Showing posts with label Census. Show all posts
Showing posts with label Census. Show all posts

30 September 2011

Is counting the census more difficult that it seems?

We're often told - quite rightly - that if one plays and jumbles with numbers enough they can be manipulated to say whatever one wants.

This is to be expected from political parties and retailers as they seek to garner more attention and loyalty, but it isn't expected from the United States Census Bureau.

Even so, this is precisely what has been done to the figures from the 2010 census in regards to the number of same-sex households in the nation.  The National Catholic Register has the story, with my emphases:

Original data published by the 2010 Census reported that the number of same-sex households in the U.S. was 901,997, including 349,377 same-sex “married couple households,” and 552,620 same-sex unmarried partner households.

On Tuesday, however, the Census Bureau admitted that those numbers were significantly revised downward after staff discovered inconsistency in file statistics “that artificially inflated the number of same-sex couples.”

The number initially reported was about 40% higher than what the Census Bureau believes to be accurate.

The Census Bureau now estimates that there are 131,729 same-sex “married couple households,” and 514,735 same-sex unmarried partner households in the U.S. for a total of 646,464 same-sex households. The Census Bureau had, therefore, misrepresented the number of same-sex couples by 255,533 or 39.5% [more, with good analysis].

That's no small miscalculation.  It could just be a complex [such a miscalculation could hardly be called simple] and honest mistake, but it seems more likely to have been done for political purposes yet to be uncovered.

Now, one might wonder, given the data collected by the census bureau it shouldn't be terribly difficult to count up the number of same-sex households; why give an estimate and not an actual figure?