BJS memo

	U.S. Department of Justice Office of Justice Programs Bureau of Justice Statistics
	Washington, D.C. 20531

To: Brecht Donoghue, OJJDP

Benjamin Adams, OJJDP

From: Howard Snyder, BJS

Allen Beck, BJS

Date: September 28, 2016

Re: Comments on OMB Package for the 2016 JRFC

Washington, D.C. 20531

As requested, we have read the OMB Package you provided. Our reactions to the package are found in this memo and the comments embedded in its attachment.

Part A of the package is not in bad shape. It needs editing to bring it up-to-date and to make it less generic. Our notes on these and other points are embedded in the text and comment bubbles.

We believe Part B could be substantially improved. It has all the needed sections headers but the materials presented under each do not include the detail that is generally expected by OMB. Along with the comments and recommendations found attachment, there are some areas that need extra attention:

• There could be more detail on how Census systematically maintains the universe list. Currently, the text reads as if it is almost an informal process with no guarantee that the changes in the universe will be identified.

• Normally, a survey instrument (whether it be a new survey instrument or the repeat of a prior survey instrument) should be reviewed prior to each administration by experts in the field for change recommendations and then tested with a small sample of likely respondents to assess the viability of the items and the burden. [The package should also include the names of these individuals.] The instrument testing should also include some cognitive interviewing to obtain the reactions from likely respondents. All this may or may not lead to changes in the instrument. If changes are made, they should be detailed in the package. We assume there was little to no expert review or testing of the 2016 survey. If not, we suggest describing as best you can the last time this work was done and the effect it had on the survey content, possibility demonstrating the lack of need for revision of the survey.

• The package could use a more detailed data collection protocol, with definite dates for each step along the way. Also, any planned communication with the respondents should be described in the text with copies of the proposed correspondence and telephone scripts attached. Also the package would benefit from more detail about the nature, frequency and termination of your non-response follow-up procedures; this should include any plans for a greatly reduced critical item survey.

• While there is an imputation document attached to the package, a detailed summary of the processes should be included in the text. The attachment also would be better if it were written as a technical document rather than as a memo. [The current memo could also be edited to remove marginal text (such as the discussion about “makeshift beds”) and focus on the procedures that will be used to impute and weight the 2016 survey responses, and not what was done in 2014.

• Finally, as detailed in the attachment, the package should present a detailed budget for the work. This should include a breakdown of costs for the prime contractor and all subcontractors, along with the estimated personal costs for OJJDP staff working on the project.

We hope to find these comments useful. If you believe we can be helpful in the future, we would be willing to do so.

Attachment: Final JRFC Supporting Statements 5-31-16_w_BJS_notes.docx

Page 1 of 1