2008 Survey of COFA Migrants - Supporting Statement B 05232008

3

SUPPORTING STATEMENT – PART B

U.S. Department of Commerce

U.S. Census Bureau

2008 Survey of Compact of Free Association (COFA) Migrants

OMB Control No. 0607-XXXX

Collection of Information Employing Statistical Methods

1. Description of Universe and Sample Selection

The Department of Interior (DOI) has requested the Census Bureau to produce estimates of the number of qualified nonimmigrants (also referred to as Compact of Free Association (COFA) migrants) from the Republic of the Marshall Islands, the Federated States of Micronesia, and the Republic of Palau residing in Guam and the Commonwealth of the Northern Mariana Islands (CNMI) in 2008 as required by the COFA Amendments Act of 2003.

These estimates for Guam and CNMI will be based on data collected from a representative sample of the addresses in each area in 2008. The Guam and CNMI estimates are expected to have the same level of reliability as an estimate of COFA migrants residing in Hawaii that will be derived from data collected in the American Community Survey (ACS), which is expected to have a 15% relative error (i.e., one standard error/estimate).

The sample universes for Guam and CNMI are all of the addresses (both housing unit and group quarters) on these islands in 2008 and the target population will be the COFA migrants residing at these addresses. A sample of these addresses will be selected in several stages. First, the Census 2000 tabulation blocks in Guam and CNMI that had no addresses in Census 2000 will be combined with larger neighboring blocks, thus giving each tabulation block with no Census 2000 addresses a chance of being selected for the 2008 Survey of Compact of Free Association (COFA) Migrants. Next, the blocks and combined blocks in each area will be assigned to one of four strata that we have labeled as “None”, “Low”, “Moderate”, and “High” based on the Census 2000 count of the number of people who were born in the Republic of the Marshall Islands, the Federated States of Micronesia, and the Republic of Palau. Blocks with a high level of growth for these populations between the 1990 Census and Census 2000 will have their Census 2000 totals adjusted appropriately to reflect projected continued growth between 2000 and 2008. The preliminary definitions of these strata are as follows:

Stratum	Census 2000 Count of People Born in the Republic of the Marshall Islands, the Federated States of Micronesia, and the Republic of Palau
None	0
Low	1-14
Moderate	15-29
High	30 or more

The counts of blocks and combined blocks for each stratum are as follows:

Stratum	Total Blocks and Combined Blocks in Guam	Total Blocks and Combined Blocks in CNMI
None	343	87
Low	378	136
Moderate	142	60
High	209	116

The blocks and combined blocks within each stratum will then be ordered based on the Census 2000 count of the number of people who were born in the Republic of the Marshall Islands, the Federated States of Micronesia, and the Republic of Palau. A representative sample of these ordered blocks and combined blocks will then be selected within each stratum using a higher sampling rate in the “High” stratum and a decreasing sampling rate for each subsequent stratum. The within-stratum sampling rates will be determined to result in an expected relative error of 15%. These rates are expected to result in an oversample of blocks and combined blocks from the “High” and “Moderate” strata, and an undersample from the “Low” and “None” strata. While the exact sampling rates are unknown at this time, we expect to select about 45 blocks and combined blocks in Guam and about 30 blocks and combined blocks in CNMI. We expect to identify about 3,300 sample addresses from the Guam sample blocks and about 2,000 sample addresses from the CNMI sample blocks.

The sample blocks and combined blocks will be canvassed by enumerators to identify every address currently in the blocks and combined blocks, and the occupants at each address will be interviewed using a questionnaire containing questions comparable to those used in the ACS and Census 2000 (see Attachment 4). The addresses in sample blocks and combined blocks that contain a large number of addresses will be subsampled and only the addresses in the subsample will be interviewed.

In addition, a reinterview sample will be selected from households completing the questionnaire in Guam and CNMI. Each enumerator in Guam and CNMI will have three cases from their workload randomly selected for reinterview. Households selected will be administered a paper questionnaire by supervisory enumerators to detect discrepancies or data falsification (see Attachment 5).

2. Procedures for Collection of Information and Estimation Procedures

Based on the sample sizes and sample designs for Guam and CNMI, we expect a 68 percent confidence interval relative error of + 15 percent for the estimates of the number of COFA migrants residing in each of these areas in 2008. This relative error is being targeted for these estimates so the reliability level will be comparable to what is expected for the estimate of COFA migrants residing in Hawaii that will be derived from data collected previously on the ACS.

In the majority of the sample blocks and combined blocks, residents in all housing units and group quarters addresses will be administered the paper questionnaire via personal visit. In very large blocks, the list of addresses will be subsampled to help defray the costs of enumeration. Only the housing units and group quarters addresses selected for the subsample will be administered the questionnaire via personal visit. The very large blocks that will need to be subsampled will be identified once the sample of blocks and combined blocks has been selected.

A maximum of three personal visits will be made to each sample address to complete the interview. Enumerators will attempt to collect telephone numbers during the first visit and when possible, conduct the follow-up interview attempts by telephone. In addition, a quality control reinterview will be conducted for approximately 400-500 sample cases.

The survey estimates for Guam and CNMI will be based on weighted data. The weights for producing these estimates will be calculated using a three-step procedure. First, the base weight (BW), which is the inverse of the probability of selecting a sample block or combined block, will be calculated and assigned to each sample address in the block or combined block. Next, the subsampling factor (SF), which is the inverse of the subsampling rate used in a subsampled block or combined block, will be calculated and assigned to each of the addresses selected for the subsample. Addresses in sample blocks and combined blocks that are not subsampled will be assigned a subsampling factor of 1.0. Finally, a nonresponse adjustment factor (NRF), which is the ratio of the occupied sample addresses divided by the interviewed occupied sample addresses, will be calculated and assigned to the interviewed occupied sample addresses. Separate nonresponse adjustment factors will be computed for the major geographic entities within Guam and CNMI. The weight for each interviewed address will be the product of these factors as expressed by the following formula:

Weight = BW * SF * NRF

3. Methods to Maximize Response

We will maintain high levels of data accuracy and response rates through interviewer instruction, professional training, and close monitoring of the data. We expect a response rate of approximately 98%, which is based on the response rate received for the American Community Survey in Hawaii in 2006 (97.8%).

Survey procedures in Guam and CNMI include the initial personal visit supplemented by up to two follow-up visits to addresses where no respondents were present at the initial visit. Staff will conduct telephone follow-up interviews when possible to help encourage responses to the survey and to obtain additional information or clarification about reported data.

4. Testing Procedures

The language of the Survey of Compact of Free Association (COFA) Migrants questionnaire is similar to the Census 2000 long-form questionnaire and the 2007 ACS questionnaire, reflecting questions that were tested from 2004 – 2006.

5. Contacts for Statistical Aspects and Data Collection

The Census Bureau will collect and process these data. Within the Census Bureau, the following individuals should be consulted for further information on their area of expertise.

Statistical Aspects:

Dennis Schwanz

Demographic Statistical Methods Division

U.S. Census Bureau

(301) 763-1984

[email protected]

Overall Data Collection:

Melissa Therrien

Population Division

U.S. Census Bureau

(301) 763-6065

[email protected]