The authors tested the validity of the LATCH breastfeeding assessment tool, controlling for intervening variables in 133 dyads. LATCH scores, mother's evaluation of an index feed, and intended duration of breastfeeding were assessed postpartum and followed 6 weeks. Women breastfeeding at 6 weeks postpartum had higher LATCH scores (mean +/- SD = 9.3 +/- 0.9) than those who weaned (mean +/- SD = 8.7 +/- 1.0), due to only one measure, breast/nipple comfort. Women who weaned before 6 weeks reported lower breast/nipple comfort (1.5 +/- 0.5) than those who were still breastfeeding at 6 weeks (1.7 +/- 0.5, P < .05). Total LATCH scores accounted for 7.3% of variance in breastfeeding duration. Total LATCH scores positively correlated with duration of breastfeeding (n = 128; r = .26, P = .003) and to mothers' scores (n = 132; r = .58, P = .001). Correlations among LATCH measures ranged from .02 to .51. The LATCH tool is a useful identifies the need for follow-up with breastfeeding mothers at risk for early weaning because of sore nipples.