Purpose/Hypothesis: Previous studies provide mixed support for the idea that older adults with lower balance confidence necessarily restrict their activity. Importantly, studies of confidence and activity neither have directly recorded free-living ambulatory activity nor have considered the potential influence of gait speed. We sought to address these limitations by examining relationships among balance confidence, monitored ambulatory activity, and gait speed in a cross-sectional sample of seniors living in the community. Subjects: 65 adults (78.6 +/− 6.2 years old, 58% female, 85% white, non-Hispanic, BMI = 26.6 +/− 5.0). 24 reported difficulty in either climbing 10 stairs or walking 0.5 miles. Materials/Methods: Balance confidence was measured using the Activities-specific Balance Confidence (ABC) Scale. Ambulatory activity data were collected continuously for 2 weeks using a StepWatch3 Activity Monitor (Cyma Corporation, Mountlake Terrace, WA). Ambulatory activity was quantified in terms of average daily steps, bouts, and minutes. Gait speed was measured at usual and fast pace over a 2.4 m walkway. Controlling for age, sex, race, mobility difficulty, and height, we used canonical correlations (alpha = 0.05) to examine relationships between the domains of (1) balance confidence and ambulatory activity, (2) gait speed and ambulatory activity, (3) balance confidence and gait speed, and (4) balance confidence plus gait speed and ambulatory activity. Results: Subjects were relatively confident in their balance (ABC = 82.3 +/− 15.5), had mildly diminished gait speed (usual pace = 1.05 +/− 0.27 m/s; fast pace = 1.51 +/− 0.38 m/s), and ranged from sedentary to very active (average daily totals: 8713 +/− 3594 steps, 65.3 +/− 15.1 bouts, and 313.1+/− 94.9 minutes). There was no significant relationship between either the domains of balance confidence and ambulatory activity [canonical correlation = 0.22, p = 0.43] or gait speed and ambulatory activity [canonical correlation = 0.36, p = 0.08]. Balance confidence was significantly associated with gait speed [canonical correlation = 0.50, p = 0.0003], with fast pace speed accounting for more of the association. The combination of balance confidence and gait speed produced a significant correlation with ambulatory activity [canonical correlation = 0.44, p = 0.05], with lower confidence and higher gait speed tending to be associated with more steps and minutes of activity. Conclusions: Our data help to resolve previous uncertainty in the literature by reinforcing the idea that balance confidence, although positively associated with gait speed, indicates relatively little about actual ambulatory activity in community-dwelling seniors. Gait speed appears to modify the relationship between balance confidence and activity. Clinical Relevance: Physical therapists should be cautious about interpreting the ABC scores of community-dwelling seniors to make inferences about ambulatory activity. More confident seniors are not necessarily more active. Less confident seniors do not necessarily restrict their activity.