Tuesday, July 14, 2009

Comments on EEO Data Collection Submitted

I may have failed to note USDA was requesting comments on a data collection of race/ethnicity/sex data. The notice was originally published on May 13, 2009, as FR Doc E911109, then amended in a notice published June 19 as E9-14444. Comments are still open until July 31, and you can see the documents and file comments at http://www.regulations.gov/

Anyhow I finally got my comments composed and submitted and here's the text. You get the advantage of some working links--the regulations.gov site doesn't recognize links.


"My comments are based on my past service in ASCS and FSA, with an early connection to the Service Center Information Management System (SCIMS). Since my retirement in 1997 I've been aware of the Pigford case and some of the GAO reports in this area.

I understand the June 19, 2009 amendment to the original request for comments to constitute a backing off from a specific process of data collection, which is good, but I'm still confused on some points.

I'd think the request for comments could have been improved by:

  1. citing the relevant governing documents issued by OMB. Based on my limited research, the OMB guidance issued on October 30, 1997 for the 2000 Census, that was to apply to other Federal programs by January 1, 2003, seems to be still the latest. (This is the change separating "Hispanic" as ethnic, rather than racial and permitting multi-racial categories.)
  2. mentioning the SCIMS database now being operated by FSA. It was originally intended to provide common producer/customer/employee data for the service center agencies. I'm not clear on the extent to which this intention has been achieved, but my interpretation of paragraphs 177-179 of handbook 1-CM is that the system already collects personal identity information according to the OMB guidance. Because the FR notice doesn't specifically mention SCIMS, it's not clear to me whether the planned information collection supersedes SCIMS, is part of SCIMS, or what. As published and corrected, the data collection doesn't really make sense--it mentions only collecting race/ethnicity/sex data but I assume it's tied to an individual, meaning the person's name and address must also be collected.

Specifically in response to the four points:

1 Is the collection necessary for performance of USDA functions? Given the criticism directed at USDA (Pigford, etc.): I'd agree USDA needs to be able to point to valid statistics on race, ethnicity, etc. But given that SCIMS currently collects some data for many of the universe of respondents, it's difficult to know whether this collection would represent an improvement over current practice.

2 Evaluate the accuracy of the agency's estimate of burden of the proposed collection of information, including the validity of methodology and assumptions used: Given the modifications in the June 19 notice, it's hard to evaluate the proposal. If the idea is that the business process already has the individual's name and address, so the collection is limited to race/ethnicity and sex, most individuals could respond in a few seconds, not 2 minutes. Even individuals who might need to decide whether to reply as multi-racial are not likely to increase the burden greatly. 2 minutes might be reasonable for a collection of name, address, plus race/ethnicity and sex, but you don't describe your collection that way.

I don't know how you came up with 14 million respondents--that's way above any estimate of producers and landowners I've seen. There should be a large degree of overlap between the three agencies, FSA, NRCS, and RD.

3 Enhance the quality, utility, and clarity of the information to be collected: The notice does not specify what information is to be collected . I assume the collection will follow the OMB guidance, but assumptions should be spelled out.

Given the suspicions of USDA among minority farmers I think it's important to increase public confidence in the data collected. I'd suggest running comparisons of the USDA data against Census data--for example, in areas with a high percentage of minority inhabitants, do rural zip codes show comparable percentages--does FSA have 10 percent Hispanics while Census has 25 percent? I'd also have OIG look specifically at existing data and have the FSA county office reviewers do spot checks.

You might also consider an annual mailing of the customer statement to everyone for whom FSA/NRCS/RD have data. That's the sort of thing Social Security Administration does, and it would give producers an opportunity to review their race/ethnicity/sex data.

4 Minimizing the burden of collection. Inasmuch as SCIMS has been operational for some time, I'd suggest a careful review of its operation would be instructive. What do USDA service center offices think of the process? What do those customers who use e-Authentication and the USDA e-gov process think? Should the data collected be shown on the USDA customer statement?

Whatever is going on behind the screen of the computer, I think it is most important that employees in the service center and those using the Internet application feel there's a seamless application for recording the data for a new customer. In other words, I'd drop any idea of a separate collection just for race/ethnicity/sex. Incorporate it in a standard process for registering new customers which can bridge into individual agency processes for their own programs.

Additional comments:

The new team at USDA might consider elevating the SCIMS function (or a revised and improved customer relationship) to the departmental level to provide greater visibility and clout. And, given the mandates in the 2008 Farm Bill for dealings with producers, USDA might investigate customer relationship management software. Finally, given the subsidies given to crop insurance companies, I'm not sure why they aren't included in the data collection. Surely it's a question whether there is outreach to minority farmers in this area.


Bill Harshaw
2420 Cloudcroft Sq.
Reston, VA 20191
bharshaw@hotmail.com"

No comments: