Infants from low-socioeconomic status (SES) households hear a projected 30 million fewer words than their higher-SES peers. In a recent study, Hirsh-Pasek et al. (Psychological Science, 2015; 26: 1071) found that in a low-income sample, fluency and connectedness in exchanges between caregivers and toddlers predicted child language a year later over and above quantity of talk (Hirsh-Pasek et al., "Psychological Science," 2015; 26: 1071). Here, we expand upon this study by examining "fluency and connectedness" in two higher-SES samples. Using data from the NICHD Study of Early Child Care and Youth Development, we sampled 20 toddlers who had low, average, and high language outcomes at 36 months from each of 2 groups based on income-to-needs ratio (INR; middle and high) and applied new coding to the mother-toddler interaction at 24 months. In the high-INR group, the quality of mother-toddler interaction at 24 months accounted for more variability in language outcomes a year later than did quantity of talk, quality of talk, or sensitive parenting. These results could not be accounted for by child language ability at 24 months. These effects were not found in the middle-INR sample. Our findings suggest that when the quality of interaction, "fluency and connectedness," predicts language outcomes, it is a robust relation, but it may not be universal.