (Chetty et al. 2003, Hartviksen et al, Mobile Learning Design 3 2002, Lennie et al, 2005). One of the fundamental beliefs of action research is that complex social systems cannot be easily simplified for meaningful study. In reality, factoring of an intricate real-life setting into a number of research variables often does not generate valuable knowledge about the whole system (Baskerville, 1999). In this study, the learning model investigated was applied in a relatively isolated social setting drastically different from contexts for which conventional technology solutions are usually developed. Therefore, this study chiefly concerns the identification and specification of an idiosyncratic setting where problems exist and furthermore the investigation of future technology-design possibilities through the development of research alliances with those who drive technological change. Henfridsson & Lindgren (2007) well noted these research features in their paper focused on action research for developing and testing a new technology product. The research model in this study is based on four stages of action (Strategize, Apply, Evaluate, and Reflect) and six guiding investigative criteria (Situation Specificity, Cultural Sensitivity, Practical Usability, Theoretical Applicability, Economical Scalability, and Viable Sustainability). This framework was adopted from a literacy development mobile learning concept named “PocketSchool” ( Kim, Miranda, & Olaciregui, 2008). In the early stages of conceptualizing the study, the social setting introduced a mixture of unusual variables and conditions (e.g., the migrating nature of the population, extreme poverty, indigenouslanguage-speaking parents, the absence of structured formal education systems, etc.). Given these significant challenges, this study focuses particularly on increased understanding of the multivariate context and its implications for the mobile learning model’s design. Through iterative actions over a 15 month-period, the investigator documented discussions with constituencies at multiple levels and domains, observed migrant indigenous children at 5 different villages, and pilot-tested various mobile devices. Mobile Learning Design 4 Critical reflections from these early pilots have refined the overall understanding of the context, the learning model and its constraints, and future opportunities. Background Literacy as Freedom Literacy is a foundation to capacity building. Key international conventions and declarations recognize literacy as a human right (UNESCO, 2006). Fundamental to individual development, literacy is one of the most essential learning tools and a basic learning need of every person (UNESCO, 1994). Literacy has numerous benefits for individuals, families, and communities and nations, through personally empowering individuals (Easton, 2005) and enabling them to participate in the local and global social community (Stromquist, 2005). Literacy efforts also contribute to increased life expectancy, reduced child mortality, and improved children’s health and education, among other cultural and economic benefits that promote diversity, equality, and development (UNESCO, 2006). The Marginalized Populations in Latin America. In the highly unequal society of Latin America, children of different social backgrounds do not have equal opportunities to learn and reap its benefits (Reimers, 2000). Many are still denied their right to an education and find themselves unable to break from the cycle of poverty. Inequality is particularly acute for indigenous populations; wherever they live, many indigenous people are among the poorest of the poor in that country (Hall & Patrinos, 2006; Psacharopoulos & Patrinos 1994; Tomei 2005). In Latin America, approximately 50 to 60 million underserved indigenous people reside, mostly in Bolivia, Ecuador, Guatemala, Mexico, and Peru (UNDP, 2004). Major differences exist in literacy rates and access to formal education for non-indigenous and indigenous populations (UNESCO/OREALC, 2004). For instance, in Ecuador 18% of Mobile Learning Design 5 non-indigenous people benefit from a full-time education, while the figure for indigenous people is a mere 1% (Gradstein & Schiff, 2006). In Mexico, many indigenous children live in communities so small that no school is provided; and another group (consisting of about 400,000 to 700,000 school-age children) travel with their parents every year for the harvest, never staying in one place long enough to be enrolled (Schmelkes, 2000). Without an innovative intervention to counter the adverse effects of globalization and technological advancements, inequality will only intensify, further excluding the uneducated from society and leaving the extremely poor without the necessary skills to secure their well-being. Mobile Learning and Literacy Development The development of mobile technologies has opened up a huge array of possibilities for the domain of literacy development and language learning (Joseph & Uther, 2006). In recent years, there have been numerous studies and projects using mobile technologies for language learning and both formal and informal literacy development (Brown, 2001; Cabrere, 2002; Chinnery, 2006; Joseph, Brinsted, & Suthers, 2005; Kadyte, 2003; Kiernan & Aizawa, 2004; Levy & Kennedy, 2005; Norbrook & Scott, 2003; Paredes et al., 2005; Thornton & House, 2005; Ogata & Yano, 2004). To date, there is no formal theory or model for mobile language learning (Joseph & Uther, 2006). Consequently, the vast majority of mobile literacy projects are experimental. Current attempts to use mobile devices for language learning range from vocabulary or grammar learning to story reading and pronunciation practices. Some mobile programs rely on prepackaged content loaded in a handheld machine, while others employ weekly podcasted shows which are downloadable from the Internet. Given that most underserved indigenous people lack adequate Internet access, prepackaged content on a light and inexpensive mobile device is the better option. Emerging mobile technologies increasingly Mobile Learning Design 6 empower other potential learning solutions via conceptual models that are interactive, ubiquitous, and convenient. Mobile Learning for the Disadvantaged Because mobile learning devices now have the potential to achieve a large-scale impact due to their portability, low cost, and versatile features, recent innovations in mobile technology offer promising opportunities to combat the deep-seated chasm of inequality entrenched in Latin America (Roschelle, 2003) and many places on earth. Also, a convergence of rapid advancements in information and communication technology (ICT)— in particular, increases in processing power, memory, and connectivity for handheld devices—have resulted in an explosive growth in media richness, ubiquitous access, and highly personalized learning solutions (Pea & Maldonado, 2006). Today’s inexpensive mobile device can store and deliver a vast amount of information, including substantial amounts of K-12 curriculum. Equally important, such devices are capable of reaching even the most isolated audiences (Attewell, 2004). For these reasons, informal learning through mobile technology seems a highly viable option worth exploring for migrant indigenous children in extremely underserved conditions. Method Participants Over 250 migrant indigenous children, along with about 20 parents, 9 local and regional representatives, 4 government officials, 5 technology corporate leaders, and 6 support staff for the investigator, have contributed to the study. The participating children and parents were sampled from 5 different villages in a rural region where large-scale agricultural farms were being operated by corporations. The migrant children with their parents came from rural mountain communities located a few thousand miles away. Mobile Learning Design 7 Many of the indigenous children travel back and forth seasonally and some of them stay in the shelters provided by the farm owners. One-way trip between the hometown communities and the village shelters can take up to several weeks by bus. The people often move between village shelters as harvesting work becomes available at different locations. Setting The region where the 5 villages were located was at least 300 miles away from major cities. In fact, 3 of the villages were at least 15 miles from the center of provincial towns where an intercity bus was available. One of the villages was only a mile away from a small town where interactions between the migrant people and the local town people were possible and frequent. The migrant people could make trips to the town only by foot, and the migrant children had no stable access to a formal education (in part because the existing schools in the relevant areas were not adequately equipped or consistently run). In one of the villages the researcher visited, there was a one-room school-like facility, and class meetings were sporadically held. On these occasions, over 30 children of different ages (i.e., 3 year – 14 years old) were together in the room, creating extreme challenges for a teacher to manage them all at once. Procedure As shown in Table 1, the AR model employed in the present study relies on four distinct action stages (Strategize, Apply, Evaluate, and Reflect) to investigate various aspects based on 6 core investigative criteria (Situation Specificity, Cultural Sensitivity, Practical Usability, Theoretical Applicability, Economic Scalability, and Viable Sustainability). Between December 2006 and March 2008, all of the 5 villages were repeatedly visited 5 times at an approximately 4 month-interval. At each visit, a series of iterative formal and informal meetings, conferences, direct and indirect observations, and interviews took place; Mobile Learning Design 8 interactions included the introduction of the mobile learning model, technical training, and on-site observations (direct and shadow). The observations and interviews occurred mostly outdoors around farm worker shelters, inside the shelters, or in farms. The children were interviewed individually or as a group (i.e., 3-8 children at a time). In parallel, numerous meetings were held with corporate representatives, government agency officers, local NGO leaders, university partners and other constituencies relevant to the project. The multiple constituencies provided multiple viewpoints on the research findings. Especially, the local NGO leaders and university partners confirmed the findings based on their own visits and observations. Various types of data—including diaries from observations, documents, audio recordings, photos, and video recordings—were collected through a series of interactions with all constituencies involved in this study. All field notes were typed up as possible and organized into temporary categories with labels. Through iterative analyses using data mapping software, the categories and labels evolved to better organize the overall data. Mobile Learning Design 9 Table 1. Action research model Stages Strategize Apply Evaluate Reflect Activities Generate (new) questions Involve (new) constituencies & supporting resources Strategize (new) actions and enhance system designs Apply (new) systems changes or (re) implement (Re) provide system tutorial Add/remove peripheral stimuli Gather & analyze (new) qualitative and quantitative data Interviews, observations, diaries, or video recordings Document and record (new) phenomena, patterns, or differences Compare with early assumptions, hypotheses, or findings Identify (new) problems & opportunities Share findings Underlying Consideration Criteria Situation Specificity, Culture Sensitivity, Practical Usability, Theoretical Applicability, Economic Scalability, & Viable Sustainability Repeat Overarching Assumption The overarching assumption of this study builds on Goldenberg (2005): “Influences on children’s development (interactions with others, playing, watching television, reading, counting, etc.) are embedded in the routines of family life that themselves are embedded in a larger ecological and cultural niche” (p. 24-25) Unfortunately, possible “influences” are extremely limited for migrant indigenous children for many reasons; for example, most of the children do not have access to radio, television, Mobile Learning Design 10 or books. Unless there is an innovative and entrepreneurial counter intervention, these children will continue to grow without stable access to a formal education. Far too many children are in similar or worse situations in many parts of the global community. The Mobile Learning Device and Content Several mobile devices were employed in the study. All devices functioned in similar ways and were able to empower the learning content prepared for the study. Minimal differences among these devices were in their color, casing material, speaker volume, battery life, and screen interface. The device shown in Figure 1 is one of early prototypes (not presently available on the consumer market) used in the study. It has a USB port, power on/off switch, 6 buttons, built-in speaker, microphone, 1.5 X 1.8 inch color screen, from one gigabyte to multiple gigabytes of memory, and a camera. With a charge, the battery continuously runs for about 12 hours. This device can be charged by a computer with a USB port, or with AC charger, or with a solar cell battery charger with a USB cable, or with a hand-crank charger specifically devised for the research. The device can store camera images or recorded sounds. In addition, it can play other multimedia contents. One multimedia Alfabeto lesson, several dozen short stories in different reading levels, and multiple sing-along animations were created to fit in the mobile device. The Alfabeto lesson was about 2.4 MB in size, and each of the stories and animations was about 1.2 MB. With even as little as a half gigabytes of memory space, one device can hold about 400 short stories (e.g., narrated for approximately 2 to 3 minutes) or 200 longer stories designed for various reading levels. The Alfabeto lesson presented each letter of the alphabet and also a sample word starting with that letter. Each letter and sample word was read once, and then the device Mobile Learning Design 11 displayed the next letter. After going through A to Z, the play button can be pressed again to replay this sequence, or the arrows keys can be pressed to play other learning contents. In later stages of the study, some devices were able to rewind, fast forward, pause, or repeat the content at any segment of the story. Figure 1. Prototype mobile learning device front and back The alphabet and corresponding words were read loudly in a female voice. Regarding the short story content, each story had sequenced animations and corresponding text. As shown in Figure 2, the narrator read each page in sequence. Figure 2. Mobile Learning Device with Multimedia Dictionary and Short stories Mobile Learning Design 12 Findings and Discussions Reporting findings from an on-going action research study is often challenging because it involves summarizing massive amounts of data in a compact form. Nonetheless, the findings most relevant to the designing of a mobile learning model are presented below in a summary format based on the six core investigation criteria set forth earlier. Situation Specificity Living Condition. With the exception of one village located closest to a provincial town, the migrant indigenous children in four other villages were highly isolated, making interactions with the townsfolk physically impossible and limiting possible literacy exposures (e.g., posters, signs, newspapers, telephones, and television). In almost all cases, the families of 4 or 6 members lived together in a 100-150 square foot shelter with an indoor woodburning oven. Generally, during the day children under 14 spent all of their time around these shelters, unless they joined their parents in harvesting during busy times (although this is illegal in most cases). Considering the confined living spaces and absence of television or radio in most of their homes, there are many opportunities with the mobile device that extended beyond literacy content. For instance, such mobile devices could educate all members of the family with various learning content, such as health, environment, numeracy, and micro-financed businesses. Hygiene and Health. The general hygiene and health conditions observed in these villages were substandard. In all five of them, the children drank unfiltered underground well water where there was no sewage system, and many of them had ear infections. In fact, there was no sewage system in the entire region nor a hospital within a 200-mile range. Many mothers as young as 13 were found in the villages. The mothers had very limited knowledge about their own or their children’s health, safety, or environmental conditions. Mobile Learning Design 13 Vital health and safety literacy content could be easily introduced to the whole family using mobile devices. Working Parents. Those who worked were able to bring home up to $6 $8 a day on a good harvesting day during that season. Those who worked (usually age 14 and up) woke up around 4 AM to go to work and went to sleep around 8 PM; in some cases, they worked at night as well. Those who worked were provided transportation by bus to move between farms and shelters. Mobile devices seemed potential learning solutions for the parents while they are in transition (e.g., either on bus or walking). Technology Infrastructure. Except for one village located close to a small town, televisions, radios, or any other electric appliances were not present in the villages. A key factor, of course, was that there was limited or no electricity. The village close to a small town had a few owners of barely working radios or televisions. Except for just a few cases, most of the children did not own a single book. However, like many other places in developing countries, technology was becoming more available even in the provincial towns. Cell phone towers were being built and a few town residents in recent months had even acquired ADSL Internet access. The adoption rate of mobile phones was phenomenal even among the poor. Mostly, they carried a second-hand phone as a clock, calculator, calendar, camera or photo album more than a phone. Education. The parents indicated that education is important for the future of the children. A classroom facility situated at one of the villages had highly sporadic class meetings. In the town overall, there were no signs or advertisements that could trigger any reading. No designated transportation was available for the village children to attend schools located in towns where more stable classes were held. Even if the children had stable access to the town schools, the highest education they could possibly obtain in such highly Mobile Learning Design 14 heterogeneous classrooms (i.e., one teacher in one class for all age children) would be about 5 grade level. Nevertheless, in all cases, the children wanted to learn if the content was engaging (e.g., interesting stories). The children highly enjoyed activities such as singingalong, short-story telling, and playing with outsiders. Among the five villages visited, most of the more literate children were found in the village closest to the small town. Although there was no formal literacy assessment for the parents or children to date, it was obvious that the parents had minimal educational experiences or literacy exposure. More details about literacy opportunities and experiences are discussed below. Cultural Sensitivity Although the investigation in this category was limited, some important phenomena were found relevant to future mobile learning design. The indigenous immigrants seem to have preserved their cultural heritages along with historic, political, religious and linguistic backgrounds. Some of the parents shared their indigenous cultural artifacts, costumes, folktales, arts, crafts and dances. Some people from multiple tribal origins were present in the villages; and two distinctive indigenous tribes of Zapateco and Mixteco were dominant in the 5 villages. In most cases, the parents could not read Spanish. Those who could speak Spanish had a wide range of proficiencies. If these could not speak Spanish, they relied on their indigenous languages. Newer immigrants typically spoke less Spanish. In some cases, the new immigrants required translation aids by more experienced immigrants during the interviews. Some of the immigrants spoke multiple indigenous languages as well as Spanish. Since there were no books and the most parents were illiterate, the children were not expected to have much exposure to literacy (e.g., bedtime reading opportunities). Also, vocabulary sets linked Mobile Learning Design 15 to modern societies such as bathtubs, fast-foods, or electronic appliances do not fit in their cultural profiles. Practical Usability Initial adoption stage. The children who participated in the study varied in ages (3 to 13 years old). Except children with extreme learning limitations due to psychological disability symptoms (such as serious attention deficit disorder), in most cases the children of all ages learned to operate the mobile device within minutes The younger children responded to the aural stimuli more than the visual animations. Children at this age group listened attentively and learned to push the Play button often to repeat the same story. The 6-8 year old children responded to both the aural and visual animations and learned to browse stories by operating Up and Down arrow buttons, leading to different stories and animations. The children 9-13 years old not only responded to the stories, but also explored the mobile interface more thoroughly, leading to mistakes and dead ends in the menu choices. Overall, the children asked for minimal help in the first introductory stage, and in most cases children over 6 in most cases kept on playing different stories alone and did not need further help. This led the investigator to believe that a mobile learning device with a manageable number of buttons for operation would not need substantial training or ongoing help (i.e., leading to rapid adoption). Interaction stage. At various stages of the research, the device was given to either an individual child or a group of children for observations. In the individual-use case, each child had one device and spread out to play with the device alone. The children had complete autonomy to play with the device in any way he or she liked. As a result, the children often looked for comfortable places to spend time with the device. Since they spent much of their time outside of their shelters, the children needed a built-in loudspeaker. The Mobile Learning Design 16 loudspeaker also induced them to spread out more because they did not want to hear noise from peers’ devices. In most cases, the older children group wanted private space or comfort zones and therefore spread out to convenient and more secluded areas as shown in Figure 3. Figure 3. Children listening, watching, or reading short stories in private spaces. In the group-use case, 3-5 children at a time shared one device. Since the device had a loudspeaker, the children were able to sit around and listened to the stories together as one child used the device to shift among stories. In other cases, siblings in one family shared a device. When this occurred, the older child typically controlled the device and helped younger siblings to either listen or watch the short stories. When the devices were played indoors, bystanders (e.g., peers, siblings, parents) were able to easily hear the content as well. The only case in which a headphone for each individual may make more sense would be when the children in close proximity trying to independently use devices. For example, if the devices were used in classroom settings, an individual headphone would be a better choice. The screen size of the prototype mobile learning device was 1.8 inch long diagonally. The limited screen size meant that animations and corresponding texts needed to share a confined screen. The hardest part in creating experimental content was to fit both graphic Mobile Learning Design 17 animations and readable text in one small screen, yet maintain the visual appeal. The font size of the text was 11 (i.e., Arial font type). When necessary, the text had the first priority over any animations to ensure readability of the stories. In order to test if text over the animation was as readable as text on printed story sheets (with font size 11 in both cases, as shown in Figure 4), a few children with moderate reading ability were asked to read the text on the device and the sheet alternatively. These children read both texts at the same speed. How much distraction this busy and small screen would introduce to the children with low reading ability requires further investigation. The initial prototype had push-and-hold buttons (i.e., one button used to do two distinctive functions based on one push versus push-and-hold), as well as one push button. The concept of push-and-hold buttons was introduced to minimize the number of buttons needed for device operation. However, the push-and-hold button was not an easy concept to teach to the children, especially the younger ones, and thus the benefits of minimizing number of buttons by using the same button for dual functions were not realized. Maintenance stage. The children had minimal knowledge about the sensitivity and fragile nature of the digital devices. The devices were easily used as other tools (e.g., hammer, simple toy object) for younger children (i.e., under 5 years old). The screens of the devices were easily dirtied and scratched. In this regard, the screens would have had a longer life if they were flushed in to minimize abrasive contacts when dropped or when rubbed against the ground or walls. Overall, the devices had to be shock, water, scratch, heat, and dust resistant. A touch screen would not have survived even a day. In the later stage of the observations, multiple mobile devices were introduced to the children to investigate their preferences on shapes, colors, and casing materials. Majority of the younger children (3-5 years old) mostly favored a white plastic casing more than metal, Mobile Learning Design 18 black, or other colored casings and shapes, whereas devices cased in a thin metal casing were more preferable among older children (10 or older). Although a formal analysis was completely absent in the casing design issues, it appeared that younger children preferred a white plastic toy-like casing that provided a good grip; this might have provided a sense of safety, confidence, or feeling of better tangibility. For older children, the mobile learning device in a thin metal design may have been also a “fashion statement,” a phenomenon found by Katz and Sukiyama (2006) in developed countries. Theoretical Applicability Pedagogical dilemma. In earlier relevant studies, Revelle et al. (2007) loaded Sesame Street videos on mobile phones and engaged 3 to 4 year old children at home, mostly in lowincome communities in California, in learning alphabet 3-4 times per week for 8 weeks. In her study, she reported that mobile-phone delivery made it extremely easy to incorporate literacy activities into their daily routines and parents attested that their children found the mobile phone controls easy to use and could play or replay videos on their own without parental assistance. Also, preand post-interviews indicated a significant increase in the frequency with which parents reported engaging their children in literacy activities after participating in this study. The pedagogy, probably oversimplified, considered in the present mobile learning project was to devise a learning solution to provide the disadvantaged children with as much as possible enjoyable literacy exposure. The conditions of use (highly heterogeneous age groups, illiterate parents, lack of exposure to literacy, mobile nature of the population, extreme poverty, and isolation from civilized communities) certainly create difficult challenges for any capacity development projects. Acknowledging the drastic differences between California low-income communities and the setting here, some of the findings and Mobile Learning Design 19 critical assessments made in this study, especially those linked to pedagogical issues, are discussed below. Attention span and independent learner. Most of the short stories loaded in the mobile device were less than 2-minutes in length. The length of the short stories presented by the device seemed to fit nicely with the short attention span of the young children, given the many challenges to their focusing on any particular stimulus. Even in the context of developed countries, mobile learning options that accommodate learning in snippet sizes of time are valuable for people frequently on the move. Thus, considering the general conditions of mobile learning such as the numerous extreme circumstances of these disadvantaged children, abundant distractions in the environment at any given time, just-intime or just-as-needed informal learning chances, the limited battery-life, the pedagogies and any applicable theories selected for this context must be much more innovative than the those for conventional settings in developed countries. This concept of innovative pedagogy or implementation strategy is almost imperative because what is simply supplementary material or an optional learning resource for children in developed communities may be the primary or “only option” for extremely underserved children. Moreover, due to the fact that there is limited or no access to structured and stable school with teachers, these disadvantaged children are forced to be completely independent learners. Age-dependent reactions to system feedback. The younger children paid more attention to the acknowledgement-sound emitted from the device when a choice was made, while the older children did not care about the feedback sound. For example, when a story is selected, the device made a series of bell sounds. The task of selecting a story and the feedback sound was a rather simple one, but nonetheless the younger children seemed to enjoy the process of activating the device as almost a novel task. They often made “a-ha” facial expressions at Mobile Learning Design 20 such moments when they received the sound feedback. Older children did not find the story activation task much of an undertaking. As a result, they did not pay much attention to the feedback sound. Skill-dependent reactions to content. Some of the older children (i.e., 11 year or older) with extremely limited literacy skill preferred to spend more time with basic alphabet animations than with short stories. They also liked to be alone in isolated areas while reviewing the basic alphabet materials. While trying the materials, they often moved away from the peers who were more literate. In contrast, children with more developed literacy skills wanted to try many stories and wanted to read for others as well, as if they wanted to show off their reading ability. The children with developed literacy skills welcomed any questions about the stories and competed against peers in answering questions raised by interviewers. Learning content. Since the device articulated the stories (slow narration by native speaker), children with minimal spoken language skill in Spanish were able to understand the content. The children who were more literate had both text on the screen and the narrated voice to understand the stories. However, the short stories had no distinctive level differentiations (i.e., due to limited content development resources), and presenting the content based on their performance and literacy level was not implemented. In addition, there was no automatic assessment function to determine learners’ progress. Typically, 5 year old or younger children had no problems when the same set of stories was repeated. In contrast, children 6 years and older wanted to access many more new stories and actively explore content navigation and control options. Overall, numerous other approaches could enhance this mobile learning model and students’ literacy experiences. First, the children with early literacy development probably would have benefited greatly from more age-dependent content. Also, it would have been Mobile Learning Design 21 more beneficial if a simple level of artificial intelligence was employed to automatically and dynamically respond (e.g., provide adaptive learning content based on user response) to the learner at different stages of literacy development. Second, if the device kept detailed usage or interaction logs for later data analysis (e.g., time used, content completed, questions answered, and learning objects shared), this would have been a valuable assessment tool and resource to examine student learning in many different dimensions. Of course, records of the learning objects shared assume an ad-hoc peer-to-peer network component integrated in the model. Third, student-centered content organization and presentation along with a personalized interface (e.g., addressing the names of the children with sound and text) would have made the learning experience more personally engaging. Fourth, it would have been more meaningful for the children if the cultural aspects were taken into account for content preparation. By observing their religious artifacts in their shelters and their religious practices, it seemed logical to cater the content to include common religious stories with which they may be already familiar. Perhaps words from lyrics for folk music or ancient hero stories would have been much better than some of the terms presented to the children. Words such as bathtub, skiing, or microwave oven are not common artifacts the children have experienced. Economic Scalability Mobile learning ecology. From the repeated observations and interactions with constituencies involved, a mobile learning ecology, if there could be one, for the extremely underserved community would be certainly different from the one possible in developed countries. There are just too many missing pieces in the underserved region; for example, little or no access is possible to electricity, telephone, Internet, or technical support. However, considering the rapidly increasing number of wireless phone subscribers in Mobile Learning Design 22 developing countries and underserved areas, integrating a mobile learning model with existing wireless communication ecology seems an eventual, plausible option. Interviews with wireless communication corporations (from chipset manufacturers to handset manufacturers and from application developers to wireless service providers) suggest abundant opportunities at the theoretical level, but also overwhelming challenges at the technical level. Because the technical challenges would mushroom if the learning contents involve high-bandwidth consuming data (e.g., multimedia animations, videos, etc.), instructional materials must be completely preloaded on such a device before distribution or must be scheduled to be downloaded at off-peak times. Also, because there are substantial limitations in what radio technologies can provide today, for mobile learning to be integrated with mobile phone services, wireless communication standards and network protocols such as GSM(Global System for Mobile communications), GPRS(General Packet Radio Service), EDGE (Enhanced Data GSM Environment), HSDPA (High Speed Downlink Packet Access) and even newer concepts such as LTE( Long Term Evolution) or WiMax (Worldwide Interoperability for Microwave Access) should be examined in depth (e.g., spectrum regulations, region-specific standards, interoperability issues). Some of these newer standards and devices are highly attractive options for discussions on a future mobile learning ecology. In particular, during the project period, a small scale WiMax infrastructure has reached feasibility as an infrastructure. WiMax seemed to be a cost-effective, easy-to-implement wireless communications model as a mobile convergence option for underdeveloped regions, allowing voice communications (e.g., through Voice Over Internet Protocol), Internet browsing, digital broadcasting, or even distance education. For a remote small village with electricity, a wireless outdoor broadband solution such as WiMax or its variant coupled with highly affordable mobile devices may Mobile Learning Design 23 provide “last-mile” Internet coverage while accommodating various communication needs. However, without a global company’s significant investment supported by a business model addressing the needs of people with extremely low income, the design and implementation of a mobile learning solution remains unlikely. A Stanford University research project Programmable Open Mobile Internet 2020 (POMI), suggests a future technology infrastructure with highly accessible, usable, and scalable open mobile computing solutions (Clean Slate, 2008). In POMI 2020, one of the objectives is to demonstrate how the new programmable open mobile Internet model can be the catalyst towards open mobile computing and education, breaking down structural barriers leading to inequalities for the underrepresented in underserved areas. However, at the present time the mobile learning ecology for the underserved children must start with a “highly self-sufficient model” not depending on existing infrastructure or local resources. Per-unit-cost. The cost of the mobile device prototypes used throughout the study period varied from $20 to $300, yet all devices were capable of presenting the short story animations through a color screen and external speakers. What differentiated the devices from acceptability versus unacceptability mostly had to do with a combination of elements such as speaker quality, durability, battery life, connectivity options, and price. Among all elements considered, storage capacity was the least significant issue because even the cheapest device ($20) had 2 gigabytes of storage, which allowed at least several hundred short stories and relevant literacy support content. The cheapest device would probably be the most economically favorable option, but the cheapest device was the poorest in terms of durability, battery life, and speaker quality. The more expensive devices had very good quality loud speakers and had a very long battery life (e.g., 12 hours) with plenty of storage space (e.g., 8 gigabytes of space which can hold Mobile Learning Design 24 over thousand multimedia stories in various lengths), but obviously would be less economically favorable to scale up. Through iterative interviews with various device manufacturers, one thing was clear: the larger the initial order is placed, the lower the manufacturing cost. Regardless of the range of elements contributing to the cost, sustained manufacturing volume and system features are the major price-determining elements. Advancement of technology. Since the mobile device market is one of the fastest moving markets in today’s economy, many of the technical recommendations in this study will require revisiting within 6 months or less. For example, some of the mid to high-end mobile devices had extra features such as satellite-based DMB (Digital Media Broadcasting) television reception feature, FM radio, flash games, a few language dictionaries and text translators, calendar, scheduler, built-in mini keyboard, and wireless networking capabilities. Some of the devices were in-between a mobile phone and an ultra mobile personal computer (UMPC) with a proprietary Internet browser. More advanced features can certainly make more exciting and innovative learning experiences possible. However, as more advanced technologies become available and affordable, whether more mobile projects and more educators devising learning options for the disadvantaged or unserved children is uncertain. Viable Sustainability Assuming an educationally sound mobile learning model is devised and implemented at a large scale; could such model attain sustainability? Many elements will contribute to the destiny of such model. In this section, some sustainability issues linked to observations made in this study are discussed . Life-sustainment necessities. The parents of the indigenous children had been asked about the value of education in their children’s present and future days. Parents indicated that education is very important for their children. At the same time, making money is Mobile Learning Design 25 extremely essential for their family. The parents are fully aware that high literacy or numeracy skills often enable them to earn more money. They also discussed among themselves about where and how to get better pay. Some people moved from farms to urban regions or well known tourist destinations to earn more money; for some, such cyclical migration was a routine. However, for almost all, what they earn for a day had to be spent during that day to sustain their life. Saving money is not impossible, but is extremely challenging. Due to the migratory nature of their life, storing resources is extremely shortterm. Surprisingly, some of the parents offered the researcher eggs or vegetables they kept in their shelter for thanking us for coming a long way and doing what we were doing. The findings above raise several questions related to sustainability: How many of these people will be interested and able to afford a mobile learning device, even if it were loaded with high quality learning contents they may perceive valuable and the cost was as low as $20? Will a low price in this case enable a rapid market adoption, or would devices priced at a higher level be more effective and have a longer lifetime? In this regard, Abed (2008) suggests that charging even small fees would be better than not charging because if people pay even a fraction of the value, they still expect certain level of quality and demand service or help. Overall, sustainability must be tightly linked to daily needs and activities, and price point certainly also influences sustainability. Further research is needed on the topic of pricing point to perfect a sustainable distribution model. Hybrid business model. The children found the mobile device a highly joyful reading companion, and the parents perceived the device as good for learning. However, what distribution and sustainment model (e.g., free-to-own, lease-to-own, purchase-to-own) would be most suitable for the children in the context needs further exploration. In addition, Mobile Learning Design 26 which acquisition model would lead them to best maintain the device and continue to use it is unknown. Throughout the research period, constituencies such as government agencies, non-profit foundations, private investors, and device manufacturers were interviewed. Overall, no single entity is interested in or capable of subsidizing the major part of the cost of the mobile learning model for its sustainment. What arose from the discussions as a more plausible solution was a hybrid model of business structure (e.g., non-profit philosophy combined with for-profit sustainment strategy). In such model, local entrepreneurs would be funded through social entrepreneur organizations with micro-financing options. The local entrepreneurs would distribute the mobile learning solutions and also provide support, since they would be the closest entities to the communities in the present context. In such a model, a non-profit organization would supply and promote a proven mobile learning device to these entrepreneurs. Academic research on social entrepreneurship is still in its infancy (Dorado, 2006), but increasingly more entrepreneurs and ventures are bridging profit and service goals in new and creative ways (Eakin, 2003). Therefore, an innovative hybrid model may soon be conceived by young social entrepreneurs currently in social innovation research centers in leading universities. A good model would not only provide a literacy development solution, but also a long-term capacity-development strategy leading to career development and placement (e.g., helping the people to secure a job such as globally outsourced customer support workforce helping Spanish speaking customers in U.S.). Such life-altering solution would help to counteract the vicious cycle of poverty and inequality deeply rooted in Latin American societies. Design innovations. During the research period, various technical concepts to increase feasibility and sustainability were tested. For areas in which electricity is a scarce resource, Mobile Learning Design 27 solar cell chargers and hand-crank charger were integrated. Although technology for such options has advanced drastically, the solar cell charger required many hours or even days to sufficiently charge the mobile device under a bright sunlight condition. A solar cell charger based on today’s technology is a questionable option considering that mobile learning is about moving around with the device and keeping the device mostly in pockets. In regards to the hand crank charger, on average it usually takes twice the time to charge the device as it does to operate it. For example, if a child wanted to play the device for 2 minutes, the handle must be cranked for 4 minutes. Worst of all, the battery can never be drained completely. Otherwise, it takes enormous amount of time and effort to bring the battery back to an operational state to power the digital device. An integrative approach, such as combining the mobile learning device with a flashlight was also explored, because a dual-purpose artifact may increase overall sustainability. However, in such case, the battery was often drained quickly, and the digital device had to be charged all over again. In terms of dual or multi-purpose design for increased usability and sustainability, it seems that combining the mobile learning device with a toy or playing activity appropriate for the target age group would be a better idea for effortless, joyful, unconscious battery charging. For example, playing activities such as jumping ropes and pulling cart boxes with wheels were often seen in the villages. As shown in Figure 4, jumping rope activity or cart wheels could be the source of motion energy to charge a mobile learning device. (i.e., use rotational motion to spin an AC electrical motor to charge the battery). Overall, innovative design strategies of integrating the mobile learning device with children’s daily activities seem to offer the potential to increase sustainability. Mobile Learning Design 28 Figure 4. Potential sources of motion energy in playing activities Edutaining content. It is quite common in developed countries to see children today with numerous mobile devices. The devices vary from electronic dictionaries to mobile phones and from game players to mp3 music or mp4 video players. For different devices, children seek and maintain a variety of entertaining contents. Children often share contents whether they have been downloaded or ripped or even created by themselves. For children in similar interest groups, highly entertaining content is always popular and heavily reused. However, for the children in the present context, short story content was highly entertaining since they had not watched television or accessed Internet content. As shown in Figure 5, the children often shared their screen with siblings and peers, as they found the content entertaining. In some cases, the children often deeply concentrated on various short stories for an incredibly extended time. Overall, if a home entertainment system came with a library of entertaining digital videos, perhaps several thousands, such an entertainment system would probably be used for a long time on an on-going basis. In the same vein, if there were no other means to add more content, sustainability of mobile learning solution in the present context would depend significantly on the amount and the presentation fashion of the edutaining content preloaded. Mobile Learning Design 29 Figure 5. Children sharing the screen with siblings (left), and a child concentrating on an activity (right) Conclusion This study employed an action research model with cyclical stages to examine issues around designing, testing, and enhancing a mobile learning model for extremely underserved migrant indigenous children in rural areas. The researcher investigated the social phenomena and possible effects of a mobile learning model introduced in the context. Although this research is continuously being refined, this study presents some key findings from iterative data analyses and critical reflections thus far. The migrant indigenous children are, in almost all cases, the most severely disadvantaged among the underserved populations in Latin America because they do not have equal access to public services, such as education or healthcare. As documented in this study, the indigenous children do not have access to reading materials. Since almost all indigenous parents interviewed in this study could not read, they would not be able to read stories to their children even if they had a book in their hands. Mobile Learning Design 30 The mobile learning device seems to promise to expose the children, and perhaps their parents, to an extended set of vocabulary that they usually would not encounter. The findings support that the simple mobile learning devices minimize the initial learning curve for immediate adoption. In fact, children can get started in minutes. Regarding readability, children with some literacy were able to read the stories presented on the mobile device display at the same speed as they would with papers, although the display size was probably not the most ideal. Thus, a mobile device loaded with easy short stories is shown to be a viable jump-starter for children who are at the early stage of reading proficiency. Also, as an interesting and joyful activity, listening, watching, and reading numerous short stories or animations with the mobile learning device is shown to provide substantial learning moments for the extremely marginalized migrant indigenous children. Although more interviews and observations are needed, among various prototype devices employed in the study the majority of the younger children preferred a toy-like design with plastic casing, whereas older children preferred a slim design with metal casing. Future Design and Research Considerations For ultimate capacity building of the children, a variety of learning content covering multiple subjects and topics including health, environment, and micro-business could be designed, tested, enhanced, and provided through the mobile learning model. For this, the content design requires repetitive testing to see what approach would maximize literacy and eventually readability while maintaining appealing aesthetics of the content. At the same time, to avoid overwhelming experiences, content must be easily retrievable in a granular fashion. Instead of using directories and subdirectories that may be complicated for children, the use of colored or icon-based categories to navigate through the lessons can be tested. As choices are made and buttons are pressed, voice guidance could Mobile Learning Design 31 possibly further assist users. Overall, the learning experience must remain fun and accessible for children; technology should not obtrude upon it (Sharples, Corlett, & Westmancott, 2002; Parsons & Ryu, 2006) For the hardware design, a careful consideration is required because of numerous constraints in the context. Given that technicians will not be available in remote rural areas to service the device should it become inoperable, it should be highly shock, water, dust, heat, and scratch resistant. For example, a design using fewer buttons sealed with rubber casing materials would be ideal, unlike a touch screen button that would be easily dirtied and scratched. The device should also render itself accessible to children by having easy to comprehend features and uncomplicated functioning (Papanikolaou & Mavromoustakos, 2006). The user-interface and buttons should provide the learning content as quickly and with as few operations as possible (Low & O’Connell, 2006) without unnecessary complexity (Parsons & Ryu, 2006). 1 In this regard, a keyboard-like input device may not be appropriate because it would limit the usability of the device to older children with basic literacy skills or at least third grade or higher grade education experience (See Bartholome, 1996; Fleming, 2002). In terms of affordability, mass production can lower the price of the digital devices dramatically. With mass production, large scale distribution becomes possible. Also with the help of social entrepreneurship foundations and non-profit organizations focused on content development and device design, local entrepreneurs could be developed to offer the learning devices and necessary support to the disadvantaged children at a low acquisition cost. In regards to content development, situation specificity or cultural sensitivity seems to be such an important issue for a global scale education project. Without cultural Mobile Learning Design 32 sensitization, learning contents many not make any sense if not offending or even disruptive. In this regard, Martin (2008) shares his anecdotes in social entrepreneurship projects in underserved communities where he observed his colleague organizations handing out vitamin supplements made of pork as gifts to Islamic populations as well as distributing a massive number of pairs of plastic sandals to regions where selling homemade sandals was lifeline small business among poor families. In addition, he saw them donating free bikini swimsuits to Afghanistan woman groups and handing out diet meal packages to starving Ethiopians. Some of Martin’s accounts on situation and cultural insensitivity may be extreme, but it is important for educators to keep in mind that certain words, images, concepts, artifacts or practices often may not be understandable if not offensive or disruptive. In terms of practical steps of a scaling up process, Abed (2008) suggests a model he considers highly effective. In his social entrepreneurship model, one must start a project in small scale, test it and try to make it effective, become efficient in the operation of the model, and scale up while routinizing only the essential elements and discarding the rest. Abed’s perspectives on a successful model of an entrepreneurship project seem to be also valuable considerations for a research design because a series of progressive attempts to perfect a scalable and sustainable learning model (i.e., leading to better understanding of the context and better impact as research outcome) is perhaps an ideal pattern in action research. In sum, future iterations of this study will certainly pay more attention to scalable and sustainable learning design issues, developing and involving local entrepreneurs, microfinancing and micro-business opportunities, and providing and maintaining proper support for mobile learning models. Acknowledgement Mobile Learning Design 33 Special thanks go to Claudia Olaciregui of Colombia, Natalia de la Chica of Spain,Grace Hyunkyung Kim of Korea, Pablo Ohm, Alberto Gárate, Laura Carrillo and TeresitaHigashi of Mexico, Talia Miranda, Alejandro Toledo and Ana Maria Romero-Lozada ofPeru, and Luz María Camey of Guatemala for making this study possible. I also want tothank university colleagues such as Martin Carnoy and Claude Goldenberg of StanfordUniversity, and Curtis Bonk of Indiana University as well as Chris Dede of HarvardUniversity for reviewing this paper and sharing invaluable insights on relevant topics to thisstudy. This Pocket School project has benefited tremendously from generous contributionsmade by organizations such as IFL (Innovations for Learning), Nokia, iRiver, KT, AT&T;,CETYS Universidad, and many individual entrepreneurs.ReferencesAbed, F.H. (2008). Thinking big and scaling up. Stanford University, Stanford, CA. RetrievedJune 21, 2008. from http://itunes.stanford.edu Attewell, J. (2004). Mobile technologies and learning: A technology update and m-learning projectsummary. Technology enhanced learning research centre. Learning and Skills DevelopmentAgency. London, UK. Retrieved March 20, 2007, fromhttp://www.lsda.org.uk/files/pdf/041923RS.pdfBartholome, L. (1996). Typewriting/keyboarding instruction in elementary schools. Retrieved March20, 2007, fromhttp://www. usoe.k12.ut.us/ate/keyboarding/Articles/Bartholome.htmSBaskerville, R. (1999). Investigating information systems with action research. Communicationsof the Association for Information Systems. 2(19).Brown, E. et al. (2001, January 8). Mobile learning explorations at the Stanford Learning Mobile Learning Design 34 Lab. Retrieved.http://sll.stanford.edu/projects/tomprof/newtomprof/postings/289.htmlCabrere, J.C. (2002). Third generation telephony: New technological supportfor computer assisted language learning. International Journal of EnglishStudies, 2(1), 167178.Chinnery, G.M. (2006). Emerging technologies. Going to the MALL: Mobile AssistedLanguage Learning. Language Learning & Technology, 10(1), 9-16.Chetty, M., Tucker, W. D., and Blake, E. H. (2003). Using voice over IP to bridge the digitaldivide — a critical action research approach. In Proceedings of South AfricanTelecommunications and Networking Application Conference (SATNAC 2003).SATNAC, Telkom, CD ROM Publication.Clean Slate (2008). An interdisciplinary research project. Stanford University, Stanford, CA.Retrieved May 28, 2008. http://cleanslate.stanford.eduDorado, S. (2006). Social entrepreneurial ventures: different values so different process ofcreation, no? Journal of Developmental Entrepreneurship, 11(4), 319–343.Goldenberg, C., Gallimore, R. G., & Reese, L. (2005). Using mixed methods to exploreLatino children’s literacy development. In Weisner, T.S. (Ed.), Discovering successfulpathways in children’s development: Mixed methods in the study of childhood and family life.Chicago: University of Chicago Press.Gradstein, M. & Schiff, M. (2006). The political economy of social exclusion withimplications for immigration policy. Journal of Population Economics, 19, 327-44.Eakin, E. (2003). How to save the world? Treat it like a business. The New York Times, 20December, p. 7. Mobile Learning Design 35 Easton, P. B. (2005). Literacy and empowerment: Raising key issues. Paper presented at the SixthMeeting of the Working Group on Education for All, UNESCO, 19-21 July 2005.Fleming, S., (2002). When and how should keyboarding be taught in elementary school? RetrievedMarch 20, 2007, fromhttp://www.eduteka.org/curriculo2/WhenHowKeyboardingSmith.pdfSHall, G., & Patrinos, A. (eds.). (2006). Indigenous peoples, poverty and humandevelopment in Latin America. London: Palgrave Macmillan.Hartviksen, G., S. Akselsen & A.K.Eidsvik (2002). MICTS: Municipal ICT schools – Ameans for bridging the digital divide between rural and urban communities. Educationand Information Technologies, 7(2), 93–109.Henfridsson, O. & Lindgren, R. (2007). Action research in new product development.Integrated Series in Information Systems. 13(2). pp.193-216.Joseph, S., Binsted, K., & Suthers, D. (Nov., 2005) PhotoStudy: Vocabulary learningand collaboration on fixed and mobile devices. IEEE International WMTE, 206-210.Joseph, S. & Uther, M. (2006). Mobile language learning with multimedia and multi-modal interfaces. In Proceedings of the fourth IEEE International Workshop onWireless, Mobile and Ubiquitous Technology in Education. (ICHIT '06).Kadyte, V. (May, 2003). Learning can happen anywhere: A mobile system for languagelearning. In J. Attewell, G. Da Bormida, M. Sharples, & C. Savill-Smith (Eds.),MLEARN 2003:Learning with mobile devices (pp.5051). London: Learning and SkillsDevelopment Agency. Retrieved March 20, 2007, fromhttp://www.lsda.org.uk/files/pdf/1421.pdfKatz, J.E., Sukiyama, S. (2006). Mobile phones as fashion statements: evidence from studentsurveys in the US and Japan. New Media & Society, 8(2), 321-337. Mobile Learning Design 36 Kiernan, P.J., & Aizawa, K. (2004). Cell phones in task based learning: Are cell phonesuseful language learning tools? ReCALL, 16(1), 71-84.Kim, P., Miranda., T., & Olaciregui, C. (2008). Pocket School: Exploring mobile technologyas a sustainable literacy education option for underserved indigenous children inLatin America. International Journal of Educational Development, 28(4). pp. 435-445 .Lennie, J., Hearn, G., Simpson, L., & Kimber, M. (2005). Building community capacities inevaluating rural IT projects: Success strategies from the LEARNERS project.International Journal of Education and Development using Information and CommunicationTechnology, 1(1), 13-31.Levy, M., & Kennedy, C. (2005). Learning Italian via mobile SMS. In A. Kukulska-Hulme & J. Traxler (Eds.), Mobile learning: A handbook for educators and trainers. London:Taylor and Francis.Martin, R. (2008). The balancing Act: Speed, agility versus cost. Stanford University,Stanford, CA. Retrieved June 21, 2008. from http://itunes.stanford.eduNorbrook, H., & Scott, P. (2003). Motivation in mobile modern foreign language learning.In J. Attewell, G. Da Bormida, M. Sharples, & C. SavillSmith (Eds.), MLEARN2003:Learning with mobile devices (pp.50-51). London: Learning and Skills DevelopmentAgency. Retrieved March 20, 2007, fromhttp://www.lsda.org.uk/files/pdf/1421.pdfOgata, H., & Yano, Y. (2004). Knowledge awareness map for computer-supportedubiquitous language-learning. In Proceedings of WMTE, 19-26.Parsons, D., & Ryu, H., (2006a). A framework for assessing the quality of mobile learning,In: Dawson, R., Georgiadou, E., Lincar, P., Ross, M., Staples, G. (Eds.), Learning andreaching issues in software quality, Proceedings of the 11th International Conference for Process Mobile Learning Design 37 Improvement, Research and Education (INSPIRE), Southampton Solent University, UK,April 2006, pp. 17–27.Parsons, D., Ryu, H., (2006b). A framework for assessing the quality of mobile learning. In:Proceedings of the 6th IEEE International Conference on Advanced Learning Technologies,Kerkrade, The Netherlands.Paredes, R., Ogata, H., Saito, N., Yin, C., Yano, Y., Oishi, Y. & Ueda, T. (2005).LOCH: Supporting informal language learning outside the classroom with handhelds,Proceedings of the IEEE Wireless and Mobile Technologies in Education, 182-186.Pea, R., Maldonado, H. (2006). WILD for learning. In: Sawyer, E.K. (Ed.), The Cambridgehandbook of the learning sciences (pp. 427-441). Cambridge University Press, New York.Psacharopoulos, G., & Patrinos, H.A. (1994). Indigenous people andpoverty in Latin America: An empirical study (World Bank Regional and Sectoral Studies).Washington, DC: World Bank.Reimers, F. (2000). Unequal schools, unequal chances: The challenges to equalopportunity in the Americas. Harvard University Press.Revelle, G., Reardon, E. Green, M.M., Betancourt, & Kotler, J. (2007). The use of mobilephones to support children's literacy learning. Lecture notes in computer science. 4744,pp.253-258.Roschelle, J. (2003) Unlocking the learning value of wireless mobile devices. Journal ofComputer Assisted Learning, 19(3) 260-272.Sharples, M., Taylor, J., & Vavoula, G. (2005). Towards a theory of mobile learning. In:Proceedings of mLearn 2005 Conference, Cape Town.Schmelkes, S. (2000). Education and Indian populations in Mexico. The Failureof a policy. In Fernando Reimers (Ed.), Unequal schools, unequal chances. Mobile Learning Design 38 Cambridge, Mass: David Rockefeller Centre for Latin American Studies, HarvardUniversity.Stringer, E. (1997). Action research. A handbook for practitioners. Sage publications, USA.Stromquist, N. (2005). The political benefits of adult literacy. Background paper for EFA GlobalMonitoring Report 2006.Tomei, M. (2005). Indigenous and tribal peoples: An ethnic audit of selected povertyReduction Strategy Papers, Geneva, International Labour Organization.Thornton, P., & Houser, C. (2005). Using mobile phones in English Education in Japan.Journal of Computer Assisted Learning, 21, 217228.UNDP. (2004). Human Development Report 2004: Cultural liberty in today's world.New York, UNDP.UNESCO. (1994). Education for all: Status & trends. Paris.UNESCO. (2006). EFA Global Monitoring Report 2006: Literacy for Life. Paris, UNESCO.UNESCO-OREALC. (2004). Education for all in Latin America: A goal within ourreach, Regional EFA Monitoring Report 2003. Santiago, UNESCO/OREALC.