Tom & Mark,

The following pages illustrate my suggested staffing pattern masking methodology as you requested. The procedures outlined are an attempt to balance issues related to the level of detail presented (the customers always want more specific data) and those of the validity and confidentiality of releasing data with relatively small cell numbers.

Table 1 and Appendix 1 of "Introduction to Wyoming’s Staffing Patterns and Wages 2003" shows the staffing pattern with the masking procedures discussed in this memo already in place. The industry "334000 – Computer and Electronic Product Manufacturing" was chosen as it has a large number of occupations that are non-disclosable. The first thing to note from Table 1 is that WY Emp (Wyoming Employment) equals zero for most occupations. We will release an adjusted percent (WY Percent) of employment for the occupation within the industry and the calculations are the subject of this memo. Also note, that while wages were non-disclosable for most of the occupations with suppressed employment the next available wage level is presented and defined by the Wage Source column as follows.

1. WYXInd - Wyoming and Industry Specific, the wages a person employed in the occupation, in the industry, and in Wyoming averages.
2. WYXAgg - Wyoming Across Industries, the wages a person employed in the occupation in Wyoming averages.
3. USXInd - US and Industry Specific, the wages a person employed in the occupation, in the industry, in the US averages.
4. USXAgg - US Across Industries, the wages a person employed in the occupation in US averages.
5. SSXAgg – Although not included in demonstration tables provided with this documentation – Sub-State or Wyoming Regional wages for the specific occupation.

The adjusted percents are calculated by ordering the occupations within the industry by their employment descending and averaging the occupations percent of employment with the next two occupations percent of employment (demonstrated in Appendix IV, Table 2.) Subsequent to assigning every occupation within the industry the new rolling average percentage the sum of the Adjusted Percent is calculated (Appendix IV, Table 3). Logically, the sum of the rolling average no longer represents 100 percent and an adjustment factor is needed to compensate. This adjustment factor is "what the total should be" 100% minus "what the total is" 86.07% which equals 13.93% divided by what the total is 86.07% for the final adjustment of 16.19%. This adjustment factor is added to the rolling average percent proportional to the occupations rolling average percent. For example, the first occupation’s (SOC = 51-2092) rolling percent equals 10.17 or (15.20+9.20+6.11)/3. The adjustment factor outlined in Appendix IV, Table 3 equals 16.19 when applied proportionally the releasable percent for SOC 51-2092 in NAICS 334000 becomes 10.17 % + (16.19 % * 10.17 %) or 11.82 percent.

This procedure was developed in attempt to release a tool, which is applicable at a level specific enough to be useful, and at the same time does not release confidential information. By using the rolling average methodology the validity of the staffing pattern is maintained, whereby the most predominate occupations remain in the correct sequential order and maintain an appropriate magnitude. Questions and concerns should be addressed to me using one of the following contact methods (preferably e-mail).