Response to OMB Comments Part 2

*Our responses to your questions are in red

Response Rate:

Would you please clarify what the expected response rate is? The original submission indicated an expected response rate of 50 – 80% but the supplemental materials indicate that it may be as low as 25%. All of which are quite low for a federal survey.

Please describe the specific measures you will use for non-response follow-up.

Given the low anticipated response rates, please describe your plans for non-response bias analysis. These plans are critical given the low expected response rates.

We apologize for the confusion. We anticipate at least a 25% response rate based on N=1,000 but still hope for a response rate of 50-80% considering our audience of local Planning Directors. So our goal is to get at least 250 responses which would have been impossible under the original plan with N=250, but hope for 500 to 800 responses based on a revised N of 1,000.

If indeed we have a low response rate, we will conduct a non-response bias analysis. That will consist of comparing community characteristics (size, state, region, housing characteristics, demographic characteristics) of the respondent communities with those of the non-respondent communities. We already have that information for all the CDBG-eligible communities in a data base. We will identify each community as a respondent or non-respondent (code of 1 or 2) and conduct difference of means or proportions analyses.

Frame Development:

Do you expect there to be any differences (i.e., socioeconomic, demographic) between the CDBG-eligible communities for which you can find contact information compared with those that you cannot. What are your plans for descriptive analysis comparing these groups? (HUD should have some information in-house that could be used to identify potential differences.)

We do not have any reason to believe that there will be any differences between the CDBG-eligible communities for which we find contact information versus those that we do not. However, just as with the non-response bias analysis, we will assign each CDBG-eligible community a code signifying either contact or non-contact and compare on community characteristics using difference of means or proportions analyses.