Errors in estimates of women employees and production and nonsupervisory employees

The Current Employment Statistics (CES) program discovered processing errors in estimates for women employees (WE) and production and nonsupervisory employees (PE) series in five smaller industries within financial activities, professional and business services, and leisure and hospitality, noted below. The timeframe of the errant data ranges from November 2012 until May 2017; however, not all data during this timeframe are errant. In some months, errors in WE impacted higher aggregate WE employment estimates up through total nonfarm, and errors in PE impacted higher aggregate PE employment estimates up through total private. The errors in PE also resulted in minor errors in hours and earnings for aggregated PE, because hours and earnings are weighted by PE employment in the aggregation process. All industries and data types impacted by the errors are listed in Tables 1 and 2, respectively.

Errors were relatively small in nature and tended to offset from month to month and, therefore, did not change the underlying trends over the period.  At the highest level, the following table shows the maximum absolute error in any one month for each basic data type.

Corrected data were published February 2, 2018, in the CES public database.