Abstract Social indices were developed to assess breeding productivity of waterfowl based on weekly roadside surveys of social groupings (i.e., pairs, lone M, flocked M). We calculated social indices for mallard (Anas platyrhynchos) populations breeding on 16 study sites in the Canadian parklands from 1993 to 1998 using 7 previously developed indices. We also calculated duckling:pair ratios from our roadside counts, and we obtained independent measures of nesting effort, nesting success, female success, and fledging rate for these same 16 sites from a concurrent telemetry study. Social indices were correlated (r2 = 0.28-0.67) with telemetry-based measures of breeding productivity in 5 of 7 cases, with the strongest relationships deriving from indices that emphasized renesting effort. The 2 ineffective social indices (r2 ≤ 0.13) both measured early onset of nesting activity. Duckling:pair ratios could be calculated more easily from the same survey data and also were correlated (r2 = 0.26-0.48) with measures of breeding productivity. Because surveys measuring late-nesting effort also can enumerate early hatched ducklings, we recommend that waterfowl researchers use duckling:pair ratios rather than social indices because ducking:pair ratios are more easily interpretable. Development of sightability-adjustment factors for pair and duckling surveys could further enhance the utility of duckling:pair ratios as indices of breeding productivity in mallards.