Update on Response Rates Show Very High Risk of Young Child Undercount Census Tracts Still Trail National Rates

Dr. Bill O'Hare

We have now had almost two months when households could self-respond to the 2020 Census and it is worth noting that the self-response rates in census tracts where a very high risk of young child undercounts are predicted trail overall response rates by a substantial margin.

Table 1 shows the average response rates in very high risk of young child undercount census tracts in  the 34 counties with the largest number of young children living in such census tracts. The data represent self-response rates through May 7, 2020. These 34 counties account for slightly more than half of all young children living in such tracts based on the Population Reference Bureau database of census tracts in 689 large counties. Based on self-response rates through May 7 2020, on average, across all 34 counties, the response rates in these very high risk of young child undercount tracts is 51.1 percent compared to a national response rate on May 7th of 57.7 percent; a difference of 6.6 percentage points.

Of the 34 counties shown in Table 1, seven counties had self-response rates for very high risk of young child undercount census tracts above the national self-response rates on May 7th.  Six of these counties are in California and one in Maryland. On the other hand, there were 10 of the 34 counties where the self-response rates in these targeted tracts are 10 percentage points or more below the national rate.  The ten are concentrated in the Northeast region of the county.   The very low response rate for Hidalgo County, Texas, is probably due in part to the fact that many households in that county are supposed to be counted in the update leave operation which has been postponed or delayed.

It is also illuminating to look at changes in self-response rates in these 34 counties since April 7th.   On April 7th, the average rates for the very high risk of young child undercount census tracts was 39.4 percent compared to national rate of 45.7 percent   In other words, the 6.3 percentage point gap on April 7, was 6.6 percentage points on May 7th – almost no change. Because the rate at which households are self-responding is tapering off, the gaps evident on May 7th, suggests a trajectory for census tracts where young children are most likely to be missed that is not promising.

Table 2 shows the counties ranked by change in response rates between April 7 and May 7th.  There is a lot of variation across counties. The counties with the biggest increase was Orange County, California, which increased 16.1 percentage points while Shelby County, Tennessee, only increased by 6.3 percentage points.  Many of counties that increased the most between April 7th and May 7th are in California which is consistent with the high rates shown in Table 1.

Data from the 2010 Census show young children had net undercount of 4.6 percent compared to an overcount 0.7 percent for adults.  Young children had a far higher net undercount rate than any other age group and young black and young Hispanic children were missed at a much higher rate.  The data provided in these tables indicate the net undercount of young children is likely to be high in 2020 without focused and robust intervention in the next few months.  The data provided here indicate exactly where the Census Bureau and census advocates should focus their outreach efforts over the next few months to get a better count of young children in the 2020 Census.