Bloomberg Businessweek has called One Laptop Per Child (OLPC) a flat-out failure. PCMag published a slightly scathing column titled. “One Laptop Per Child Doesn’t Change The World.” Now, the Inter-American Development Bank (IDB) has released a study claiming the program hasn’t at all improved math or language test scores. To top it off, the program hasn’t affected time allocated to doing homework.

The study is based off data the IDB collected after 15 months of implementation in 319 primary schools in Peru. Implementation might just be the problem, however, given that teachers said OLPC hasn’t provided them with enough guidance on how to effectively integrate the computers and tablets into their courses.

Two positive notes: The program has increased the ratio of computers per student from 0.12 to 1.18. Because of that, children are now more comfortable with basic computing skills and tasks, such as word processing.

OLPC’s mission has always been to “provide every child in the world access to new channels of learning, sharing and self-expression.” Those positive notes, then, appear to support their goals, contrary to what the negative numbers say.

In an interview earlier this year, OLPC CTO Ed McNierney said, “Why have 43 countries around the world bought two and a half million of these laptops? There’s a reason. There’s a reason they didn’t buy something else.”

Will people keep buying their laptops and tablets if test scores aren’t improving, though? Will implementation need to be taught? In January, with the support of the Knight Foundation, OLPC donated 500+ students at Holmes Elementary School in Miami laptops, along with in-house training at the school for parents, teachers and students on how to use the computers to advance students’ learning. If the training can help improve test scores, could that training then be brought outside of the United States to countries like Peru?

Although the numbers are disheartening, if properly integrated into a school’s curriculum, OLPC could see a rise. But, what do you think? Do you think OLPC is making enough of a difference?