Omitted variable bias: A threat to estimating causal relationships