Insect predators in North America suppress Aphis glycines Matsumura (Hemiptera: Aphididae) populations; however, insecticides are required when populations reach economically damaging levels. Currently, insecticides used to manage A. glycines are broad-spectrum (pyrethroids and organophosphates), and probably reduce beneficial insect abundance in soybean, Glycine max (L.) Merr. Our goal was to determine whether insecticides considered reduced-risk by the Environmental Protection Agency could protect soybean yield from A. glycines herbivory while having a limited impact on the aphid's natural enemies. We compared three insecticides (imidacloprid, thiamethoxam, and pymetrozine,) to a broad-spectrum insecticide (λ-cyhalothrin) and an untreated control using two application methods. We applied neonicotinoid insecticides to seeds (imidacloprid and thiamethoxam) as well as foliage (imidacloprid); pymetrozine and λ-cyhalothrin were applied only to foliage. Foliage-applied insecticides had lower A. glycines populations and higher yields than the seed-applied insecticides. Among foliage-applied insecticides, pymetrozine and imidacloprid had an intermediate level of A. glycines population and yield protection compared with λ-cyhalothrin and the untreated control. We monitored natural enemies with yellow sticky cards, sweep-nets, and direct observation. Before foliar insecticides were applied (i.e., before aphid populations developed) seed treatments had no observable effect on the abundance of natural enemies. After foliar insecticides were applied, differences in natural enemy abundance were observed when sampled with sweep-nets and direct observation but not with yellow sticky cards. Based on the first two sampling methods, pymetrozine and the foliage-applied imidacloprid had intermediate abundances of natural enemies compared with the untreated control and λ-cyhalothrin.