When I worked at the Independent Evaluation Group of the World Bank, I evaluated the outcomes of completed Bank-financed projects in an FTI partner country in Francophone Africa. Around 2004 the country could be considered the testing capital of Africa. Students had been assessed through a French-financed achievement test, the UNESCO Minimum Learning Achievement tests, and PASEC. All tests showed low and potentially deteriorating results. The reasons were quite clear: Students were supposed to learn literacy in French, a language they did not know and that has complex spelling rules. The textbooks financed by the projects had been delayed, and many were stolen from the few schools that eventually received them. Students spent the class time copying the missing textbooks. Teachers themselves had limited command of the basic skills they were supposed to teach or were absent much of the time.

Many carried out the time-honored practice of interacting with the 3-4 students who could perform and neglecting the rest. The result was evident without a probability sampling design. Most students were illiterate all the way to grade 6. So how useful was the feedback from the tests in helping schools improve student performance? Some officials would consider the question a joke.

Fast forward to 2011. The World Bank has released widely the film‘When Students Learn, Nations Prosper“. It is an inspirational film aimed at demonstrating the power of measurement in improving student outcomes in Singapore, Brazil, Colombia, and South Africa. Failing students in Singapore received vocational training. And South African teachers trained in assessment declared themselves better able to teach their students. “We need our students to become critical thinkers” said one of them. But what exactly did they learn to do in the first five minutes or the next 10 minutes of a class? The film tantalizingly asked the question. But the answer was poetic words about prosperity – and beautiful pictures of cheerful, some dancing kids.

Testing may not help improve many low income countries’ practices, but it has greatly improved the bottom line of the consultant industry.

Could it be that when nations measure, testing specialists prosper?

Individuals and firms, often hired through designated donor funds, conduct sophisticated assessments in the furthest corners of the earth. International comparison tests, in particular are expensive and complex exercises that may use up the time of the best-qualified people in Ministries of Education. By contrast, instructional methodology gets no billing. “Pedagogy” is often viewed as a detail which schools will somehow figure out if they get paid on the basis of performance.

The evidence for such motivation theories is dubious, but the evidence on memory mechanisms is quite reassuring. People process information according to rather well known “rules”. For example, students must read and calculate fluently early on in order to free up working memory for more complex problems. The class activities to automatize these skills are fairly clear and should take priority over everything else. And classroom time spent on linking information – presenting it, analyzing, synthesizing, comparing will help create complex networks of knowledge. If schools focus on these activities and do plenty of them, kids will do great on tests.

So does the World Bank’s ambitious“Learning for All” strategy boil down to “testing for all”? Certainly not, but some critics may think so. One difficulty is that learning research is relatively unknown in the education sector. By contrast, everyone in high positions has passed tests. We tend to remember fondly our parents getting feedback, our children’s well-educated teachers meeting to pore over standardized test results. Testing seems like a natural strategy to promote.

And should FTI partner countries test or not? Most certainly, but the answer may lie in the proverbial 80-20% tradeoff. Smaller, quicker, cheaper tests can give valid and reliable feedback on basic skills acquisition. These are one-minute reading fluency tests (just 1 the 8-9 EGRA subtests), 3-minute math and writing tests. They measure fluency in the building blocks and therefore readiness for more complex work. The poor tend to fail in that early stage, and yet the methods to teach them are fairly well understood. When students spend their class time creating complex networks of knowledge, nations are sure to prosper.