A colleague of mine did a neighborhood survey of a set of US Census tracts. One of the variables he asked is household income. He wants to know if the median income in his sample is significantly different than the median household income as reported by the US Census. How can he pool the estimated median household income and its standard error? He would have the median for each tracts, the total household income for each tract, the number of households in each tract, the margin of error for each tracts, and perhaps some of the different percentiles in the distribution.