This study was designed to determine the number of days of pedometer monitoring necessary to achieve reliable and valid estimates of a 1-year average of step counts in adults based on either consecutive days (CD) or random days (RD) of data collection. Twenty-three participants (16 women; M age = 38 years, SD = 9.9) wore a Yamax SW 200 pedometer and recorded their step counts for 365 consecutive days. Nine measurement periods of various lengths were selected (2, 3, 4, 5, 6, 7, 14, 21, and 30 days). Each measurement period was randomly selected 10 times each for CD and RD from the larger database. For reliability and validity, two-way intraclass correlation coefficients and mean absolute percentage error (MAPE) were calculated, respectively. The year-round average was considered the criterion measure of the "true" habitual physical activity. Data were analyzed separately by CD and RD. At least 5 CD or 6 RD were necessary to achieve an ICC of .80. A minimum of 30 CD or 14 RD were necessary to achieve an MAPE lower than 10%. These findings provide researchers and practitioners with useful information to determine appropriate measurement length and the method of data selection based on a desired level of reliability and validity.