From UBC Wiki

Solo Edit by Nicholas Graves

Minimally Invasive Education is a pedagogic method developed by Sugata Mitra that attempts to use the learning environment to intrinsically motivate students to group learning with little to no inducement and support from the teacher.


Minimally Invasive Education (MIE) arose from Sugata Mitra’s study of and experiments in Indian educational practices. In particular it arose out of his “Hole-in-the-Wall” experiments that he conducted in rural Indian villages, as well as other rural villages throughout Africa. Mitra defines Minimally Invasive Education as a “pedagogic method that uses the learning environment to generate an adequate level of motivation to induce learning in groups of children, with minimal, or no, intervention by a teacher.” (Mitra, et al 2005, p. 409) The name borrows the term “Minimally Invasive” from the medical term minimally invasive surgery and follows the idea that education can be unobtrusive in individuals’ lives.

One of the most striking and important features of MIE is that it seeks to build an education system based on the student’s motivation and is focused on voluntary participation. This is in direct opposition to most pedagogies, and most current education systems that rely on mandatory attendance, structured curriculum, and the teacher’s ability to disseminate and regulate learning. Mitra has based MIE around the idea that “… any learning environment that provides an adequate level of curiosity can cause learning among groups of children.” (Mitra, et al 2005, p. 410)

Another key feature to MIE is that it allows and encourages students to form their own social networks within their learning environments. Expanding on Mitra’s work, Dangwul and Kapur, have noted that the real strength of the MIE model is that it relies on both individual and group thinking. The group setting provides motivation and enthusiasm, creates a collaborative environment where students can learn from one another, and develops social practices wherein students are more active learners and more receptive to learning from their peers. (Dangwul & Kapur, 2009, p.19) On the individual level MIE allows a student to consolidate learned information by providing them with time necessary to assemble and strengthen what they have studied. (Dangwul & Kapur, 2009, p.20)

The main demographic that Mitra has focused on are students in rural environments that lack the infrastructure to provide a well-developed education system to participate in. The success of Minimally Invasive Education should not, however, be restricted to rural communities as the ideas behind it are applicable to, and can be used to complement current education systems as well. In this context MIE should be seen as a way to get students to initiate and continue the process of learning outside of the classroom setting. (Mitra & Rana 2001, p.223)


There are many reasons to deploy Minimally Invasive Education, but the primary reason for its development was to create effective teaching methods for remote communities. Remote communities suffer from two main problems that directly impact the education of students: teacher quality and teacher motivation. In their study “Effects of remoteness on quality of education: A case study from North Indian schools” Mitra, Dangwal and Thadani, discovered that quality of education in Indian schools decreased with the distance from the nearest urban area. They also discovered that this was not related to the number of students per teacher, nor the number of students per classroom, but had deeper connections to the training of teachers and the motivation of the teachers. The purpose of Minimally Invasive Education is to overcome these obstacles by ignoring them, and providing educational technologies that allow the learner to pursue educational goals independent of teachers. It is further noted that these technologies would also be useful in complementing formal education systems by giving them a greater control over their own learning experience. (Mitra & Rana 2001, p. 223)

Hole-in-the-Wall Experiment

Computer Skills

  • Experiment Design

One of Sugata Mitra’s initial experiments with Minimally Invasive Education was in the acquisition of computer literacy in students 6-14 years old. He postulated ‘if given appropriate access and connectivity, groups of children can learn to operate and use computers with none or minimal intervention from adults”(Mitra, et al 2005, p.2)

To test his hypothesis he built twenty-one learning stations in playgrounds throughout India, from the Himalayas in the north to the tip of the Indian peninsula in the south, and from the Ganges Delta in the East to the Rajasthan deserts in the west.[1] (Mitra 2005, p.72)

Most if not all of the students in these villages had never seen a computer before. Other than a caretaker appointed to turn on and off the computer everyday no other assistance was given in the acquisition of computer skills. On the first day of instillation in each location fifteen students were chosen at random and tested on their computer skills, these students would act as the focus group. After the testing the students were left alone for nine months to interact as much or as little with the computer as they wanted. At the end of the nine months the fifteen focus group students and another fifteen students, selected on frequency of usage, were tested on their computer literacy. (Mitra 2005, p.77)

  • Participants

The students of the villages were divided into two groups, Focus groups and Frequent Users. To test the results students from nearby villages were chose to act as control groups, and two groups of traditional learners were formed from schools in New Delhi, India.(Mitra 2005, p.76-77)

  • Design of Structure

Brick enclosures were built to hold the learning stations and provide protection from the elements. The computers were put into the enclosures at such a height to emphasize usage by students and make it uncomfortable for adults to use. They were also placed in safe public areas to limit vandalism, theft and usage of the computers to access undesirable material. (Mitra 2005, p.73-5)

  • Testing Procedure

To accurately assess the knowledge of the students a new test called the Icon Association Inventory was developed. The test was devised to assess the user’s ability to determine the function of common icons on a Windows system. The test included seventy-seven commonly used icons and the user was asked to explain their purpose. Before implementing the test it was given to adults in computer related occupations to decide its usefulness. Through these early tests it was discovered that users who took the test were able to identify icons in software that they used frequently independent of whether or not they actually used the icons. In this way it was determined that the test provided a good indication of a users computer literacy. (Mitra 2005, p.76)

The test was given to the Focus Group at the beginning and at the end of the experiment, but was only given to the Frequent User Group and the Control Group at the end. The Traditional Learner Groups were also given the test at the beginning and the end of the experiment. It was decided not to give the test to the Control Group before the end of the experiment to avoid influencing their behavior between testing periods. (Mitra 2005, p.76)

  • Results

The Focus groups’ initial testing on day one of the experiment produced a national average score of 6.65%. By the end of the experiment the national average score of these students had improved to 43.07%. The Frequent Users group produced a national average score of 43.73% and the Control group produced a score of 6.94%. The two groups of traditional learners improved form 10.44% to 35.96% and 11.96% to 49.17% respectively during the same period of time. (Mitra 2005, p.78)

From these results Mitra concluded that MIE initiatives like the “Hole-in-the-Wall” experiment appear to be effective and low cost methods of developing computer literacy in regions were traditional forms of education are limited or non-existent. (Mitra 2005, p.79)

Hole-in-the-Wall Experiment – Math Skills

  • Purpose

Parimala Inamdar and Arun Kulkarni recognized the successes of the “Hole-in-the-Wall” experiment in developing computer literacy, and decided to see if it could also be used to improve performance in Indian school examinations. These exams are conducted twice a year in the subjects of English, Science, and Math. In developing their own experiment they hypothesized:

“If given appropriate access, connectivity and content, groups of children can learn to use computers and the 
Internet to achieve a specified set of the objectives of education, with none or minimal intervention from adults.” And “Academic performance will be impacted by the frequency of use of the MIE kiosk.” (Inamdar & Kulkarni 2007, p.172)

  • Experiment Design

The experiment took place in the Sindhudurg District of India and involved two villages, Shirgao and Kuvle. The Shirgao School had an MIE kiosk in its playground while the Kuvle village had no computers at all. In total there were 161 students from age 13-14 in both villages, and all were included in the study.

The students were given three tests, Raven’s Standard Progressive Matrices, an intelligence test, Cattell’s High School Personality Questionnaire, a leadership and creativity test, and a Frequency of Usage Test, an informal survey of how often a student used the kiosks. The intelligence and personality tests were given to determine any differences between the two villages and subsequent student achievement. The Frequency of Usage Test allowed the researchers to divide the students from Shirgao into frequent users and infrequent users and give them a better sense of the impact of the kiosk on their learning.

The results from school examinations were collected from the exam period before the instillation of the kiosk in March 2002, and two and a half years after the instillation in October 2004. The student data was then organized into two data sets, Shirgao students and Kuvle students, and frequent Shirgao users and infrequent Shirgao users. (Inamdar & Kulkarni 2007, p.172)

  • Design of Structure

Each of the kiosks in the Sindhudurg district had two computers. They were built into a similar structure as in the previous experiments. The computers were connected to the Internet and also had educational material installed for offline use. These included educational games and videos in the subjects of Math, English and Science. The Math games had different activities on learning numbers, shapes, sizes, basic addition, subtraction, division, multiplication and basic algebra. The computers were all installed with English language Windows operating system. (Inamdar & Kulkarni 2007, p.171)

  • Results

The experiment returned mixed results in how the MIE kiosks impacted student achievement. The 2002 data showed that both Shirgao and Kuvle had similar levels of achievement. The 2004 data also showed a similar level of achievement between the two villages, with the exception that students from Shirgao were achieving slightly higher in sciences than Kuvle. The researchers determined however, that the difference was too slight to attribute to the MIE kiosks. (Inamdar & Kulkarni 2007, p.174)

In a comparison of frequent users and infrequent users in Shirgao no noticeable difference was found in English and Science exam marks. There was though significant difference between frequent and infrequent users in the Math exam marks, with frequent users nearly doubling the scores of infrequent users. With these results the researchers believe that the kiosks have shown to have a definite impact on the students’ math abilities. (Inamdar & Kulkarni 2007, p.176)

In reflecting on the final scores and after discussion with local teachers Inamdar & Kulkarni believe that the difference in content and teaching style played a role in why no significant change was seen in the Science and English marks. These two subjects in India are taught mainly through memorization and the exam is focused on answering textbook questions. They speculate that the MIE kiosks are not very good at teaching this type of learning. The math scores on the other hand, where problem solving plays a much more important role, could have been impacted because the kiosk allowed students to strengthen their basic mathematic skills, and as a result improve their overall problem solving skills. They concluded that the end results do show that MIE kiosks can impact student achievement if the kiosks are calibrated to supplement the school curriculum and assessment procedures. (Inamdar & Kulkarni 2007, p.177)

Improving Pronunciation

  • Purpose

In India the ability to speak English is one of the main determinants in living standards and occupation. Most parents in India send their students to private English language schools in the hopes that they can attain a better life. Many students do learn English, however they also develop strong accents that make them difficult to understand. One of the main factors in the development of this accent is that the local teachers they learn from also have an accent. This experiment set out to test two hypotheses, first whether the measurement of words correctly recognized by a Speech Recognition Program parallels human judgments about pronunciation, and to see if given speech recognition software if students can instruct themselves without a teacher to improve their pronunciation. (Mitra et al 2003, p.76-7)

  • Location

A school was selected in the slums of Hyderabad, India that was populated mainly by students from low-income families. The student’s mother tongue was either Urdu or Telegu, but the school’s curriculum was taught in English and most students could read and speak with different levels of competency. (Mitra et al 2003, p.77)

  • Experiment Design

Eight boys and eight girls were randomly selected and organized into four groups. The students were given a computer that had a speech recognition program, Dragon Naturally Speaking, and an English language learning program,Ellis, plus the appropriate hardware to use these programs. The computer also included four movies, The Sound of Music , My Fair Lady, Guns of Navarone and The King and I . A timetable was also developed to give each group access to the computer three hours a week for a five-month period.

The students were shown the basics on the speech recognition software and where to find the movies. The only instructions they were given was to improve their English pronunciation by reading their English textbooks into the Dragon Naturally Speaking and to get the program to understand what they were saying. They were told that they could use Ellis and or watch the movies if they wanted to. They were also encouraged to share information between the groups, but other than maintenance activity there was no organized adult intervention in how they spent their time. (Mitra et al 2003, p.78)

  • Testing Process

The experiment, started in September 2002, began with the students reading a passage from their textbooks into the Dragon Naturally Speaking speech recognition software. These readings were also video taped so that the human judges could watch and rank the pronunciation of the students. At the start of each month a new passage was added to the testing, until January 2003 when the students were reading four passages and the experiment concluded. The rationale behind having the students repeat the passages was so that the first passage, which they had read four times, could be compared to the fourth passage which they read for the first time in January.

A control group was necessary to verify the intentions of the experiment so sixteen students were randomly selected from the school that the experiment participants attended. These students were asked to read three passages from their textbook into Dragon Naturally Speaking and were video taped. Since the control group and the experimental group went to the same school the only difference between the students was that the experimental group had worked on the computer for five months. (Mitra et al 2003, p.78)

  • Data Analysis

The analysis of the data took place at the end of the experiment. Four judges randomly observed the video taped recordings of the students and gave a subjective assessment of the students’ clarity on a ten-point scale, ten being high and one being low. To make sure there was consensus between the judges Kendall’s coefficient of concordance was determined for the readings. Each student was also ranked for each reading on the percentage of words correctly recognized by Dragon Naturally Speaking. Kendall’s coefficient of concordance was also calculated for Dragon Naturally Speaking since it was also used to assess student performance.

To avoid bias in the analysis of the control group four new judges were selected. They were provided with forty-eight video clips, sixteen from the experimental group recorded in September 2002, sixteen from the control group recorded n January 2003 and sixteen from the experimental group recorded in January 2003. As with the experimental group, the judges were asked to mark the recordings on a ten-point score, and Kendall’s coefficient of concordance was calculated. (Mitra et al 2003, p.80)

  • Results

The experiment’s results show that the experimental group’s results increased from thirty percent in September 2002 to seventy-one percent in January 2003. The control group on the other hand achieved an average score of thirty percent. The results for the human judges and Dragon Naturally speaking differ in how they scored clarity, but both scores when graphed produce a similar curve indicating that there was agreement in the improvement of the student clarity. Mitra, Tooley, Inamdar, and Dixon note that it is not possible to identify whether these improvements came from using the speech recognition software, the English language learning software, or the movies, but conclude that using similar interventions students can acquire any kind of accent depending on how the intervention is constructed. (Mitra et al 2003, p.81-2)


To test the limitations of Minimally Invasive Education Mitra decided to develop a MIE kiosk to teach molecular biology to rural students in Kalikuppam. The task seemed phenomenally difficult, but within seventy-five days of unsupervised learning the students had a mean score of thirty percent on the government developed test. Students of the same age in the wealthy regions of New Delhi had a mean score of fifty-three percent. The difference between the two scores illustrated to Mitra that without a mediating force behind the education, the students in Kalikuppam would never be able to surpass those from New Delhi. (Mitra & Dangwal 2010, p.681)

There was no qualified molecular biology teacher in the vicinity of Kalikuppam to tutor these students and help them raise their scores. Lack of educational infrastructure however, was one of the reasons Mitra developed MIE in the first place, so he sought new ways to tackle this problem. Mitra hired a mediator to praise the students and encourage them to go deeper into their investigations. The mediator had no knowledge of the subject and had no training as a teacher, she was told to just act like a grandmother and praise the students on their achievements. As a result of this interaction the students of Kalikuppam were able to match the scores of students in New Delhi. (Mitra & Dangwal 2010, p.680)

Minimally Invasive Education is not meant to replace teachers, but to reach students who have little to no educational access. One of its main limitations is the lack of a knowledgeable teacher to help guide the students, but this is also one of its strongest features as well as it allows students to form learning communities and engage with learning materials at their own pace. Mitra has begun to investigate ways to blend the teacher into MIE learning through mediation. His most current experiments seek to combine Skype with the MIE kiosks so that adult mediators can videoconference with the students and compensate for the lack of a teacher. These mediators seek to augment the MIE kiosks and improve their outcomes through encouragement and support. (Mitra & Dangwal 2010, p.683)

Further Reading

“Hole-in-the-Wall” Website

Sugata Mitra shows how kids teach themselves - TED Talk Video

Sugata Mitra: The child-driven education - TED Talk Video

An Interview with Sugata Mitra

Sugata Mitra's Blog


Dangwal, Ritu and Kapur Preeti (2009) “Learning through teaching: Peer-mediated instruction in minimally invasive education”. British Journal of Educational Technology 40(1): 5-22

Mitra, Sugata and Rana, Vivek (2001) “Children and the Internet: Experiments with minimally invasive education in India”. British Journal of Educational Technology 32 (2): 221-232

Mitra, Sugata et al (2003). “Improving English pronunciation: an auto- mated instructional approach” Information Technologies & International Development, 1, 1: 75–84.

Mitra, Sugata et al (2005) “Acquisition of computing literacy on shared public computers: Children and the ‘Hole in the Wall’”, Australasian Journal of Educational Technology, 21(3): 407-426.

Mitra, Sugata (2005) “Self organizing systems for mass computer literacy: Findings from the ‘Hole in the Wall’ experiments”. International Journal of Development Issues, 4(1): 71-81

Mitra, Sugata, Dangwal, Ritu and Thadani, Leher (2008) “Effects of remoteness on the quality of education: A case study of North Indian Schools”. Australasian Journal of Educational Technology, 24(2): 168-180

Mitra, Sugata and Dangwal, Ritu (2010) “Limits to self organizing systems of learning – the Kalikuppam experiment”. British Journal of Educational Technology 41 (5): 672-688

Inamdar, Parimala and Kulkarni, Arun (2007) “‘Hole-in-the-Wall’ Computer Kiosks Foster Mathematics Achievement – A comparative study”. Educational Technology & Society, 10 (2): 170-179