Measuring underreporting and under-ascertainment in infectious disease datasets: a comparison of methods