The present study further explored the age-related changes in attentional selection, first reported by Jackson and Balota Jackson and Balota (Psychology and Aging 28:3, 2013). To this end, response stimulus interval (RSI) was manipulated in in a Stroop paradigm that allows to isolate different components of the overall Stroop effect. The originally reported reduction of the overall Stroop effect in short (vs. long) RSI was successfully replicated. In line with the original claim that the ability to maintain the relevant task set consistently across time is preserved in older adults, this reduction failed to interact again with age – a null result that was uniquely supported in the present study with Bayesian evidence. Although healthy aging affected all components of the overall Stroop effect, the present study further identified interference and a disproportionally larger semantic conflict effect, as being the primary mechanisms behind the larger overall Stroop effect that older adults produce compared to younger adults. These results therefore directly fuel the original conclusion of Jackson and Balota Jackson and Balota (Psychology and Aging 28:3, 2013) that age-related differences in attentional selection are due the degradation of task set quality, namely a lesser ability of older adults to control interference. The present study additionally showed this reduced control of interference directly and identified where in the processing stream it occurs. The immediate implication that challenges the dominant unitary models of Stroop performance, is that future interventions designed to boost interference control in older adults should most likely target semantic, as opposed to response conflict.