0990-0220

B. COLLECTIONS OF INFORMATION EMPLOYING STATISTICAL METHODS

1. Statistical Survey Methodology

In line with the previous data collection effort, DHHS expects to conduct web-based surveys on a 36-month cycle. While offices with few members in the population will do a census, larger grants or contracting offices will take a sample of no more than one-half the population to achieve plus or minus 3 percent precision at the 95 percent confidence level; with the largest offices surveying about 1,200 recipients or vendors. All sampling will be based on systematic random sampling—selecting every nth one from an alphabetical list of the population names. These procedures will be repeated for every survey cycle. Stratified sampling will be used whenever answers to survey questions are expected to differ significantly from one stratum to another in the grant recipient or contractor population (e.g., commercial firms vs. non-profit organizations). Each of our OPDIV offices will work with Dr. Dillman’s Tailored Design Method to achieve a high response rate, targeted at 80 percent. Sufficient OPDIV and Departmental resources have been allocated for the efficient and effective management and use of this survey information.

DA’s Acquisition Balanced Scorecard Users Group and OG’s Executive Committee on Grants Administration Policy (ECGAP) will continue to ensure that surveys are conducted in accordance with the required statistical methodology.

In addition, at OMB’s request, we would be glad to provide a copy of LMI’s “Government Manager’s Guide to Satisfaction Surveys and Performance Improvements” [second edition]. It was written by Larry Schwartz –one of DHHS’ survey consultants; and exemplifies the kind and quality of consultant support given to DHHS and its Users Groups. The paper’s guiding principles will help the OPDIVs to conduct the requested contractor and grant recipient surveys.

2. Procedures for Collecting Information

Each contracting and grants office will continue to follow a hypergeometric distribution by taking a 50 percent sample for populations of 1,000 or fewer, to achieve precision of plus or minus 3 percent at the 95 percent confidence level. A normal distribution would be used for any larger populations. Stratified sampling will be used when needed to ensure a representative sample. However, we believe stratification has limited applicability for our OPDIVs, because survey answers are not expected to differ greatly among population strata. DHHS will call for OPDIV data collection approximately every 36 months to minimize the burden on respondents, obtain timely feedback, address grant recipient or contractor concerns promptly, and encourage continuous improvement efforts.

3. Response Rates

DHHS and its OPDIV contracting and grants offices will continue to use Dr. Dillman’s Tailored Design Method for web-based surveys, to balance survey costs against the need for adequate response rates. Each of our OPDIVs or its support contractor plans to contact its own survey participants several times to ensure high response rates, targeted at 80 percent. For web-based surveys, our survey consultant—LMI—tracks this information electronically. Survey participants will continue to be contacted as follows:

• Alert message, alerting the selected sample to the upcoming survey.

• First cover message and survey instrument.

• Second cover message and survey instrument.

• First reminder and survey instrument.

• Second reminder and survey instrument, and

• Third reminder and survey instrument.

Post-survey analysis shows that OPDIV response rates under our previous data collection method averaged in the mid-60 percents, and were as high as 83 percent. To help ensure that we receive an 80 percent response rate across the board, we have instituted the following procedures:

(a) We are instructing our OPDIVs to “scrub” their mailing lists on a regular basis, to make them more current, accurate and complete. This will resolve the major obstacle to our achieving higher response rates.

(b) We have added a mandatory email address data element to our Departmental Contract Information System to collect and store email addresses.

Using the two new procedures stated above, we believe that DHHS can meet OMB’s target response rate.

4. Tests of Procedures

DHHS has pretested its web-based survey instruments to eliminate or revise questions, as necessary. In addition, we used the pretest to add material that the representative respondents strongly believed should be included. Further, DA and OG remain in close contact with industry and the research communities to obtain their views on the availability, disclosure, and reporting of this information. Essentially, DHHS continuously strives to minimize burden on respondents.

5. Consulting

We continue to retain the services of the Logistics Management Institute (LMI) to help us properly analyze and use the results of this survey effort. LMI points-of-contact is Isabela Castaneda (703-917-7533).