When you checked the boxes on your census form, you probably did not pay much attention to the form design. But it is something the Census Bureau researches and gives a lot of attention. Something as simple as using vertical instead of horizontal boxes can have a big impact on how you read the form and the data we gather.
If one person’s eye misreads the form and checks the wrong box, it may not change percentages of statistics very much. But if even 1 percent of a large population checks the wrong box on a question, it could lead to an inaccurate picture of a smaller population.
For example, you’ll see in the video below that if you have 1,000 circles and 100,000 squares and just one-percent of the squares are counted as circles, then that’s one thousand squares now listed as circles. This mismarking would result in responses that double the amount of circles.
A similar data capture error affected the 2010 Census results for same-sex couple households. On the form used by census takers, the boxes for sex were placed vertically, and the wrong gender box was marked for a small percentage of opposite-sex partners, artificially increasing the percentage of same-sex couple households.
When we discovered this inconsistency in the responses, we developed a better set of estimates to provide a more accurate measure of same-sex married and unmarried partner households. These estimates are now available.
At the Census Bureau, we strive to eliminate these kinds of errors and modify our survey questionnaires to gather the best possible data on the nation’s people, places and economy.
This brief video illustrates how a small error in a large population can create a large error in a smaller related population, such as the same-sex couple population.
For more information, visit our online press kit.