research on computer use and child outcomes

One topic that we've been discussing at length recently is the wisdom of efforts such as One Laptop Per Child, which aim to greatly expand home computer use among low-income children, especially in the third world. It has been taken as fact that computer access is beneficial to children, but is this really so? Are computers an important priority, or are they a distraction from more pressing needs?

A new study offers an interesting perspective on these questions. Economists Ofer Malamud and Cristian Pop-Eleches studied a Romanian program that offered 200 Euro vouchers to tens of thousands of families with children. The circumstances of the program offered the researchers a unique opportunity to compare demographically similar families where the only variable was whether or not the family had received a voucher. The results showed that the program was a tremendous success at improving access to computer technology, but having a home computer actually had a negative result on the child's academic situation - time spent doing homework declined, as did GPAs. Children with computers were 13% less likely to want to go to college and, surprisingly, were no more likely to want to pursue higher education in computer science.

There are a few important caveats. First, and perhaps most importantly, parental involvement greatly reduced the negative effects of winning a computer voucher. This was especially true of families with a stay-at-home mother. Secondly, although the Romanian Ministry of Education provided free educational software, only 30% of families reported having such software installed on their computer. Only 20% of children reported using educational software more than once a week, versus 70% for games. Any new large-scale efforts to provide computers to low-income families need to focus on educational uses - something OLPC is doing, to their credit - to avoid the negative outcomes reported in this study.

Third, it is clear from this study that demand for home computers was very high with this low-income Romanian population. Over 90% of families used their vouchers towards a computer purchase - even though the voucher only covered about 75% of the cost of a computer. By my calculation that would mean that families spent about $80. If we consider that families needed to make less than $16 a month per household member to receive a voucher, than families were likely spending more than a month's income on their computer. If the demand for computers among low-income families in the United States even approximates that in Romania, than digital inclusion efforts shouldn't be afraid to charge modest sums for computers.

Finally, it is important to note that only 6% of those surveyed here had Internet access at home. Fortunately, Internet access is more widely available in American cities than it is in rural Romania, but if a computer with no Internet access serves primarily as a distraction for schoolchildren, will the potential benefits of Internet access also be outweighed by the potential to distract?

Of course, this study only measures the impact of technology access in one area - school performance. It may be the case that computer access is a net benefit for adults - offering access to new resources like distance learning, for example, at the expense of time that would otherwise be spent watching television. Additionally, having a home computer in the United States, where computer knowledge is expected and the Internet is a major part of culture, may be relatively more beneficial than owning a computer in a third world country.

Having said all that, this study should give technologists the opportunity to examine carefully whether we really know as much as we think we do about problems like third-world poverty and education. We ought to consider whether computer technology and the ideas computer scientists embrace (open source or constructivist educational theory for example) may not be appropriate in all circumstances.

Source: http://www.columbia.edu/~cp2124/papers/computer.pdf