Accuracy of traditional legendre estimators of quadrupole ratios for the N → Δ transition

We evaluate the accuracy of traditional estimators often used to extract N{yields}{delta} quadrupole ratios from cross-section angular distributions for pion electroproduction. We find that neither M{sub 1+} dominance nor l{<=}1 truncation is sufficiently accurate for this purpose. Truncation errors are especially large for R{sub EM}, for which it is also essential to perform Rosenbluth separation. The accuracy of similar truncated Legendre analyses for E{sub 0+},S{sub 0+}, and especially M{sub 1-} is even worse.