The following statistical model was used: : where, [Y.sub.ijl] indicates the set of assessed traits; [mu] is the general mean; [b.sub.i]: the block effect;
popj: the effect associated with the j-th population level; [e.sub.ij]: effect of the error between (experimental error); [d.sub.ijl]: effect of the error within (sampling error), which are data obtained by evaluation of the plants of the experimental units.
Table 5: Mayotte imports (2000-2010) Model without wharf occupancy and with corrected population (1) (2) Intercept -8.75 -1.17 (4.87) (2.70) GDPi 0.32 * 0.56 *** (0.15) (0.10) GDPj 0.85 *** 0.85 *** (0.09) (0.09) POPi 1.04 * (0.48) POPi Corrected 0.29 (0.24)
POPj -0.28 *** -0.28 *** (0.08) (0.08) dist -0.41 *** -0.41 *** (0.04) (0.04) Colony 6.29 *** 6.29 *** (0.51) (0.51) Demand for asylum (DA) Expulsions Wharf occupancy (mar) Wharf occupancy/Gross tonnage imported EFj Yes Yes EFk Yes Yes EFt No No Obs.
where S is the aggregate grant to all provinces for delivery of some service i, i is the type of service, j is the eligible subset of residents, and
popj represents the proportion of a province's residents in group j.