Optimizing Testing

I have lately written a series of posts on a proposed American Baccalaureate Certificate that would grant public funding for three years of higher education, combined with automatic admission to our public universities. This would be earned by satisfactory performance on a series of assessments in the last two years of secondary school, with a natural focus on the end of the last year. There would be a lot of testing here, as there is in many countries with good secondary schools, especially in Europe; but how much testing is too much testing?

As always I stress comparative methods, and have lately been investigating systems' attitudes towards testing. At one extreme I find Shanghai High School,one of the oldest high schools in China's largest city, where pupils can count on 64 important assessments per year (eight per subject, or about one a month, in the eight subjects they study). At the other extreme I find, among public school systems, those of the Nordic countries, where pupils traditionally have not been tested, nor received any grades in any subjects, before the eighth grade; and, beyond them in terms of freedom from public examination accountability systems, private schools in southern Africa, Canada, the United States, and the champions, in my view, Swiss private schools for (rich) foreign boarders, whose family money may well obviate a need for either university or vocational education, and who therefore have the temptation to extensive play before possibly settling down to a sinecure within the family business (which may well be in ruling a non-democratic country).

Significant questions arise, about which I have become increasingly curious: what is the optimal (not maximal or minimal) amount of testing pupils should experience? And what are we losing as we, like many other countries under the spell of OECD economists (these are the devisers of the PISA assessments comparing national educational achievement in ways that have politicians in almost all democratic countries antsy as they contemplate accounting to their electorates), move towards ever more testing? These are issues I expect to be exploring in the coming days, and about which (especially the issue of testing time optimization) I would love to discover useful research.