ICTlogy

Paper Session: Gadgets

Impact of Low-Cost, On-Demand Information Access in a Remote Ghanaian Village
Clifford Schmidt, Trina Jean Gorman, Michael Shayne Gary, Andrew Azaabanye Bayor

Context: low income, low or none literacy, no electricity.

Try to see how to get on-demand information, which is difficult if you have to access a kiosk and wait for the operator to be there, for the information to be ready or available, etc.

The Talking Book is the the world’s most affordable, durable, audio device designed specifically for people who cannot read and who live without electricity. Local experts spread knowledge reliably and easily with no information loss. Rural teachers complement their lessons with interactive applications and audio books.

The Talking Book allows for listening to and recording of content.

How to implement it?

The chief of the village will be approached and he will provide contacts with the relevant people and institutions around.

Once the target users are identified, an average of 45′ training is needed to operate the device.

After a year of usage of the Talking Book to get information on agriculture, an impact assessment was performed in order to see how had their harvests changed and whether these changes had had any origin in the usage of the Talking Book. It appeared that most people had been using the Talking Book and applying its knowledge to their daily practices and that their harvests had significantly been better in comparison with those that had not used the Talking Book. The impact was even more evident indeed because most people applied the advice of the Talking Book only to a part of their crop, thus the comparison was even more easy to test.

One of the problem is that it had been younger people the ones that had been using more intensively (or at all) the devices. There was not, though, any difference at the education level: unlike what was expected, less literate people did not think that the device was for “smarter” people, but used it them too.

On-demand access to locally recorded information can have a positive impact and the Talking Book is a low-cost good option for that.

Managing Microfinance with Paper, Pen and Digital Slate
Aishwarya Lakshmi Ratan, Sunandan Chakraborty, Pushkar V. Chitnis, Kentaro Toyama, Keng Siang Ooi, Matthew Phiong, Mike Koenig

There are people that prefer paper and others digital supports.

• Paper: tangible artefact; handwritten pen-based entry; familiar form.
• Digital: automatic digitization, ease of aggregation and distribution; real time prompts and corrections.

86M women participants i 6M microfinance self-help groups across India, linked with banks, decentralized, autonomous, self-run. They are active in leading social, political and economic initiatives,but limited financial leverage/growth.

The problem is that they do all their accountability records on paper, which means that sending data to the central station takes a lot of time, that there are recording errors, calculation errors, legibility errors, data are incomplete, etc.

How to retain the familiar work practice while improving the system? Improve meaning locally digitising adn processing, real time prompts for error correction and completeness, establishing of a single point of entry… and keeping it low cost.

A pilot was developed: a digital slate where you write on paper, but where writing is recognized and digitized automatically. For the field trial, three methods were compared: paper-only, digital slate and touch-screen only.

Digital slate and touch-screen prover to increase the amount of “paperwork” done per day. The average meeting transaction recording time also decreased, even more in the case of the touch-screen.

Concerning the user, 88% of them stated that they preferred “the machine to speak” (the machine saying the figures and passbook entry) and not having a writer saying the figures and passbook entry.

The device can also be applied in or used for legal records, in educational testing, healthcare records etc.

Paper Session: Rarer Themes in Education

Beyond Strict Illiteracy: Abstracted Learning Among Low-Literate Users
Indrani Medhi, S. Raghu Menon, Edward Cutrell, Kentaro Toyama

Text-free user interfaces increase the success of use for a given amount of time training. What else is required for non-literate uses to reach the usage level of ICTs of literate users?

Videos have no text and thus do not require reading while providing text-like information.

In order to to perform an experiment, a community of female domestic helpers were chosen, with very low literacy levels, and to test whether videos on how to use a modern vacuum cleaner had any impact in the acquisition of skills by these illiterate women. Will users benefit from diversified examples as a way to learn abstract concepts?

Variants analysed were whether the users was or was not literate, and whether the user was or was not familiar with (a) the vacuum cleaner and (b) a specific vacuum cleaner. And videos included also these variables.

Diversified video (e.g. showing more than one type of vacuum cleaner) proved to be helping literate users, but not illiterate ones.

Beyond strict illiteracy, other aspects affected comprehension of video content: cognitive skills, social standing, intimidation by technology, visual organization, efficient processing of information, language taks, self-efficacy, etc. Even for tasks that do not require reading at all and where there is the context, there seem to be cognitive barriers that impede use in non-literate users.

Discussion.

Q: Won’t literate people have cognitive barriers too? A: Agreed. But technology and treatment of information imply a fool range of cognitive barriers that go from technological illiteracy to abstract thinking, etc.

Technology, Teachers, and Training: Combining Theory with Macedonia’s Experience
Laura Hosman, Maja Cvetanoska

Some factors behind the ‘computers in the classroom’ concept: technology changes but human nature does not; computers in the classroom… mission accomplished; major struggle in ICT4ED projects; Education and Psychology scholars theorising and writing; policy makers not listening… and as a result, teachers blamed over and over for tech project failures. Maybe the real problem is not acknowledging that innovation is a years-long process of change, not a one-time event; that teachers are key change agents but are often not treated accordingly; and that teachers need ongoing support and must be stakeholders in the innovation-adoption process.

Now, the issue of computers in the classroom has spread from developed to developing countries, with the added problem that (a) resources in developing countries are even more scarce but, notwithstanding (b) computers in the classroom are being introduced at an imprecedented speed and level.

Macedonia Connects is a USAID-led initiative to provide one computer lab per school in Macedonia, after the country succeeded at breaking the telecom monopoly and bringing affordable broadband wireless to the entire country. Prior to the technology deployment, all teachers were provided with technology and methodology training.

As most teachers’ concerns advance predictably, most of them can be addressed as they arise by leaders/change facilitators.

Key findings:

• 65% have not used a computer in class in previous two months
• 86% believe that the class is the place where to learn to use a computer.
• 72% use ICT for preparing teaching materials and tests.
• 51% spend a few hours a day with a computer.
• 30% use ICT for working with students.

Recommendations: set up a yearly ICT plan; involve teachers as stakeholders; recognize that change is a years-long process; don’t press for overnight success; support teachers in managing change.

SPRING: Speech and PRonunciation ImprovemeNt through Games, for Hispanic Children
Anuj Tewari, Nitesh Goyal, Matthew K. Chan, Tina Yau, John Canny and Ulrik Schroeder

MILLEE project: Mobile and immersive learning for literacy in emerging economies.

Pilot project in a school in California targeted to the Hispanic students (20 in total) with low English skills. Instead of mobile phones, it was decided that laptops would be used instead.

Challenges faced were key problems with English, issues with reading and writing, resistance to learning English, etc.

To do so, two games were designed (Zorro, based on Mario, and Voz.Guitar, based on Guitar Player) according to the needs and profiles of the students (that had previously been analysed). Movements required speech to be commanded and a speech recognizer was embedded so to tell whether the student was using the correct pronunciation.

two metrics were gathered: acoustic score gain percentages (measuring the improvement in the pronunciation of correct words) and word gain (correctly pronounced words). Score Acoustic and Word gains improved a little bit (though significantly) between control and treatment group.

Gender and pre-existing knowledge didn’t seem to play a role or be a factor.

Discussion

Ismael Peña-López: why pronunciation of English words was in English standards and not Spanish standards? Why (for surprise) put ‘ser-prize’ instead or ‘sur-prais’, which would have been the Spanish transcription? A: Some of the transcriptions were added ex-post and used the acknowledge standard. But, certainly, in future editions, there is a need to adapt the transcriptions to the linguistic realities of the target community.

Q: What was the teacher proficiency in English pronunciation? Q: The project was performed in a public classroom in California and had extended English teaching experience.

Paper Session: Technology Sharing in Education

MetaMouse: Improving Multi-user Sharing of Existing Educational Applications
Kurtis Heimerl, Janani Vasudev, Kelly G. Buchanan, Tapan Parikh, Eric Brewer

We use to design for individual users, but in disadvantaged regions, most computers are shared by several students. This ends up with a single student e.g. controlling the mouse. And it you do not control/use the mouse, the problem is not only that you cannot participate, but that you end up not learning (and not only not learning how to use the mouse).

The solutions so far just allow for multiple mice to be used, but they have problems with most games and educational material.

MetaMouse uses multiple cursors controlled by multiple mice and based on the key idea of “location-voting”: it is assumed that users agree on a button / option on the screen if their cursors are in the same location, thus requiring all (consensus) or most (majority) users to agree.

An evaluation was made with three scenarios: single mouse, MetaMouse based on consensus, and MetaMouse based on majority. Findings show that the majority variant enables laggard students to participate much more than with the traditional mouse variant and a little bit more indeed than with consensus.

MetaMouse, though, created discussion from “come to this” to “why this”, thus making everyone more involved. On the other hand, MetaMouse actively encouraged students to teach their peers. There was less fighting, more participation or equality.

Discussion.

Q: What happened with gender dynamics after the experiment? A: There was no evaluation on that. Nevertheless, the experiment was participated by extremely dominant girls too, so levelling might have worked both ways.

Evaluating an Adaptive Multi-User Educational Tool for Low-Resource Environments
Emma Brunskill, Sunil Garg, Clint Tseng, Joyojeet Pal, Leah Findlater

Help teachers provide quality education.

Work on interfaces & HCI proved that it could improve interactions using multiple inputs.

With MultiLearn System each students works on a different exercise using their own numeric keypad. Students compete to be the first to answer 12 questions correctly.

How to keep everyone engaged? How to prevent one student domination the session?

MultiLearn+ personalized the problem selection, including a dynamic selection of the problems according to the (right or wrong) answers provided by the student.

3 scenarios were then compared for analysis: control (group spelling computer game), MultiLearn (non-adaptive grup math game) and MultiLearn+ (adaptive grup math game).

The adaptive condition showed that game dominance was reduced. Reducing dominance is a side benefit of a feature designed to improve educational effectiveness. on the other hand, there was no significant effect of adaptivity or math software over control condition, though there was a good correlation on student performance. In summary, adaptive, multi-user software has potential to help improve education.

Collage: A Presentation Tool for School Teachers
Saurabh Panjwani, Aakar Gupta, Navkar Samdaria, Edward Cutrell, Kentaro Toyama

ICTs alone cannot impact learning by much; engaging teachers is necessary. On the other hand, ICTs can help good teachers teach better. How can software be designed to facilitate classroom instruction via a digital projector? What’s a good presentation software for teachers?

PowerPoint and variants were never designed for teaching, but for general-purpose presentations (though it is used by +6M teachers worldwide). Document cameras are a digital substitute for overhead projectors (OHP), but they can not be programmed and they are quite costly.

Collage aims at being specifically designed for teaching in schools.

A first analysis of urban public and private schools in India showed that teachers used intensively paper and the blackboard, they had a strong desire to use PCs for teaching and, despite the desire,usage was low, especially because of lack of tools to prepare content.

Collage features: textbook page interactions (zoom, highlight, overlay), inking (through graphics pad), scratch space.

Collage has few features, but it is very easy to use and even allows real time authoring. Indeed, textbook scans proved to be highly beneficial, especially because of a better recall of visual information.

Key takeaways:

• Using paper in digital presentations is valuable. Both teachers and students seem to benefit.
• Teachers like to author content in real time. Collage caters to this liking.

Deployment of Collage still a challenges. Content assembly takes time, and using it has the associated hardware cost of the PC and the projector.

Notes from the the II Encuentro Internacional TIC para la Cooperación al Desarrollo (Development Cooperation 2.0: II International Meeting on ICT for Development Cooperation) held in Gijón, Spain, on February 10-12th, 2009. More notes on this event: cooperacion2.0_2009. More notes on this series of events: cooperacion2.0.

Round table: Innovative Uses of Mobile ICTs for Development

Merryl Ford, Emerging Innovations Group of the Meraka Institute of the Council for Scientific and Industrial Research (CSIR).

Mobile voice Wikipedia (MobiLed: you send an SMS to the mobile Wikipedia with the name of an article, and a voice reads you the whole article on your mobile phone.

Remote tutors that, through SMS, help kids in their Maths homework.

We have to de-skill the process of mobile application creation, thus why at Meraka they’re creating open source platforms for people to code mobile applications without much knowledge on developing applications.

Kentaro Toyama, Microsoft Research India (MSR India)

(disclaimer: he’s going to play the devil’s advocate)

The focus on putting development first, and then technology. If we speak about “M4D”, we’re putting technology first. Unless you have a strong interest in mobile phones (e.g. you’re working for Nokia) you should focus on what’s available, not just on a specific technology. Development is about human and institution capacity.

On the other hand, there’s also even more simple and broadly accepted technology than mobiles: television, community radio… so we should also focus in these if our arguments are cost, simplicity, broad usage and so.

There’s hype around M4D as there was hype about telecentres 15 years ago. It just does not make sense to fund projects that explicitly (ex ante) have to be run by mobile phones. And this happens. And this is hype.

Oleg Petrov, e-Development Thematic Group of World Bank

(in kindest answer to Toyama) The mobile phone is the new sibling, the new tool in the development toolbox. Enthusiasm vs. hype. It’s just enthusiasm, and other technologies, just like siblings, are “jealous” of the newcomer. But it’s a powerful tool indeed. We have not for forget about everything but mobile phones, but as a new tool, it deserves special exploration to determine its real potential and weaknesses.

The World Bank has plenty of projects that follow this excitement to test M4D: for health, for education, etc. We need a community of practice, ways to test this technology. Raise awareness and also move forward in the next direction of building toolkits, making it simple.

If you just look at the human side of development, you’re likely to miss (or not be up-to-date) state-of-the-art technological developments that might give you hints or ideas on how to solve human problems.

Jan Blom, Nokia Research Center – India

Anecdote: 6 months ago, a cab driver in Bangalore, using the mobile phone to SMS and as a GPS, but knowing nothing about what e-mail was. So: M4D is absolutely real.

There’s a dire lack of public information available (sometimes it does not even exist). M4D can focus on making this public information available, in providing location based services. There is much utility in taking local data, uploading to a central server, and publish it online (like Ushahidi does).

Stéphane Boyera, Device Independence Working Group of W3C

(also answering Toyama) It’s not that it’s a hype, it’s that technologists are approaching development in their daily lives, which is new and it’s great. Of course, when coming from the Development Cooperation field one must focus on humans, but the thing is that techies are approaching humans through tecnology, and the specific technology of their specific fields.

Q & A

Najat Rochdi: we have to know all the technologies available to be able to make the best decisions.

Stijn Vander Krogt: what is the role of governments in M4D? Isn’t it to analyse all that’s out there? Petrov: absolutely, this is one of the key roles of the World Bank, to provide advice on what can be used to solve any kind of problem. Thus why organizations have to know, and raise awareness, of the different applications of tools for human development.

Manuel Acevedo: What do we do when some laptops are really cheap and simple, and some mobile phones become increasingly complex and expensive? Thus, our task is to inform people of all the options available. Can we build multidisciplinary teams (as in research) at the government level? Can we build multi-institutional approaches in ICT4D?

Q: are we confusing needs for development? Are we artificially generating needs for gadgets? Aren’t we trying not to develop rural areas, but the broaden the target market of telcos?

Notes from the the II Encuentro Internacional TIC para la Cooperación al Desarrollo (Development Cooperation 2.0: II International Meeting on ICT for Development Cooperation) held in Gijón, Spain, on February 10-12th, 2009. More notes on this event: cooperacion2.0_2009. More notes on this series of events: cooperacion2.0.

Research for Development at Microsoft Research India
Kentaro Toyama. Microsoft Reserarch India

Microsoft Research India is a computer science research lab focused in technology for emerging markets.

They promote the ICTD Conference.

Methodology:

• Immersion: ethnography; qualitative social science
• Design: iterated prototyping; design, engineering
• Evaluation: randomized control trial; economics
• Implementation: partnership; political science (this point relating implementation usually transferred to partners)

Is technology always worth it? It depends: if the increase of productivity is smaller than the cost of technology (applied to achieve this increase of productivity), then, the cost-benefit analysis results in negative returns, even if productivity increased. Microfinance in developing countries seems to be a clear example of this, at least at the front-end (though further research is needed because in some cases it might pay back).

Examples:

• Microfinance and Technology: raise of productivity by adding technology. Might not be cost-effective.
• MultiPoint in Education: increasing education quality with several kids using same PC and by means of multipoint devices (i.e. each child their own mouse to interact with the screen). No significant impact on education
• Digital Green for knowledge transmission in the field of agriculture: storyboarding with video. Results: 7 times more adoptions of new techniques, 10 times more cost-effective.

Key lessons

• Development first, then technology. The goal is not to close the digital divide, but to achieve a development goal.
• Expend time with communities, not with “experts”.
• Multidisciplinary of teams, not individuals.
• Quality through great people, not processes
• Sustainability is case-by-case, there’s no magic bullet. And sustainable models are often very different among them.
• Impact as the goal, ideology has to be set aside

Notwithstanding, there are three counterexamples (e.g. the mobile phones) for the first three dots in this list!

Q & A

Q: Why should Microsoft invest in basic research of this kind? A: Knowledge is created, even if (in principle) roughly related with your direct interests. It’s good for the morale of the company and their workers. Being concerned about the future of the World makes you aware of the future of your own company. And, sometimes, Microsoft (Microsoft-not-research) is up to do the follow up and scaling of a prototype or pilot project.

Q: is research tied to business models? A: no, it’s basic research bound to find impact. Implementation (and their related business models) come after. But if the impact is positive, and there’s a justification to go on with implementation, business models will come.

Q: How are research projects chosen? A: It’s up to the researcher. There’s an abolute trust on their criterion.

Notes from the II Encuentro Internacional TIC para la Cooperación al Desarrollo (Development Cooperation 2.0: II International Meeting on ICT for Development Cooperation) held in Gijón, Spain, on February 10-12th, 2009. More notes on this event: cooperacion2.0_2009. More notes on this series of events: cooperacion2.0.

How do we go forward in the field of ICT4D R+D+i?

Florencio Ceballos, telecentre.org

• ICT4D are a clear niche that can grow outside the circuit of development issues
• Capacity building happens locally, and this means building confidence, trust.
• Institutional independence has to be promoted to enable real capacity building.
• Focus on networking: promoting open networks for capacity exchange

It’s not as much as how you design agendas, but how you make them evolve, how to shift the paradigm. And this shift of paradigm is towards openness.

Caroline Figueres, International Institute for Communication and Development

There is a need for a research to ground some “evidences”, and showcase successes in the field of ICT4D under the rigour of scientific analysis.

People in the South should be put in the agenda of ICT4D research, as most of the output is targetted to developing countries.

Co-creation (e.g. in the sense of Don Tapscott’s Wikinomics) is a very powerful concept. Capacity building can be enabled this way by means of knowledge workers co-creating together.

Kentaro Toyama, Microsoft Research India (MSR India)

How to do formal research in ICT4D? Several steps:

1. Immersion. Ethnography
2. Design, involving people, where technology is just one component and a cost-effective one
3. Evaluation, including finding statistical significance on the impact of a specific project or action

It’s a good idea to break the link between funding and the research agenda. The researcher should be able to pursue their own interests and not be tied (or upset) to the need for funding.

Experience in research might be as important as (or even more) than experience in development. Accuracy of the scientific process is crucial.

Andrés Martínez, EHAS Foundation.

Evidence has to be demonstrated to convince policy-makers and funding institutions that some actions are to be taken and deserve being supported (politically or economically).

• Research is needed in the impact of ICTs in welfare, health, education
• But also, research is needed on how to provide appropriate and cost-effective infrastructures, as most communities just do not have access to either hardware or connectivity
• Sometimes the context is unknown. Thus, research should focus not only on the impact of a specific project, but on what the context (sociocultural, health, education, economic) is.
• Research on services.
• How to measure empowerment and mainstreaming of technologies in specific communities and sectors (e.g. the Health sector)

The only way to promote research in the field of Development and ICT4D is to foster publication of research results in indexed publications. Despite the interest of the topic, if the work is “well done”, then it can be published. It is highly relevant to find the problem you want to deal with your research, more important than finding “the” solution.

And diffussion is absolutely worth doing it. On the one hand, results of the projects and the research undertaken. On the other hand, not only information about the results, but knowledge transfer through assistance, direct training, formal education, especially to achieve multiplier effects.

Merryl Ford, Emerging Innovations Group of the Meraka Institute of the Council for Scientific and Industrial Research (CSIR).

There’s sometimes resilience to empowerment. Capacity building is not only about specific (digital) skills, but also about changing mindsets.

• Slogan on disabilities in SouthAfrica: Nothing about us, without us. We need to make sure that we don’t do things “for” people but “with” people. Africa should take ownership of its development agenda.
• Interventions should be simple
• The cellphone is the PC of Africa
• Sustainability, replication, massification. A pilot needs to be scaled at any stage.

Q & A

Q: research on impact… is a real need or an imposed “need” of the inner structure of development cooperation, projects, agencies and so? Ceballos: The need to measure impact is real. Many policies are put into practice based on intuition, on vision. So we do need to evaluate these policies to support or reject such intuitions. Martínez: short-run projects are difficult to analyze accurately, as there’s no time to do it properly. A solution would be that everyone involved in the projects collected data and helped to analyze it.

Q: How do we cope about the cost of maintenance of cellphones in rural areas? A: There are alternatives (e.g. via radio) that do not charge per call… but the maintenance of the whole network does have a cost. Certainly, it’s not a matter of absolute costs, but a matter of cost-benefit analysis, seeing whether the project is worth running it and find out how to support the overall costs.

Q: How do we put social research together with tecnology research in development related research? A: The problems that research has to face have to be far ahead enough. And they require plenty of time. In this sense, everyone involved in ICT4D should be in a same conversation, to gather all sensibilities and be able to look far in the horizon.