Using data from the GEOS 2 electron gun experiment, we have analyzed the dependence of the dayside convection electric field at L=6.6 (averaged into 3-hour LT bins) on solar wind and interplanetary magnetic field conditions. The convection electric field does not correlate at all with the solar wind momentum flux density (correlation coefficients of <0.25). Hence viscous interaction plays only a minor role for equatorial magnetospheric convection at L=6.6. The correlation coefficients for convection electric field versus merging electric field are of the order of 0.5-0.6, thus indicating that dayside convection is predominantly driven by dayside merging. The regression coefficients, describing the (LT-dependent) transfer of the merging electric field to synchronous orbit, are of the order of 0.1-0.2. Their LT dependence follows approximately that described by a Volland-Stern model with &ggr;=2, except that the regression analysis yields a dawn-dusk asymmetry with the dusk sector electric fields being about twice as strong as those in the dawn sector. For the above shielding factor (&ggr;=2) the regression coefficients are consistent with a merging efficiency of about 20%. The regression constants (of the order of 0.05 mV/m) apparently describe an electric field stemming from the ionospheric wind dynamo. A comparison of the electric field described by the regression constants and two empirical models of quite day electric fields implies that these models still include a substantial dawn-to-dusk electric field, thus indicating the presence of solar wind dynamo action even during quite intervals.