exactly 50 days till the papers for my qualifying exams are due, and i've already started having dreams of excel. not a good sign. have reached a rather tricky point in the meta-analysis where i'm trying to model the residual variance unaccounted for by sampling and random error. i should have thought this through a little more carefully at the start -- by allowing hours of sleep deprivation to vary between studies, i now have to build a sinusoidal meta-regression with an exponentially saturating component*, which i'm not sure is even technically possible. sigh. for a while in the 90s, math actually made sense to me. those were the days.

* this is not gibberish, i swear: see borbely, 1982 in human neurobiology