Time-of-survey dependence of apparent educational participation

Bilal Barakat, Wittgenstein Centre (IIASA, VID/ÖAW, WU)

While seemingly straightforward, many common indicators of educational participation, such as net enrolment ratios, are subject to distortions that are transparent from within a demographic framework, but generally neglected in the field of educational development. With a focus on the fact that the true school-age population is misidentified when age is measured in whole years in household surveys conducted some time after the beginning of the school year, and on the interactions of this effect with the cut-off date, population growth, and drop-out, it can be shown both analytically and through simulations that the errors induced in customary participation indicators - and potentially also in statistical analyses of the determinants of education - can be considerable. At the same time, approximate corrections are possible even with the data currently available, that is, when age is measured in whole years only. In addition, the demographic perspective adopted suggests an alternative interpretation of the out-of-school rate in terms of person-time rather than headcount.