Skip to main content
. 2006 Aug;60(8):712–720. doi: 10.1136/jech.2005.042275

Table 2 Summary statistics comparing goodness of fit for different cervical cancer maximum likelihood models.

Model Degrees of freedom Deviance p Value* Adjusted R2
Age 60 1805.2
Age drift ‡ 59 270.9 0.85
Age period 55 214.6 <0.001 0.87
Age cohort 44 111.0 <0.001 0.92
Age‐period‐cohort 40 68.5 0.94

*p Values are based on the F test for comparisons between two factor model with the full age‐period‐cohort model.18 †Adjusted R2 measures how much of the variability that is explained by factors other than age, taking into account the difference in the numbers of degrees of freedom.15 ‡The “drift” parameter represents a log‐linear change in rate not exclusively identifiable as a period or cohort effect.