Table 1. Regression Model Resultsa Showing Relationship Between Weekly Area Code-Level 1–800-QUIT-NOW Call Volume and Weekly Market-Level Tips Campaign GRPs, February 4–July 8, 2012.
| Independent Variable | β Coefficient (P Valueb) [95% CI] | Elasticityc | Mean Market-Level Weekly GRP |
|---|---|---|---|
| 1-800-QUIT-NOW Television GRP (in 100s) | 89.3 (<.001) [72.9 to 105.8] | 0.219 | 33.8 |
| Smokefree.gov television GRP (in 100s) | 29.4 (<.001) [23.0 to 35.7] | 0.196 | 87.0 |
| Hispanic television GRP (in 100s) | 3.15 (.66) [−10.8 to 17.1] | 0.004 | 15.8 |
| Radio GRP (in 100s) | 17.4 (.02) [2.83 to 32.0] | 0.004 | 5.5 |
Abbreviations: CI, confidence interval; GRP, gross rating point.
a
Adjusted R 2 = 0.74.
b
Calculated from standard ordinary least squares regression t tests for each regression coefficient.
c
Elasticity is the percentage change in calls for a given percentage change in GRPs. Models control for weekly time trend, area code population size, percentage of media market population that is African American, percentage of media market population that is Hispanic, percentage of media market population that has bachelor’s degree or higher degree, media market median income, media market smoking prevalence, state fixed effects, and GRPs for Tips ads aired through separate state media buys.