2007 Economic Census

Data Manipulation

Effect of "D'd" Records on Sorted Data

When using the sort tools in American FactFinder to rank data, records with 'D's fall to the bottom of the ranking.

Data users need to know:

• these D'd records could be one of the higher rankings and
• that they need to calculate the amount of the total that these D'd records account for (Total - published rows = residual) to determine if the D’d records could be in top.

Effect of Unpublished Data on Sorted Data

Data often have to meet minimum criteria to be published in the data sets.

Data users need to know that:

• this minimum criteria used is often unpublished,
• the sum of the data published is often less than the totals and
• that just because an industry or geography is not published, it does not mean that there is zero activity in that industry or geography.

Ranking "Small" Data Cells Along With "Larger" Data

Data users need to know that:

• these small cells of data often are not statistically significant enough to allow for comparison with larger, more significant industries,
• small changes in these small industries can often result in large (but statistically insignificant) swings in year-to-year and other ratios and
• with the limited resources the Census Bureau has, these smaller industries/geographies may not be as thoroughly reviewed as others.