Summary of the Sample Design for the Panel

ATTACHMENT 5.
SUMMARY OF THE SAMPLE DESIGN FOR THE PANEL

5-1

ATTACHMENT 5.
SUMMARY OF THE SAMPLE DESIGN FOR THE PANEL
First Stage: Primary Sampling Units (PSUs, One or More Counties)
Sampling Frame

All Counties in the 50 States and the District of Columbia

Stratification

Nine Census Divisions

Type of Selection

Probabilities Proportional to Size Measurea
PSUs Will Be Proportionately Allocated to the Nine Strata

Sample Size

80 PSUs

Second Stage: Census Block Groups (CBGs, One or More Adjacent Census Blocks)
Sampling Frame
CBGs in Sampled PSUs
Stratification

None or TBDb

Type of Selection

Probabilities Proportional to Size Measure

Sample Size

3 CBGs/PSUs

Sampling Frame

Stratification

None

Type of Selection

Equal Probability within Each CBG; Equally Allocating HUs to Each
CBG

Sample Size






Selected HUs: 36,390
Occupied HUs: 34,571 (95% occupied HU rate)
Screened HUs: 29,385 (85% screening rate)
Eligible HUs: 4,444 (15,2% HU eligibility rate )

Stratification

Fourth Stage: Tobacco Users Age 18+
Roster of Eligible Persons (adult tobacco users) in Screened and Eligible
Hus
Age Group (18–25, 26+) by SES Status (low SES, non-low SES)

Type of Selection

Stratified Random Sampling (Oversample 18–25 Years Old)

Sample Size

 Tobacco Users Selected: 4,444
 Recruited Tobacco Users: 4,000 (90% recruitment rate)

Sampling Frame

a

Third Stage: Housing Units (HUs)
Address-based Sampling (ABS) CBGs: Residential Mailing Addresses
Supplemented with the CHUMc
Field Enumerated CBGs: Listing Housing Units

Size measure will be the number of tobacco users in a PSU.

b

Stratification at the second selection stage can be socioeconomic status (SES) of CBGs (low SES CBG or non-low
SES CBG), or high or low tobacco use prevalence rates (CBGs with high-density tobacco use or CBGs with lowdensity tobacco use), will be determined with consultation of the FDA Project Officer.
c

Address-based sampling (ABS) will be used in 202 CBGs (84%) where the expected ABS coverage rate is 90% or
more. ABS and the Check for Housing Units Missed (CHUM) procedure will be used in 26 CBGs (11%) where the
expected ABS coverage is 50%–90%. Field enumeration will be used in 12 CBGs (5%) where the coverage is
estimated below 50%.

5-2