Users may notice a large shift in the population counts for some cohorts between 2009 and 2010. This shift happens because our demographics data from 2000-2009 are based on the 2000-2009 Census Population Estimates dataset, which are modeled figures released by the Census Bureau based on the 2000 Census.
Since the 2010 Census was published, Population Estimates have been modeled off the 2010 Census. The dramatic shift from 2009 to 2010 happens because the modeled Population Estimates data from 2001 to 2009, based on the 2000 Census, is correcting to match the actual data in the 2010 Census. The mostly smooth trends from 2000-2009 are sharply broken when the new Census comes out, revealing that the Census projections didn’t accurately project the population.
The Census and Population Estimates are reported in terms of race and ethnicity cohorts. Race refers to categories such as Asian, Black/African American, White, etc. Ethnicity refers to whether an individual is Hispanic or non-Hispanic. Race and ethnicity are combined, so each race has a Hispanic and a non-Hispanic ethnicity category. There was confusion on the 2000 Census about reporting race and ethnicity, and the Census postulates that this may contribute to the incorrect projections made in the 2001-2009 Population Estimates data. The sharp spike between 2009 and 2010 data is most noticeable in the non-White, Hispanic Race/Ethnicity categories.
Here is an example of this trend for several US counties:
We publish this data, shift included, to match published data from the Census. Our demographic projection methodology does not incorporate data prior to 2010 in calculating future trends, so Emsi Burning Glass’s modeled demographics data is untouched by this anomaly.
A similar situation may occur when the 2020 Census data becomes a part of our demographics dataset. In the event that the 2019 Population Estimates deviate noticeably from the 2020 Census data, we may incorporate a smoothing algorithm into the projection methodology to approximate more accurate values for years from 2010-2019 that match the 2010-2020 censal trend more closely. For the moment, however, the Census Population Estimates are still the best data source around for historical demographics data.
Let us know what specific questions we can help you with (we may even add your question to our knowledge base).
Let us know what specific questions we can help you with (we may even add your question to our knowledge base).