This is one of the final papers from 2011 when I was at Yale University. The One Laptop Per Child program was a very exciting initiative to get cheap and durable laptops to the world’s poorest children. However, no empirical studies had been done on the effectiveness of the program in improving educational outcomes. I identified a way to test this using existing and available data in Peru and set up the econometric experiment to test the impact of the laptops. Below is an excerpt. Read the full paper here

­­The educational impact of the XO computers in Peru

It’s not a laptop project. It’s an education project. – Nicholas Negroponte

The One Laptop Per Child (OLPC) program has tried to bring information technology to schools in developing countries. The idea behind the program is that children’s access to computers and the internet should be a right instead of a privilege. OLPC has therefore created a cheap, durable laptop, and has tried to disseminate it across the developing world. At the beginning of this year Uruguay proudly announced that their objective of giving one XO laptop to every child in Uruguay had been met, a process which cost 0.25% of the country’s gross domestic product. Similarly, Antigua and Guyana have recently signed deals to bring one laptop to every child in their respective nations.

To date, there is no rigorous empirical investigation of whether children who use the XO laptops in their education perform better academically than those who do not. This is an essential bit of information in calculating the cost-benefit analysis of the program. This project seeks to answer that particular question. Matching XO laptop deployment data from Peru with standardized test scores, I will find whether or not the use of these machines has any educational effect on the children that use them. Although the laptops are relatively cheap, they still represent a significant cost for governments who have to choose how to best allocate their small budgets. Given the qualitative analysis of the OLPC program put forward by Derndorfer (2010) which highlights the lack of infrastructure, teacher training, and government support, I suspect that the effect will be at best positive but insignificant. This would not mean that OLPC is a bad idea altogether. Instead, different arguments need to be used to support the mass distribution of laptops throughout schools.

The OLPC program and Peru

OLPC is a non-profit organization whose mission it is to deliver cheap laptops to the world’s poorest children for the purposes of education. This project was born at MIT in 2005, and by 2006 the XO laptop, at the time known as “the $100 laptop,” was released. The laptops were designed to endure harsh climate and treatment conditions. They could be charged through a variety of means, like solar power or motion power as well as through regular electrical outlets. The laptops were designed to make internet connectivity possible, so the wireless mesh network was introduced, such that every laptop with an internet connection would act as a router for other laptops which did not have a direct connection. Lastly, the laptops were not installed with Windows, choosing instead various forms of Linux like Sugar, the Operating System used in Peru. Although this presented a reduction in production costs, it also presented a challenge for teachers who had limited experience with computers and had only been exposed to Windows systems.

Mass production and deployment ensued. OLPC has been criticized for its questionable development priorities, as many believe that there are more fundamental development projects worth undertaking, such as the provision of clean water. Furthermore, OLPC does not involve itself with the implementation of laptops into the countries’ educational curricula. They have a “walk away” policy, since OLPC chairman Nicholas Negroponte believes that the technology is so easy to use that “you can give kids XO laptops and just walk away.”[1] Implementation, then, is left up to the Ministry of Education of the recipient country. XO computers are provided for school use, but they are also intended to serve as home computers. Indeed, it is expected that other members of the family, such as parents, will also benefit from having a computer in the home.[2]

This drive-by approach has important consequences. Among other things, there is a clear lack of communication between OLPC and the Ministries of Education. For example, the stated aim of OLPC is to provide a laptop to all primary students in Peru, a number they claim is around 2.7 million. [3] According to the Peruvian Ministry of Education’s census data, however, there over 3.7 million students enrolled in primary school.[4] The OLPC target is mis-calibrated by nearly 40%. All the success reports should thus be viewed under that light. Second, as Derndorfer (2010) describes, the absence of OLPC staff in the recipient country creates an infrastructural vacuum where no support exists for teachers who have to learn how to use, fix, and teach with the laptops.

OLPC nevertheless entered Peru in early 2008 with 260,000 laptops after a brief pilot test in the small village of Arahuay.[5] By 2011 there were over 850,000 XO laptops in over 8,000 primary schools. The regions targeted were Lambayeque, Chincha, Puno, Huanuco, and La Libertad. Peru is the fourth most populous country of South America with a population of nearly 30 million inhabitants. Although Spanish is the national language (84% of population), roughly 14% of Peruvians speak Quechua and 2% speak Aymara. The XO laptops, however, only operate in Spanish, and any command required to fix common issues is in English (Derndorfer: 2010). At just over 92%, Peru has one of the highest literacy rates in the region,[6] and ranks relatively high on school enrolment rates both in urban and rural areas, both for males and females.[7]

Previous Research on the Role of Technology in Education

There are many theories that attempt to explain what effect computers may have on educational outcomes such as test scores. Introduction to computers is an introduction to technology, so newly-acquired computer skills may be reflected in more efficient work in terms of time and quality. Furthermore, educational software and the internet provide users with an access to large troves of information that might otherwise be unavailable. Aubrey and Dahl (2008) believe that ICT should even be used in the Early Years Foundation Stage, which is to say children below the age of five, a claim supported by Sweeney and Geer (2010).

The assumption here, though, is that computers will get used for educational purposes. However, gaming and surfing the internet are among the most frequent activities of computer-using children. This feeds a common fear that children may become addicted to the computer, which is linked to repetitive strain injuries, eye strain, increased risk of obesity, and decreased social involvement. Time spent on a computer, then, might be time that may be better used by the child from a development perspective, such as by spending time outside and building personal relationships. It may also be time better used from an academic perspective if the child can focus better and work more efficiently if she does her homework by hand.

Is education technology, then, academically effective? Schachter’s 1999 review of the effect of technology in student achievement finds positive results like faster learning rates or higher student satisfaction in the classroom. Regarding achievement, the results are mixed. Some of the studies reviewed find positive results on achievement, and many of these tended to be comprehensive educational programs which included the use of technology. Many also did not use standardized tests. Other studies, like Macho (2005), find no significant effect. There seems to be evidence, however, that technology is more effective when learning objectives and the focus of technology use are very specific.

The results seem to be equally mixed in the 21st century. O’Dwyer et al (2005) find that language arts and writing scores increased in Massachusetts 4th graders when computers were used to write papers. Scores decreased when computers were used for other tasks (such as preparing presentations), and students’ recreational use of technology at home was negatively associated with learning outcomes. However, they were not able to randomize their samples and therefore fail to resolve endogeneity issues.[8] The same authors in 2008 find that computer use had no significant effects on standardized test scores, even when controlling for recreational use of computers.

The story seems to be different with older students. Fairlie and London (2009) find that a randomly selected group of community college students who received free laptops achieved better educational outcomes than the control group, though the effects were relatively small. Students initially living farther from campus benefitted more from the free computers, suggesting that laptops grant higher flexibility in completing tasks. In a later study (2010), Fairlie et al. find that teenagers who had access to computers at home were more likely to graduate than those who didn’t, although in their analysis they do not convincingly remove omitted variable bias, which is also present in Fairlie (2005) when he finds an association between computer ownership and school enrolment. [9]

Nevertheless, the general consensus in the literature is that the effects of educational technology are mixed (Palozzi and Sppradlin: 2006). This makes a cost-benefit analysis difficult, not least because many of the costs of the program are also not obvious (e.g. academic or telecommunications support). With regards to the benefits, most authors point to the clarity and focus of the educational programs as determinants of success. In evaluating the OLPC program in Peru, then, I am not simply evaluating the effect of XO computers in Peruvian schools but the characteristics of the particular program as well.

The trend is similar in non-English-speaking countries such as Spain, where the deployment of ICT centres are questioned (Cubillo and Rojo: 2009) and Portugal, where the presence of broadband in schools has detrimental effects on high school students’ test scores (Belo et al: 2008). Thus, there are at best mixed effects of computerized instruction on test scores, and the evidence from developed countries is not encouraging. However, this may not be the case in developing countries. As Banerjee et al. (2007) argue, computers may replace unmotivated and untrained teachers. They examine the effect of a computer-assisted learning program in India which offered children two hours of computer time per week to play games that involve math problems. They find a positive effect of computer use on math test scores, even though the effect faded over time. Nevertheless, their research suggests that closely targeted computer instruction may be beneficial.

Malamoud and Pop-Eleches (2010) evaluate a recent Romanian program which allocated vouchers for the purchase of a home computer to low-income students by estimating the computer’s effect on child and adolescent educational outcomes. They find that most of the computer time was spent playing games. They also find that children who (randomly) received a voucher achieved significantly lower school grades in maths, English, and Romanian, but scored significantly higher in computer skills and computer fluency. Furthermore, in a previous 2008 study of the same program, they find suggestive evidence that the having a computer at home was associated with negative behaviour outcomes. Importantly, they find that the presence of parental rules regarding computer use and homework appear to mitigate the effects of computer ownership, but only when it is constructive (encourages doing homework) rather than restrictive (regulating computer use).This suggests two things: the first is that parental monitoring and supervision may be important mediating factors. The second is that such monitoring probably applies to teachers in the classroom too, meaning that computers in the classroom may have positive effects on student educational outcome if teachers are properly trained.

No empirical study has been done on any OLPC program. Antje Breitkopf (2011) has conducted a qualitative analysis of 12 OLPC schools in Peru. She finds that there seems to be an adjustment period on behalf of both students and teachers after the introduction of the laptop, which is extensive given the relative lack of support from the Ministry of Education and the relative lack of initial technological know-how in the village.[10] Laptops were mostly used between 1 and 3 times a week and very rarely on a daily basis. In Grades 1 and 2 they were hardly used because teachers focused their efforts in teaching how to read, and more importantly, write.[11] There were also some significant language barrier problems, as many students spoke Aymara at home and had only begun to learn Spanish at school. Santiago et al. (2010) through the Inter-American Development Bank study find similar results, although the also find that laptop use decreases over time, though they are not able to explain why.[12] There is no doubt that the XO laptops have great potential, but the program is likely over-idealized and not country-specific enough to give children all it can give. (Yujuico: 2011)

 

Bibliography

  1. Aubrey, Carol, and Sarah Dahl, (2008) “A review of the evidence on the use of ICT in the Early Years Foundation Stage,” Early Childhood Research Unit Warwick University Institute of Education.
  2. Banerjee, Abhijit V, Shawn Cole, Esther Duflo, Leigh Linden, (2007) “Remedying education: evidence from two Randomized experiments in India.” Quarterly Journal of Economics, 122:3
  3. Belo, Rodrigo, Pedro Ferreira, Rahul Telang, (2008) “The Effects of Broadband in Schools: Evidence from Portugal,” JEL Working Paper.
  4. Breitkopf, Antje, (2011) “Informe sobre la investigacion de las laptop XO en el Peru.” Presented before Direccion General de Tecnologias Educativas del Ministerio de Educacion del Peru.
  5. Cubillo Fuentes, María Dolores, and Remedios Rojo, (2009) “Educar para innovar, innovar para educar: Los centros TIC en Andalucía,” Centro TIC.
  6. Derndorfer, Christoph, (2010) “OLPC in Peru: A Problematic Una Laptop Por Niño Program.” Educational Technology Debate. http://edutechdebate.org/olpc-in-south-america/olpc-in-peru-one-laptop-per-child-problems/
  7. Fairlie, Robert W, (2005) “The Effects of Home Computers on School Enrollment.” Economics of Education Review 24.
  8. Fairlie, Robert W, and Rebecca A. London, (2009) “The Effects of Home Computers on Educational Outcomes: Evidence from a Field Experiment with Community College Students.”
  9. Fairlie, Robert W, Daniel Beltran, and Kuntal Das, (2010) “Are Computers Good for Children: The Effects of Home Computers on Educational Outcomes,” Economic Inquiry.
  10. J. Palozzi, Vincent J. and Terry E. Spradlin, (2006) “Educational Technology in Indiana: Is it Worth the Investment?” Center for Evaluation and Education Policy, 4:4
  11. Macho, Steve, (2005) “Differences Among Standardized Test Scores Due to Factors of Internet Access at Home and Family Affluence.” College of Human Resources and Education at West Virginia University.
  12. Malamud, Ofer, and Cristian Pop-Eleches, (2008) “The Effect of Computer Use on Child Outcomes, Harris School of Public Policy Studies working paper 0812.
  13. Malamud, Ofer, and Cristian Pop-Eleches, (2010) “Home Computer Use and the Development of Human Capital,” NBER Working Paper no.15814.
  14. O’Dwyer, Laura M., Michael Russel, Damien Bebell, & Kevon R. Tucker-Seeley (2005), “Examining the Relationship Between Home and School Computer Use and Students’ English/Language Arts Test Scores,” The Journal of Technology, Learning, and Assessment, 3:3
  15. O’Dwyer, Laura M., Michael Russel, Damien Bebell, & Kevon R. Tucker-Seeley (2008), “Examining the Relationship Between Students’ Mathematics Test Scores and Computer Use at Home and at School,” The Journal of Technology, Learning, and Assessment, 6:5
  16. Santiago, Ana, Eugenio Severin, Julian Cristia, Pablo Ibarraran, Jennelle Thompson, Santiago Cueto, (2010) “Evaluación experimental del program ‘Una Laptop Por Niño’ en Peru.” Inter-American Development Bank: Education, 5.
  17. Schacter, John, (1999) “Education Technology on Student Achievement – what the most current research has to say.” Milken Family Foundation.
  18. Siann, G, H. Macleod, P. Glissov and A. Durndell, (1990) “The effect of computer use on gender differences in attitudes to computers,” Computers and Education 14:2.
  19. Sweeney, Trudy, Ruth Geer, (2010) “Student capabilities and attitudes towards ICT in the early years.” Australian Educational Computing.
  20. Yujuico, Emmanuel (2011) “Cautions from one laptop per child in marketing technological innovation to LDCs.” California management review.

Data Sources:

ESCALE Censo Escolar: http://escale.minedu.gob.pe/censo-escolar

Evaluación Censal de Estudiantes UMC: http://www2.minedu.gob.pe/umc/index2.php?v_codigo=52&v_plantilla=2

Distribución de Laptops OLPC: http://www.perueduca.edu.pe/olpc/OLPC_Dist.html

ESCALE interactive map: http://escale.minedu.gob.pe/mapaeducativo/

  1. From OLPC website: http://www.olpcnews.com/people/negroponte/new_negropontism_you_can_give_kids_xo_laptops.html ↑

  2. Anecdotal evidence from the OLPC website points to parents using the computers to connect to Google Maps in order to make their business transactions more efficient. ↑

  3. http://laptop.org/en/children/countries/peru.shtml. ↑

  4. See http://escale.minedu.gob.pe/magnitudes-portlet/reporte/cuadro?anio=2&cuadro=20&forma=U&dpto=&dre= ↑

  5. The pilot consisted of providing laptops to all 46 children in the single-classroom school in the village. ↑

  6. https://www.cia.gov/library/publications/the-world-factbook/geos/pe.html ↑

  7. See Urquiola, Miguel and Valentina Calderon (2006) “Apples and oranges: Educational enrollment

    and attainment across countries in Latin America and the Caribbean,” International

    Journal of Educational Development, Vol. 26, pp. 572{590 ↑

  8. i.e. children it may be the case that smarter children use a computer because they were smarter to begin with, not merely because of the potential facilitating effects of the computer. O’Dwyer et al do recognize their inability to randomize as a shortcoming of their paper. ↑

  9. E.g. it may be the case that children with higher levels of academic ability or children with more “educationally motivated” parents are more likely to have access to a home computer. ↑

  10. Teacher training was clearly lacking, as some teachers had problems just entering and exiting computer programs. ↑

  11. That is one skill that current laptops cannot teach. ↑

  12. The reason is probably that laptops get broken and take a long time to fix. For more on this see Derndorfler (2010) ↑