The interpretation of the observed gamma-ray burst V/V(max) statistic in terms of spatial distributions is model-dependent. Detection of gamma-ray bursts requires the counting rate in one or more detectors to exceed aa threshold C(lim) determined from a time-dependent background rate B(t). The sampling depth of the burst detector is thus time-dependent, and, if burst sources are non-uniform in space, the observed distribution will be affected by B(t). We demonstrate this effect with aa simple geometric distribution of standard candles and argue that V/V(max) statistic without information on threshold variations is insufficient for rigorous data analysis. Peak count rates and threshold values must be given separately for all events in order to facilitate aa meaningful comparison of observations with theoretical distribution models.