ERS Response to SESRC Full Study Recommendation

VI. ERS RESPONSE TO SESRC FULL STUDY RECOMMENDATIONS

Statistical Design

The doubling of the sample size for prescreening is prudent given the challenging environment for surveying business establishments. Other findings regarding differential response rates for large establishments suggest oversampling these establishments by a factor of two. We agree that the response rates across other strata do not warrant additional corrections to the drawn sample. Given the much larger size of the sample and the possibility that not all of the sample will need to be released to meet the completed survey target, the main study sample will be partitioned into replicates of roughly 5,000 respondents, each with identical sample composition across the three strata.

Prescreening Protocol

The need to double sample size compared with the initial survey plan from the Supporting Statement is the strongest evidence of a fundamental change in the survey design strategy. The assumptions in the original plan that phone intensive contact would generate a large share of completes and significantly reduce nonresponse has not been borne out. The effective strategy suggested by the pilot is a much larger mail/web distribution, phone contact to complete a small share of surveys and more effectively direct respondents to the web, and limited ability of repeated phone contacts to significantly reduce nonresponse. Given the more limited role of phone contact in this new survey strategy the value of telephone prescreening for identifying contact individuals appears to be less effective in increasing response rates. Determining eligibility through a telephone prescreening protocol also appears to be largely redundant given the high quality of the BLS sample but should be used for the proprietary sample for states not granting access to the BLS sampling frame. Determining eligibility during data collection for respondents, and using this to estimate eligibility rates of nonrespondents is a common practice in establishment surveys and will allow deriving valid sample weights.

Change Incentives to $1 instead of $2

The Jefferson $2 as a token incentive was very appealing given the relative rarity of the bills that is more likely to grab a respondent’s attention and the very close association between the Department of Agriculture and Thomas Jefferson as the nation’s most renowned farmer-president. The original plan was to include a $2 incentive in two mailings. However, given the doubling of sample size we agree that two mailings of a $1 bill as a token incentive will still be effective in increasing response rates while maintaining the overall incentive level for the study. We anticipate some incentive savings for those respondents completing their surveys over the web or sending in completed mail surveys early as these cases will be removed from

Use web first (Group 3) contact sequence for full study

We agree that results from the pilot study strongly favor the Group 3 contact sequence. In addition to demonstrating the highest response rate in the pilot, the proposed protocol is also likely to minimize respondent burden, minimize the cost of data collection, and minimize the chance of data entry errors relative to the other modes relying more on phone or mail completes.

Survey Instrument

We agree that the survey instrument performed well in all three modes and there is no need for revision for the main study. The recommendation to add the question at the end of the survey to confirm whether the respondent agrees to be contacted if there is a need to clarify a response is a prudent addition if such a need should arise.

Anticipated Impact on Respondent Burden

The respondent burden reproduced from the original Supporting Statement Part A assumed phone completes would make up a large share of completes and that repeated telephone contact would reduce nonresponse:

FULL STUDY 30 MINUTE INTERVIEW
		Estimated Number of Respondents	Responses Annually per Respondent	Total Annual Reponses	Est Ave Number of Hrs per Resp*	Est Total Annual Hrs Resp Burden	Anticipated Completes
Business Phone Screener initial sample		30,000
Completed screener interviews	80%	24,000	1	24,000	0.07	1680	24,000
Attempted interviews (number not completing)	20%	6,000	1	6,000	0.04	240
Main Survey:
Phase 1: Business Phone Interview to All respondents						0
Completed interviews	47%	11,280	1	11,280	0.5	5640	11,280
Mail Short Form for Tel. Refusals
Completed 1 pg questionnaire	2%	480	1	480	0.1	48	480
Phase 2: Business 1st Follow-up Mail Survey to nonrespondents
Completed questionnaires	9%	2,160	1	2,160	0.5	1080	2,160
Phase 3: Business 2nd Follow-up Mail Survey to nonrespondents
Completed interviews	6%	1,440	1	1440	0.5	720	1,440
Phase 4: Business Web Questionnaire
Completed questionnaires	7%	1680	1	1680	0.5	840	1,680
Phase 1 to 4
Attempted interviews/contacts	29%	6.960	1	6,960	0.476	3,313
Total responding burden						13,561	17,040

The main changes proposed are doubling the initial sample, eliminating prescreening for the high quality BLS sample, and continue telephone prescreening with the proprietary sample. The best mode sequence from the pilot survey combined mail first contact with web link to encourage web completes. Respondent burden reduced by elimination of the telephone prescreening and respondent preference for more time efficient modes should accommodate the doubling of sample size while staying within the original burden estimate:

Survey	Samp Size	Freq	Resp Count	Freq Count	Min/Resp	Burden Hours	Nonresp Count	Freq X Count	Min/ NonResp	Burden Hours	Total Burden Hours
REIS Main Study
Screener for Proprietary Sample	3612	1	2167	2167	4.2	152	1445	1445	2.4	58	209
Advance Letter and Publicity Materials for BLS Sampleand Screened Proprieary Sample	58555	1	58555	58555	2.4	2342					2342
Phase 1: Business Mail Survey to all respondents	58555	1	6,120	6,120	22.2	2264					2264
Phase 2: Business 1st Follow-up Mail Survey to nonrespondents	52435	1	4,320	4,320	22.2	1598					1598
Phase 3: Web Survey to nonrespondents	48115		4,428	4,428	22.2	1638					1638
Phase 4: Phone Interview to norespondents	43687	1	720	720	30	360					360
Mail Short Form for Tel. Refusals	42967	1	1,440	1,440	6	144					144
Attempted interviews/contacts Phases 1 Through 4	41527						41527	41527	7	4364	4845
Total	60,000		17,028								13400

3