According to the a priori hypotheses, an additional parameter is introduced in the age-period model, in order to allow for a lower case ascertainment in the oldest age group (persons aged 70 years or older) during the period 1970-79 (when this isn’t taken into account, the predictions can’t fit to the data).

The deviance is used to measure of the goodness of fit: it‘s the value of the likelihood ratio test statistic for the model fitted compared with the saturated model with a parameter for every observation unit; it‘s asymptotically distributed as a chi-square.

It is impossible to estimate the effects of each category of the factors: such models are always saturated. The age-period and age-cohort models are therefore examined. They are compared according to the Akaike Information Criterion (AIC), which takes into account the fact that the age-cohort model has more parameters than the age-period model.