07-orr09.sup

SUPPORTING STATEMENT

Annual Survey of Refugees

(Form ORR‑9)

PART A: JUSTIFICATION

1. Circumstances:

The Office of Refugee Resettlement (ORR) is requesting an extension of OMB approval of the questionnaire for its annual telephone survey of refugees in the United States. This questionnaire is similar to that last approved by OMB (OMB No. 0970‑0033) in 1993, with an expiration date of September 30, 1996. The results of the survey are used to address the requirement in section 413(b)(1) of the Immigration and Nationality Act (as added by the Refugee Act of 1980) that ORR report annually to Congress with:

an updated profile of the employment and labor force statistics for refugees who have entered under this act within the five-fiscal-year period immediately preceding the fiscal year within which the report is to be made and for refugees who entered earlier and who have shown themselves significantly and disproportionately dependent on welfare as well as a description of the extent to which refugees received the forms of assistance or services under this chapter during that period.

As in previous years, the contractor will conduct the refugee survey in September of each year, with preliminary results reported in ORR's annual Report to Congress. ORR has commissioned a survey of Southeast Asian refugees since 1975, when these refugees first began arriving in the U.S. in large numbers. Beginning with the 1993 survey, ORR expanded the refugee universe to sample other refugee groups. At the present time, the survey sample population includes refugees from every country as well as Amerasian immigrants¹ and Cuban and Haitian entrants². The combination of these groups (referred to as "refugees" for the sake of brevity hereafter) is the universe of persons eligible for ORR-supported programs.

Each year, ORR produces for the contractor reliable arrival data for the largest countries or regions of origin. The contractor must trace the sample of refugees, interview them, and produce a data tape and tabulations on a timely basis. Interviews are conducted in Russian, Creole, Spanish, Farsi, Arabic, Lao, Vietnamese, and Hmong. This sampling design is explained further in Part B.

2. Purpose and uses:

The initial use of the data collected will be to fulfill the requirement quoted above: to provide employment, labor force participation, and welfare utilization data for Congress. These data cover the following major areas:

• Characteristics of the respondents, analyzed by year of entry,

• Description and analysis of refugees' employment characteristics,

• Description and analysis of refugees' use of training programs and cash and medical assistance, and

• Analysis of secondary migration.

ORR program mangers use these data to formulate policy options for program development. Selected findings are incorporated into documents and testimony prepared for Congressional hearings on appropriations, reauthorization of the Refugee Act, and annual admissions levels for refugees. This survey is the only reliable source of such information. ORR receives numerous requests for copies of its findings from academic researchers and other interested parties.

If the findings from this series of surveys were not available, ORR officials would operate from a much poorer knowledge base. Because of the survey, ORR has been able to monitor trends in refugee adjustment. A landmark evaluation study of U.S. immigration statistics was released in 1985 under the auspices of the National Academy of Sciences, Committee on National Statistics ("Immigration Statistics: A Story of Neglect"). The ORR Annual Survey was cited in it as a useful example of a way to obtain follow‑up data on the progress of refugees, and the evaluators recommended that a similar panel study of (non-refugee) immigrants be established.

3. Improved information technology:

ORR has not identified any possibilities for reducing the burden in this information collection by using improved information technology.

4. Duplication:

The ORR project officer and other staff members maintain active liaison with other public and private agencies that collect information on refugees. A partial list of such agencies would include the Department of State, the Immigration and Naturalization Service, the Census Bureau, InterAction, the Urban Institute, the Refugee Policy Group, national offices of the voluntary resettlement agencies, public interest groups, and many academic researchers.

Through these liaison activities, ORR staff members maintain a thorough knowledge of current and planned data collection on the refugee population and portions thereof. ORR's research and data collection activities are carefully planned to complement and coordinate with research and data collection activities of other agencies. The ORR Annual Survey does not duplicate any other information collection; it is unique in its frequency, subject matter, and geographic coverage. No similar information is available which could be used or modified to serve the required purpose. ORR's other quarterly information collections are based on reporting by State refugee programs on their service and public assistance populations. These reports present aggregated data from which no detailed analysis of how the experiences and characteristics of refugees affect their adjustment can be performed. In addition, reports based only on service and public assistance populations necessarily present a biased picture--not applicable to the general refugee population.

ORR has undertaken special studies of the adjustment of selected refugee groups, based on more detailed personal interviews than are feasible in the Annual Survey. These studies are extremely useful for the sophisticated insights that they can provide into the process of refugee adjustment. However, all such surveys are necessarily limited in their geographic, ethnic, and cohort coverage. None can be generalized to produce a comprehensive description of the refugee population. Moreover, such studies cannot be completed rapidly enough to meet Congressional reporting requirements, and their cost is substantially higher than that of the Annual Survey.

Special surveys exist of the total U.S. population that provide periodic descriptions of some of the topics covered on the ORR Annual Survey. The Current Population Survey is the source of labor force and employment statistics on a monthly basis, as well as information on a variety of other topics. The Survey of Income and Program Participation focuses on the use of various types of public assistance programs. Although these two surveys utilize very large samples, the refugee population (Southeast Asian and otherwise) is still too small in terms of the total U.S. population (less than four-tenths of one percent) to appear in these samples in numbers adequate for separate analysis.

5. Small businesses:

Not applicable; respondents are individuals. ORR is pleased with the current data collection and does not intend to expand or reduce the scope of the survey instruments. for future years.

6. Frequency of collection:

If this survey were conducted less often than annually, ORR would fail to comply with the reporting requirements specified in the Refugee Act. Rapid and unexpected change is characteristic of the refugee program; admissions levels, ethnic composition, geographic placement of refugees, welfare utilization and labor force participation levels all fluctuate significantly from year to year. Because of such rapid changes in both program and client characteristics, ORR has a continuing need to update its information base at least yearly in order to monitor trends.

7. Special circumstances:

No special circumstances are involved: this information collection is consistent with OMB guidelines specified at 5 CFR 1320.6.

8. Consultations:

Consultations with a variety of experts on the availability of data and on desirable methodologies for collecting the required data are an ongoing ORR staff activity. Discussions with users of the data, including Congressional staff, are also conducted to ensure that the data meet their prime purpose of satisfying legislative requirements for the reporting to Congress of the current status and progress of refugees. The major mechanism for consultation on the design and uses of the annual survey is an informal one and involves a wide variety of ongoing interactions between ORR staff and other researchers and analysts in both the public and private sectors. Past consultations resulted in the expansion of the Annual Survey in 1993.

9. Remuneration to Respondents: None.

10. Assurance of confidentiality:

In the letter to respondents that serves to introduce the study and request cooperation, respondents are told, "The information which you provide will be used by the U.S. Government to gain greater insight into the adjustment process as you are experiencing it .... Information on your household will not be released to anyone unless required by law."

This assurance of confidentiality is in accord with ORR policy. The findings from the survey are used only to describe and make inferences about refugees as a group. The contractor maintains tracing information and sample identifiers under tight security.

11. Sensitive questions:

No questions of a sensitive nature appear on this survey instrument.

12. Estimate of Collection Burden:

The contractor times each interview. Over the past three years, the average response time has been approximately 45 minutes. The total respondent burden in FY2006 is thus 1,022 hours (1,363 respondents x 45 minutes/interview). Assuming the value of respondents' time at $14 per hour, the estimated cost to the public is $14,308. The respondents bear no direct cost other than their time to participate in the study.

13. Annual Cost Burden: There are no direct monetary costs to respondents other than their time to participate in the study.

14. Annualized Cost to the Government:

ORR estimates that the cost for the conduct of the Annual Survey in 2007 will be about the same as for FY 2006--approximately $707,000.

15. Changes in burden:

The total estimated burden for this information collection is 1,022 hours—an increase of 353 hours from the 669 hours estimated previously. This is the result of a larger number of interviews (217) and a slightly longer time for interviewing.

16. Publication:

The findings from this survey will be compiled and disseminated in the ORR Report to Congress. In addition, as in past years, selected findings will be incorporated into materials prepared for Congressional hearings, policy options papers, and other papers prepared by ORR staff members. In past years, the Report to Congress has been ready for distribution in mid-summer of the year following the end of the fiscal year.

17. Expiration date. May 31, 2010.

18. Exception to the certification statement. Not applicable.

PART B. Statistical Methodology

1. Respondent universe and selection:

The 2007 annual survey will continue as a panel study. ORR restricted the survey to a five-year cohort of refugees, beginning in 1985. This is consistent with the legislative requirement, and since the minimum waiting period for attaining U.S. citizenship is five years after arrival, it is a natural cutoff point. As a consequence, refugees sampled in 2007 who arrived before May 1, 2002, will be dropped from the panel for the 2007 sample. Refugees who arrived between 5/1/02 and 4/30/06 will be re-interviewed in 2007. The remaining 2007 sample will consist of a sample of refugees who arrived between 5/1/06 and 4/30/07. ORR plans to continue to drop refugees from the panel after their fifth year.

In practice, the universe contains a large proportion of children, and the contractor will interview the heads of their household or other adults. Since the sampling frame consists of individuals, the contractor draws a simple random sample of individuals, identifies the heads of the households to which those individuals belong, and interviews the heads. If the head of the household is unavailable during the interview period, another adult in the household is interviewed.

For the 2006 survey, 1,363 households were contacted and interviewed. Refugees included in the 2005 survey who had not yet resided in the U.S. for five years were again contacted and interviewed along with a new sample of refugees, Amerasians, and entrants who had arrived between May 1, 2005 and April 30, 2006. Of the 2,667 re‑interview cases from the 2006 sample, 1,113 were contacted and interviewed, and 86 were contacted, but refused to be interviewed. The remaining 1,468 re‑interview cases could not be traced in time to be interviewed. Of the 874 new interview cases, 250 were contacted and interviewed, another eight were contacted, but refused to cooperate, and the remaining 616 could not be traced in time to be interviewed. The resulting responses were then weighted according to year of entry and ethnic category.

In addition, of the 1,468 re-interview cases that could not be traced in time to be interviewed, four were deceased, six moved back to their native country, and 1,458 households had wrong or disconnected phone numbers. Of the 616 new interview cases that could not be traced in time to be interviewed, 36 households had wrong or disconnected phone numbers.

2. Procedures:

The basic sampling frame is the ORR Master Data File, which contains information on refugees and Amerasian immigrants at the time of arrival in the country, including the address of a relative or sponsor for most refugees. Cuban and Haitian entrants are sampled from a list obtained from the Community Relations Service of the U.S. Department of Justice. A simple random sample of each month's arrivals is drawn from the sampling frame. The random sample is drawn by a computer program that selects all eligible persons from the ORR and entrant files, creates a subfile of eligibles, generates a random start, and continues sampling one of every 250 records until the subfile end is reached. Stratification is impractical in light of the month-to-month change in the ethnic composition of the arrival population.

An E-file and printed listings of the selected sample are generated and given to the contractor. All available resettlement information on a refugee, including detailed name and address of his sponsor, is forwarded to the contractor. The contractor immediately begins tracing activity for that sub- sample, establishing contact and asking to be notified in the event of a move.

The contractor maintains records of the tracing work done in earlier surveys and the addresses and telephone numbers of persons contacted. They will rely on these records for re-contacting the earlier samples in 2006.

The 1,363 refugee households interviewed in 2006 included 8,631 refugees of working age (age 16 or older). With a sample this size, ORR can be 95 percent confident that the error of our estimates for households is at most 0.03 and that the error of our estimates for working-age adult refugees is at most 0.01.

3. Response rates:

Response rates are maximized in these surveys by intensive tracing activities. Different tracing techniques are used, depending on the information available. Where a mailing address is known, the contractor mails a letter printed in both English and in the respondent's native language to that address, explaining the survey. The letter includes a letter to be returned, indicating willingness to be interviewed and furnishing a telephone number where the respondent can be reached.

ORR’S contractor DBConsulting ofRockville, Maryland uses available information on the city and sponsoring agency of record, as well as the name, address and telephone number of the local sponsor, who is often a relative. Agencies are contacted by letter and by follow‑up telephone calls for assistance in locating the respondents. Each refugee for whom a current address is known or obtained receives a letter of explanation and a response letter. Interviews are conducted by telephone in the respondent's native language. The interviewers make up to four or five telephone calls or more as necessary to speak with the head of the household or a person designated by the family to respond. If this number of calls proves unsuccessful in reaching the desired respondent, the interview is conducted with another adult member of the household, if possible.

Tracing refugee families has roved to be a difficult task, given that many refugee families move addresses frequently, generally in anticipation in another community. Approximately 15 percent of refugee families move across State lines every years, many more from community to community within States. Friends and relatives are contacted by DBConsulting for a forward address, but few are willing to supply a new address. Friends and family are generally reluctant to divulge a forwarding address to a stranger, even one from an organization with ORR documentation, Refugee communities typically have an unyielding fear of the authorities, fearing that agencies in their countries of origin are probing into their activities through phony surveys or studies

4. Tests of Procedures:

No pretests are planned for this survey. As the most recent in a long series, it benefits from experience gained in the earlier surveys. The questions and procedures reflect 19 years of experience in conducting similar surveys. Past modification, such as the creation of a panel survey, have made ORR data collection analysis stronger, and proposed modifications, particularly the creation of a linked longitudinal file, will further strengthen this trend. The panel format made possible a direct assessment of the process by which refugees gain skills, move into the labor market, locate employment, and become self-sufficient over time. The linked longitudinal file will permit more refined judgments and increase ORR's ability to follow refugee adjustment to American life over time.

5. Statistical contacts:

For advice on statistical aspects of the research design, ORR staff consulted with Pat Brannen (401-5096), statistician of the Division of Quality Control, Office of Family Assistance, and Leonard Sternbach (401-5600) of the Office of Policy and Evaluation of the Administration for Children and Families.

Start here on Monday

1 Amerasians are children born in Vietnam to Vietnamese mothers and American fathers. Although the almost 72,000 Amerasian youth and accompanying family members have been admitted as non-refugee immigrants, they are still entitled to the same benefits and services as refugees.

2 Some Cuban and Haitian nationals arrive in the U.S. as refugees; the rest as entrants. The latter are eligible for services and benefits to the same extent as the former.

13