supporting statement Part B.wpd

B. Statistical Methodology

1. Description of the Universe

There were over 3 million housing units enumerated in the 2000 Decennial Census for New York City. The bulk of the 18,000 units selected for the 2008 New York City Housing and Vacancy Survey originated from the 2000 Decennial Census address file for New York City, and were carried over from 2005. Remaining units will be selected from New York City Certificates of Occupancy and lists of units converted to residential use, lists of Single Room Occupancy (SRO) units, and lists of In Rem (city-owned) units. The survey portion that will measure units returning as residential dwellings consists of those that were classified as "lost from the housing inventory" from the 1999 survey and not returned.

The response rate for the 1993, 1996, and 2002 surveys was 98.0 percent. The response rate for the 1999 survey was 95.0 percent. For 2005 it was 96 percent. We expect that the response rate for the 2008 survey will fall within the 90 to 98 percent range.

2. Procedures for Collecting Information

In 2008, we will select the sample from the five sources as described in Section B.1 above. Each source is considered a sampling frame. Within these frames, we will select clusters of (generally) four housing units. Each housing unit will have an approximate 0.6 percent chance of selection, or about 1 in 167. The sample is designed to produce a standard error of the estimate of the "vacant available for rent" rate for the entire city of no more than one-fourth of 1 percent, if the actual vacancy rate is 3 percent or less.

The procedure employed to estimate housing unit and population characteristics will consist of three parts: (a) inflating the sample results by the reciprocal of the probability of selection, (b) adjusting the sample results to account for noninterviews, and (c) using independent estimates of housing units and population in order to control the estimates in a manner that increases precision.

3. Methods to Maximize Response

Enumerators will conduct interviews at all selected sample units. Since the first survey in 1962, personal interviewing has been found to be the best means of obtaining the information. A letter and NYCHVS fact sheet will be sent to each selected housing unit prior to the interview.

There will also be steps undertaken to assure the accuracy of the information we obtain. Any unit classified as vacant will be reinterviewed using Form H-108. Additionally, 5 percent of the occupied units will be rechecked using the H-108. This will be done to ensure the accuracy of the original report. Other enumerators trained specifically for reinterview will conduct this phase of the operation.

Finally, all units classified as "lost from the housing inventory" will be rechecked by professional staff. There is no burden for this recheck as it is an observation (drive-by) check.

4. Testing of Procedures

There will not be any pretesting for this survey. The questionnaire items and procedures are similar to those from previous NYCHVS surveys and other household surveys such as the American Housing Survey and the Current Population Survey.

5. Contacts for Statistical Aspects and Data Collection

The sample selection and weighting specifications for the survey will be devised by staff of the Longitudinal Surveys Branch, Demographic Statistical Methods Division, Bureau of the Census. The Branch Chief, Dennis Schwanz, can be reached on (301)763-1984.

All the information in the survey will be collected by a staff of enumerators under the auspices of the Field Division, Bureau of the Census. Henry Tow, Chief of the Housing and Health Surveys Team, will coordinate this project and can be reached on (301)763-4879.

Overall coordination for the survey will be provided by the Financial and Market Characteristics Branch, Housing and Household Economic Statistics Division. The Branch Chief, Howard Savage, can be reached on (301)763-3199.

LIST OF ATTACHMENTS

Attachment A - Survey forms, respondent letter, and fact sheet