Appendix I.1 National Agricultural Statistics Service Comments

Appendix i.1

National Agricultural statistics service comments

Food and Nutrition Service

Revision to OMB # 0584-0607, School Meals Operations Study (SMO)

Doug Kilburg

NASS Reviewer

Most of the study was well documented and justified. Most of the suggested comments below are for enhanced documentation for mostly clarity purposes but may require some adjustments to the plans if the criteria aren’t met.

Part A:

A1: no comments

A2: We at NASS have found early reminders sent at short intervals are the most effective in soliciting improved response rates.

A3-A18: no comments

Part B:

B1: The proposed sampling strategy seems possible but will require some weighting strategy adjustments to make it feasible. Clearly outline how the weighting will be handled if the minimum number of reports are not met within a stratum and the reserve sample is utilized. Typically, sample quotas are only used in a nonprobability-based setting where sample weighting isn’t implemented.

B2: The documentation states, “The web survey of State CN Directors and the disaggregated administrative data collection will be conducted with a census of State CN directors. Because this study involves a census of SAs and a 100 percent response rate is expected, there is no need for sampling, weighting or nonresponse adjustments at the state level.” I’m not sure how a 100% response rate can be guaranteed from a survey of state CN directors. Even with statutory reporting, unless a penalty is enforced for not reporting. If not, it would be feasible to have a back-up plan in place for nonresponse weighting adjustments, if a 100% response rate is not obtained.

I did see a comment that telephone follow ups will be made with directors who do not complete the survey; however, a telephone follow-up is not likely to result in a 100% response rate and would necessitate a nonresponse weighting adjustment plan. It should be relatively easy to implement using the strata design.

B3. What are the characteristics of the respondents that you plan to assess for non-respondent bias across the frame, sample and respondents? You reference them, but never list them.

B4. Great responses to the results of the cognitive testing.

B5. No comments