Supporting StatementA

Supporting Statement

U.S. Department of Commerce

U.S. Census Bureau

BC-1294, Field Representative Exit Questionnaire

BC-1294(D), Decennial Field Staff Exit Questionnaire

BC-1294(CM), Coverage Measurement Exit Questionnaire

OMB Control No. 0607-0404

1. Justification

1. Necessity of Information Collection:

Retention of trained field interviewing staff is a major concern for the Census Bureau because of both the monetary costs associated with employee turnover, as well as the potential impact on data quality. Therefore, in a continuous effort to devise policies and practices aimed at reducing turnover among interviewers, the Census Bureau collects data on the reasons interviewers leave their Census Bureau jobs.

The exit questionnaires are the instruments we use to collect turnover data from a sample of former current survey interviewers (field representatives) and decennial census interviewers (enumerators and listers). The goal or purpose of the exit questionnaires is to determine the reasons for interviewer turnover and what the Census Bureau might have done, or can do, to influence interviewers not to leave. Thus the exit questionnaires seek reasons interviewers quit, inquires about motivational factors that would have kept interviewers from leaving, attempts to identify training program strengths and weaknesses and their impacts on turnover, and explores the impact of pay, working conditions and supervisory styles on employees reasons for quitting.

As the environment in which surveys take place, the demographics of our labor force and the way surveys are conducted continues to change, it is important that we continue to examine the interviewers’ concerns about their job. Information provided by respondents to the exit questionnaire provides insight on the measures the Census Bureau might take to decrease turnover, and is useful in helping us determine if the reasons for interviewer turnover appear to be systemic or localized. The exit questionnaires have been shown to be useful and, therefore, we believe it is important to continue to use them to effect program planning and management.

The collection of this data is authorized by Title 5 U.S.C. Sections 301, 2301 and 3101.

Forms BC-1294 and BC-1294(D) are the instruments we currently use to collect turnover data from a sample of former current survey interviewers and decennial census interviewers, respectively. We are adding the BC-1294(CM) to this clearance to collect turnover data from Census Coverage Measurement (CCM) listers and interviewers. CCM operations are conducted to determine the number of people and housing units missed or counted more than once in the Census of the United States and Puerto Rico. The five major CCM field operations, which are conducted independently from census operations, are: (1) CCM Listing (2) CCM Housing Unit Followup, (3) CCM Person Interview, (4) CCM Person Followup, and (5) CCM Final Housing Unit Followup. Although the questions on the BC-1294(CM) are similar in nature to those on the currently approved decennial form, we found the BC-1294(D) could not be easily reformatted to accommodate the large number of CCM operations nor modified to address the issues specifically related to the Coverage Measurement operations (e.g., the use of laptops for the CCM Person Interview).

In addition to the new form BC-1294(CM), this submission includes changes to the BC-1294 and BC-1294(D) which reflect Census Bureau policy and procedural changes to current survey and decennial operations since the last request for clearance. For example, on the current survey side, field staff have reported an increased concern for safeguarding Personally Identifiable Information (PII) and Census Bureau confidential data on laptops, which has resulted in the use of multiple passwords and data encryption software. Thus we have submitted changes to the BC-1294 to determine if the Census Bureau’s increased security concerns, for the accountable property (laptops) and the increased security requirements pertaining to the restricted data contained on the laptops, are impacting FR turnover.

Recent changes in 2010 decennial operations have also made it necessary to revise the BC-1294(D). For example, the 2010 Nonresponse Followup (NRFU) operation that was originally going to involve computer-assisted data collection will revert to being a paper-based operation. That is, a paper instrument or questionnaire will be used to collect respondent data instead of a hand-held computer. Thus we are submitting changes to the BC-1294(D) to remove questions about the use and impact of automation on the retention of NRFU enumerators. In addition, the proposed revisions to the BC-1294(D) reflect the redesign of the 2010 Address Canvassing operation from what was done in the 2006 Census Test and the lessons learned during the 2008 Dress Rehearsal.

The specific changes to the current survey form BC-1294 and the decennial operations form BC-1294(D) are shown in Attachments 1 and 2, respectively, and the new Coverage Measurement form BC-1294(CM) is shown in Attachment 3.

2. Needs and Uses:

The costs to the Census Bureau resulting from interviewer turnover are numerous including:

• Loss of time the trainers and supervisors have invested in training the interviewers who leave;

• Training pay given to interviewers who leave while they are learning the job;

• Recruiting costs and fees to find replacements;

• Cost of training replacement interviewers;

• Costs to data quality while replacement interviewers learn the job;

• Disruption to operations due to turnover; and

• Loss of knowledge the leaving interviewers take with them.

To accomplish the goal of reducing interviewer turnover, Census Bureau planners and decision makers must fully understand the relative importance and interaction of possible contributory factors. These factors include, but are not limited to:

• Job factors such as salary and benefits, career advancement and performance expectations;

• Working conditions such as unsafe areas/neighborhoods and direct interaction with the public;

• Work hours and workload;

• Use of computers/automation;

• Training;

• Supervision and training of supervisors; and

• Recruiting methods and selection criteria.

Since the initial implementation, the use of the BC-1294 and BC-1294(D) have helped the Census Bureau identify specific reasons for interviewer turnover. From these forms, we have learned that the causes of interviewer turnover are often a combination of reasons rather than one single reason. We have also learned that there are some reasons for turnover which are within the Census Bureau’s control and some which are not. In response to the reasons identified by the exit questionnaires over which the Census Bureau can exercise some control, we have initiated and continue policies and procedures designed to reduce turnover. For example:

• As current surveys moved to a computer assisted interviewing environment and/or new software on the automated instruments, there was concern about the possible impact of these technological changes on interviewer turnover. We learned from the exit questionnaire that neither the conversion to automated interviewing instruments nor new software (e.g., Blaise) had a substantial effect on current survey interviewer turnover.

• Using earlier results from the exit questionnaire, we discovered that interviewer turnover on current surveys often resulted in feelings of isolation. As a result, the regional offices initiated programs to enhance communication between/among interviewers in the field and the office, and promoted a more nurturing environment through the use of interviewer teams or a “buddy” system approach for new interviewers, in particular.

• The Census Bureau’s ability to retain a competent, motivated and representative staff of local census-takers is one of the most important factors affecting the quality, length of time required, and overall costs of the field data collection phase of the decennial census. Thus we were also able to use the results of the Census 2000 exit questionnaire to justify the use of localized pay rates, and in many large urban areas we now have pay rates that are more closely tied to the local labor market. And through the years, exit results have also led to improvements in recruiting methods and techniques, and training protocols and materials.

We require more than one version of the form because the nature of decennial interviewing differs from current survey interviewing, and to address the differences between decennial and coverage measurement operations. For example, decennial interviewing is short term, intensive and concentrated where current survey interviewing is continuous and more diverse. Although the BC-1294(D) largely covers the same topics as the BC-1294, the questions and response choices on the BC-1294(D) have been tailored to decennial operations and are more in-depth. In addition, while the questions on the BC-1294(CM) are similar in nature to the BC-1294(D), the five Coverage Measurement operations, which are conducted separately from decennial operations, warrant their own questionnaire to accommodate the various operations and the complexity of their differences. For example, the CCM operations include both automated and paper based data collection. In the Listing operation, listers canvass and identify all living quarters in selected areas in the CCM sample. During the Housing Unit Followup, listers use a paper questionnaire to resolve differences between the CCM and census address lists. The Person Interview is an automated operation during which interviewers contact selected addresses in the CCM area and ask detailed questions to determine who lived at the address at the time of the interview and who lived at the address on Census Day. The Person Followup operation, which is conducted using a paper questionnaire, resolves differences between the household rosters collected during the CCM Person Interview and during the Census.

The information collected via the three Exit Questionnaires will help the Census Bureau develop plans to reduce turnover in its current survey, decennial and coverage measurement interviewing staff. This, in turn, will allow for better informed decisions regarding the field workforce and implementation of more effective pay plans, selection procedures, interviewer training, and retention strategies for all interviewers.

Information quality is an integral part of the pre-dissemination review of the information disseminated by the Census Bureau (fully described in the Census Bureau’s Information Quality Guidelines). Information quality is also integral to the information collections conducted by the Census Bureau and is incorporated into the clearance process required by the Paperwork Reduction Act.

2. Use of Information Technology:

Exit interviews will be conducted by telephone. No automated collection techniques will be employed to gather data for this questionnaire. The forms have been designed to allow for in-depth probing, when/where necessary, during interviews, while facilitating data capture by limiting the amount of hand written comments/information through the use of predefined codes.

2. Efforts to Identify Duplication:

We will collect data from interviewers who have recently resigned from the Census Bureau. This data is not available from any other source. While similar studies concerned with identifying causes of interviewer turnover have been done by the Census Bureau in the past and the information has been used to improve management policies, the content and nature of the surveys used to collect data today, the methods of data collection (e.g., automation), the composition of our field workforce, and the climate in which we now conduct both current survey and decennial interviews are ever-changing, making past information less relevant for current management decisions.

2. Efforts to Minimize Burden:

Efforts to minimize burden include the use of telephone methodology, in lieu of a mail out/mail back strategy, and the relative brevity of the questionnaire. The use of sampling and predefined response options, based on previously administered exit interviews, also contributes to reducing respondent burden.

2. Consequence of Less Frequent Collection:

This will be a one-time request for information.

2. Special Circumstances:

There are no special circumstances.

2. Consultation Outside the Agency:

A pre-submission notice was published in the Federal Register on February 21, 2008. We received two responses to our solicitation. These responses have been summarized and addressed below. Relevant comments and/or suggestions include: 1) inform respondents that their answers are not shared with former supervisors; 2) modify questions to reflect current Census Bureau policy on the use of hand-held computers; 3) reevaluate the estimated time to conduct the interview; and 4) provide respondents (e.g., via email) with an advance copy of the questions prior to the telephone interview to allow respondents to be prepared to answer the questions. The Census Bureau response to these comments/suggestions are as follows:

1. The instructions read to the respondents tell them that their information will be protected from disclosure under the Freedom of Information Act and the Privacy Act to the extent provided by law. We will also modify the instructions to include a statement that a System of Records Notice exists which provides more details about the ways information they provide can be used. The SORN is entitled COMMERCE/DEPT –18, Employees Personnel Files Not Covered by Notices of Other Agencies. The routine uses that apply are in Section 8 as follows: “A record in this system may be disclosed to the Office of Personnel Management: for personnel research purposes; as a data source for management information; for the production of summary descriptive statistics and analytical studies in support of the function for which the records are collected and maintained; or for related manpower studies.”

2. The Census Bureau will only use hand-held computers (HHCs) for the 2010 Census Address Canvassing operation. All Nonresponse Followup (NRFU) operational questions about and/or references to the HHC in the BC-1294(D) have been removed.

3. The estimated burden time was based upon our administration of the questionnaire. We conducted several pilot administrations of the revised 2008 forms and found the estimate to be consistent. Additionally, we believe the removal of the hand-held computer questions may actually reduce the burden stated on the BC-1294(D) even more.

4. Sending the questionnaire to respondents before calling them is prohibitive since personal email addresses are not required from employees and therefore are not readily available. Although we collect home mailing addresses from employees, distributing the questionnaire through regular mail is not cost effective and past experience has indicated that this action could hinder cooperation instead of facilitating it.

2. Paying Respondents:

Respondents, all of whom are former Census Bureau interviewers, will not be paid.

2. Assurance of Confidentiality:

The information collected is voluntary. Each former interviewer contacted will be told that his/her participation is voluntary, and that the information obtained from him/her will be protected from disclosure under the Freedom of Information Act and the Privacy Act to the extent provided by law.

2. Justification for Sensitive Questions:

None of the questions in the BC-1294, BC-1294(D) nor the BC-1294(CM) are sensitive in nature.

2. Estimate of Hour Burden:

On average, the respondent burden for the BC-1294 should not exceed seven (7) minutes. The estimated number of current survey field representatives to be contacted, by telephone per fiscal year, should not exceed 500. Therefore, the maximum estimated annual burden for the BC-1294 is approximately 59 hours.

Telephone interviews with former Address Canvassing listers, former Update Enumerate and NRFU enumerators, and Coverage Measurement listers/interviewers should take no more than ten (10) minutes. The BC-1294(D) and the BC-1294(CM) each contain a few more in-depth questions about decennial and Coverage Measurement pay and working conditions on listers’/enumerators’/interviewers’ decisions to leave the Census Bureau. Therefore, on average, they take a little longer to complete than the BC-1294. The estimated number of decennial listers and/or enumerators to be contacted per fiscal year should average 1,500: 400 Address Canvassing listers, 100 Update/Enumerate enumerators and 1000 Nonresponse Followup enumerators. Therefore the maximum estimated annual burden for the BC-1294(D) is approximately 251 hours: 67 hours for AC, 17 hours for U/E and 167 hours for NRFU. The estimated number of CCM listers/interviewers to be contacted per fiscal year should average 500. Therefore, the maximum estimated annual burden for the BC-1294(CM) is approximately 83 hours.

2. Estimate of Respondent Cost Burden:

There are no costs to the respondents other than their time.

2. Estimate of Agency Cost Burden:

The time and material costs of the data collection, which will be borne by the Census Bureau, is expected to be approximately:

	Cost to Agency
Form	Materials	Time
BC-1294	1,000.00	2,700.00
BC-1294(D)	6,000.00	11,700.00
BC-1294(CM)	12,000.00	3,900.00
Total	19,000.00	18,300.00

Material cost for forms design and the estimated cost for time, which is based on burden and staff hours, would not have been incurred without this collection of information.

2. Reason for Change in Burden:

In anticipation of the 2010 decennial census, we have increased the number of respondents annually and therefore the burden hours have also increased. The current OMB inventory is 84 burden hours. The estimated inventory number for this clearance is 393 burden hours, of which 83 burden hours are attributable to the new BC-1294(CM).

2. Project Schedule:

The BC-1294 data will be collected approximately every quarter from a sample of one-half of all current survey interviewers who voluntarily resign within the sampling period covered. Respondents will be contacted by telephone to complete a questionnaire.

Data will be collected using the BC-1294(D) beginning approximately two weeks after the start of the 2010 Census Address Canvassing operation and again approximately two weeks after the start of the 2010 Census Update Enumerate (UE) and Nonresponse Followup (NRFU) operations. We expect to hire approximately 140,000 listers to complete the 2010 Address Canvassing operation, which will occur in two-10 week overlapping waves, or approximately 15 weeks from April 6 – July 17, 2009. Approximately 600,000 enumerators will be hired to complete the 2010 UE and NRFU operations. The Update Enumerate and NRFU operations will last approximately 8-10 weeks and will be conducted in the April through July 2010 timeframe. Data will be collected beginning two weeks after the start of each operation and continue until (approximately) two weeks after the completion of each operation. The proposed sample size of enumerators and listers reflects the expected high turnover rate for decennial operations.

Data will be collected for the Census Coverage Measurement (CCM) operations using the BC-1294(CM) approximately two weeks after the start of each operation and continue until (approximately) two weeks after the completion of each operation. The CCM Listing operation that takes place between September and December, 2009, will hire approximately 5,000 listers. The CCM Housing Unit Followup operation will take place between late February and early April, 2010 and will hire approximately 6,200 interviewers. The CCM Person Interview operation occurs from late August through late October, 2010 and will hire 6,500 interviewers. The CCM Person Followup operation takes place from early January through late February, 2011 and will hire approximately 5,500 interviewers. The CCM Final Housing Unit Followup operation will be conducted from April through mid-May of 2011 and will hire approximately 3,050 interviewers.

The Census Bureau does not plan to publish results or conclusions derived from the information obtained from the exit questionnaire. All analyses and conclusions are for internal use only.

2. Request Not to Display Expiration Date:

We will display the assigned expiration date on the information collection instrument. During the introduction that we read to the respondents, we inform them of the OMB control number and its expiration date.

2. Exceptions to the Certificate:

There are no exceptions.

9