According to traditional conceptualizations of the relationship between fear of falling and falls in older adults, fear of falling is considered to be predictive of falls because it leads to activity avoidance which, in turn, leads to de-conditioning that increases fall risk. The recent literature has begun to challenge such conceptualizations. Specifically, it has been argued that fear of falling and anxiety, in and of themselves, have a direct negative effect on balance. In this study we manipulated anxiety level by asking older research participants to walk either on the floor (low anxiety condition) or an elevated platform (high anxiety condition). Half the time participants carried a tray (dual tasking) and half the time they did not. Manipulation checks (involving heart rate, galvanic skin response, and self-reported anxiety measurement) confirmed that the experimental manipulation was successful in affecting anxiety level. The results demonstrate that the experimental manipulation (platform vs. floor) affected balance parameters and dual tasking performance with the platform condition resulting in a less stable gait. In addition, increased task demand (i.e., dual tasking) also had a negative effect on balance performance. Finally, the results demonstrate that the paper and pencil measures of fear can also predict balance performance (although the variance accounted for is small) even after controlling for medical risk factors for falling. Implications for models of fear of falling are discussed.