Table 4.
Strength of ability to explain variance in serum cotinine and hair cotinine for questionnaire and environmental measures of environmental tobacco smoke as indicated by reduction in Akaike Information Criteria (AIC), Cincinnati Asthma Prevention Study (2000–2003)
| Reduction in AICa |
||
|---|---|---|
| Serum cotinineb | Hair cotininec | |
| Hours in a room with smoking | 6 | 4 |
| Number of cigarettes/day | 16 | 9 |
| Maternal smoking | 21 | 1 |
| Paternal smoking | −2 | −2 |
| Parent report (4 questions above)d | 37 | 9 |
| Home sizee | 24 | 16 |
| Parent reportd+home sizee | 69 | 25 |
| Household nicotine | 91 | 13 |
| Staff perception of smokiness | 7 | 0 |
AIC values were subtracted from the AIC from a base linear mixed model adjusted for intervention status, study visit, and in the case of hair cotinine, African-American race.
From linear mixed models predicting log serum cotinine with an unstructured correlation matrix, adjusted for intervention status and study visit.
From linear mixed models predicting log hair cotinine with an exchangeable correlation matrix, adjusted for intervention status, study visit, and African-American race.
Parent report includes hours in a room with smoking (three categories as in Table 3), number of cigarettes/day (continuous), maternal smoking (yes/no), and paternal smoking (yes/no).
Home size was measured as home volume in serum cotinine models and number of rooms in hair cotinine models.