Sampling biases and implications for child abuse research.