Supporting Statement Section B

New Small Business Pulse Survey

During COVID-19 Pandemic

Request for OMB Approval

April 20, 2020

Supporting Statement Section B

As noted in Part A, the Small Business Pulse Survey is in experimental phase. Fielding this survey during the COVID-19 pandemic is designed as a proof of concept. As such, the methodology detailed here may adapt over time. As the methodology adapts, we will submit non-substantive changes to this information collection request to OMB. In addition, we will provide all information about sampling, weighting, post-survey processing, and cognitive testing results on our website to maximize transparency for the public. This transparency is especially important given the predicted value to decision makers of having weekly state-by-state estimates of the effects on business operations, finances, and expectations for small businesses.

1. Universe and Respondent Selection

The target population is all non-farm, single-location employer businesses (receipts ≥ $1,000) with 1-500 employees in the 50 states, DC and Puerto Rico, and in industries covered in the Economic Census. The Business Register contains validated e-mail addresses for 940,588 businesses in the target population. These email addresses will be updated and supplemented with emails collected via other means, such as the Census Bureau’s Customer Respondent Management tool used for respondent communication across economic surveys, to identify the best contact for the business. For further discussion of the target population and use of email address, see Attachment H.

The universe will be subdivided into nine groups. Each group of ~100k businesses will receive an email with a survey link one week out of the nine week data collection period. This will minimize respondent burden, while also providing the data needed to produce detailed estimates by sector and state and for the top 50 MSAs. We expect to complete 9 weeks of data collection. Additional weeks of collection could be possible, and we will work in coordination with OMB to make that determination. If additional weeks of collection are deemed warranted, we will update documentation through a nonsubstantive change request to this package. Efforts will be made to account for potential biases in survey estimates due to the non-probabilistic sample by applying a weighting class adjustment. We are working to develop this adjustment in parallel with initiating data collection, and we expect that the adjustment methods may change as we learn through this collection process. For further discussion of survey methodology, see Attachment H.

2. Procedures for Collecting Information

Businesses will be contacted via email based on known email addresses that have been collected across economic programs. As mentioned earlier, the sample will be split so that ~100,000 different respondents are selected weekly to reduce burden and survey fatigue. Emails will be delivered in increments of 20,000 between the hours of 9AM -6PM every Sunday for nine weeks. An hour will be allowed for the delivery of the 20,000 emails and an hour will be allowed after the conclusion of the sending batch to gauge optimal timing of email delivery. The respondent can respond at any point, but their response will be allocated to the week in which they responded. (The methodology to describe how we will handle this allowance in producing the estimates is available in Attachment H.) The email will include a Thursday due date. Business who have not responded may receive a follow-up email reminder on Friday. For the first several weeks of the SBPS, all non-responding businesses will receive a follow-up email. Once we have a better sense of the response to the survey, we may target the non-responding companies to ensure that we have adequate representation to publish comprehensive statistics.

Data collection for the new SBPS will be conducted using an initial email invitation to the in-scope population. The email invitation will describe the purpose of the survey collection and will contain the authentication code that is linked to the business’s EIN in our internal database. The website address will direct respondents to a landing page that provides motivational text to encourage response, links to releases of the data products, and a direct link to the Centurion landing page to respond to the survey. As of 4/20/20, the landing page is not live off census.gov, but screenshots from the page can be found in Attachment C. The production landing page will be found at https://census.gov/businesspulse. See Attachment I for examples of the survey invitation email and follow-up emails.

The Census Bureau’s Centurion team developed the required electronic collection infrastructure including a welcome screen, screens for providing response to survey questions, and finally a submission complete/thank you screen. Once credentials are established, the respondent will be presented with the established 16 questions. See Attachment A that provides screenshots from the collection instrument of the survey content.

Messaging on the SBPS landing page will speak to the legitimacy of the collection and will include Frequently Asked Questions (FAQs). Additionally, respondents can respond to the email address with questions that will be answered by staff at Census Bureau Headquarters.

The data collection schedule by week with corresponding due dates is as follows:

Reference Week	Dates	Due Date
1	April 26-May 2	April 30
2	May 3-May 9	May 7
3	May 10-16	May 14
4	May 17-23	May 21
5	May 24-May 30	May 28
6	May 31-June 6	June 4
7	June 7-13	June 11
8	June 14-20	June 18
9	June 21-27	June 25

Knowing that small businesses are actively dealing with the COVID-19 pandemic, the Census Bureau is optimistically aiming for a response rate of 15-20%. This data point is based on recent use of email blasts in Economic Programs like the Economic Census, which has led to increases in response. Additionally, a similar survey produced by the UK’s Office of National Statistics, but with a much smaller sample of businesses (18k), resulted in response rates of 21% for the first wave, followed by 26% and 30% in the second and third waves.¹ For further discussion of response rates and nonresponse bias, see Attachment H.

3. Methods to Maximize Response

The Census Bureau has put in a place the following to maximize response:

• Respondents will receive an email invitation with a direct link to complete the survey. They will simply need to enter the authentication code credential into the Centurion instrument.

• The collection instrument is optimized for electronic response, including the option to respond via a mobile device.

• All questions are checkbox responses in order to limit burden and maximize response.

• A landing page was developed off census.gov that provides motivational text to encourage response, links to releases of the data products, and a direct link to the Centurion landing page.

• We expect that real-time survey results displayed on the dashboard will generate interest in the products and businesses will see value in responding.

4. Tests of Procedures or Methods

Due to timing constraints, we were unable to conduct formal testing of the content on the new survey ahead of the first wave of collection. However, the questions have been reviewed by Census Bureau staff who are experts in the field of questionnaire design. The process for expert review in lieu of in-field expert review was approved by the Methodology and Standards Council at the Census Bureau. In addition, we plan to cognitively test the new content in parallel with the first week of data collection and to continue refining the questions as we identify problems with comprehension. We commit to continued review of the data collection methods, survey content, improving response rates, weighting methods, and other aspects of this experimental design. Changes to the methods and content will be documented through nonsubstantive change requests before they are implemented.

The Census Bureau will release these data under the auspices of its Experimental Statistical Product Series. Information on the Series is available at https://www.census.gov/data/experimental-data-products.html#.

5. Contacts for Statistical Aspects and Data Collection

Nick Orsini, Associate Director for Economic Programs, may be contacted (301-763-6959) regarding the statistical and data collection aspects of the Small Business Pulse Survey.

1

https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/bulletins/coronavirustheukeconomyandsocietyfasterindicators/16april2020

1