There is increasing evidence that top-down controls have strong non-consumptive effects on herbivore populations. However, little is known about how these non-consumptive effects relate to bottom-up influences. Using a series of field trials, we tested how changes in top-down and bottom-up controls at the within-plant scale interact to increase herbivore suppression. In the first experiment, we manipulated access of natural populations of predators (primarily lady beetles) to controlled numbers of A. glycines on upper (i.e. vigorous-growing) versus lower (i.e. slow-growing) soybean nodes and under contrasting plant ages. In a second experiment, we measured aphid dispersion in response to predation. Bottom-up and top-down controls had additive effects on A. glycines population growth. Plant age and within-plant quality had significant bottom-up effects on aphid size and population growth. However, top-down control was the dominant force suppressing aphid population growth, and completely counteracted bottom-up effects at the plant and within-plant scales. The intensity of predation was higher on upper than lower soybean nodes, and resulted in a non-consumptive reduction in aphid population growth because most of the surviving aphids were located on lower plant nodes, where rates of increase were reduced. No effects of predation on aphid dispersal among plants were detected, suggesting an absence of predator avoidance behavior by A. glycines. Our results revealed significant non-consumptive predator impacts on aphids due to the asymmetric intensity of predation at the within-plant scale, suggesting that low numbers of predators are highly effective at suppressing aphid populations.