Laboratory experiments are widely used to study how populations in nature might respond to various biological interactions, but the relevance of experiments in artificial venues is not known. We compiled mortality and growth data from 424 anuran populations carried out under laboratory, mesocosm, field enclosure, and field settings to determine if major differences exist amongst experimental venues and how this might influence experimental responses of tadpoles amongst venues. Our results show that there are fundamental differences in survival amongst venues, with the highest mortality occurring in field populations and the lowest in laboratory populations. Separation of mesocosm and field enclosure data based on the possibility of predatory interactions indicates that predation is an important factor leading to increased mortality in natural populations. Comparisons of size distributions across venues (although size data were limited for field populations) suggest that variation in tadpole size is low in natural populations compared to populations in artificial venues. We infer from this that mortality has a homogenizing effect on size in nature, resulting in natural populations that are not a random sample of hatched individuals. This finding suggests that populations reared under controlled laboratory conditions in the absence of predation (and other selective pressures) may not be representative of natural populations.