We investigate the drivers of 40-150 keV hourly electron flux at geostationary orbit (GOES 13) using ARMAX (autoregressive moving average transfer function) models which remove the confounding effect of diurnal cyclicity and allow assessment of each parameter independently of others. By taking logs of flux and predictor variables, we create nonlinear models. While many factors show high correlation with flux (substorms, ULF waves, solar wind velocity (V), pressure (P), number density (N) and electric field (Ey), IMF Bz, Kp, and SymH), the ARMAX model identifies substorms as the dominant influence at 40-75 keV and over 20-12 MLT, with little difference seen between disturbed and quiet periods. Also over 40-75 keV, Ey has a modest effect: positive over 20-12 MLT but negative over 13-19 MLT. Pressure shows some negative influence at 150 keV. Hourly ULF waves, Kp, and SymH show little influence when other variables are included. Using path analysis, we calculate the total sum of influence, both directly and indirectly through the driving of intermediate parameters. Pressure shows a summed direct and indirect influence nearly half that of the direct substorm effect, peaking at 40 keV. N, V, and Bz, as indirect drivers, are equally influential. Neither simple correlation nor neural networks can effectively identify drivers. Instead, consideration of actual physical influences, removing cycles that artificially inflate correlations, and controlling the effects of other parameters using multiple regression (specifically, ARMAX) gives a clearer picture of which parameters are most influential in this system.