Table 3. Genetic model-fitting results for cervical smear.
Model | −2LL | df | vs | χ2 | Δdf | P | AIC | a2 | c2 | e2 | |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | Full ACE | 2302.350 | 2688 | −3073.650 | 0.37 | 2.6−10 | 0.63 | ||||
2 | E model | 2315.051 | 2690 | 1 | 12.701 | 2 | 0.002 | −3064.949 | — | — | 1.00 |
3 | CE model | 2304.188 | 2689 | 1 | 1.838 | 1 | 0.175 | −3073.812 | — | 0.23 | 0.77 |
4 | AE model | 2302.350 | 2689 | 1 | 0 | 1 | 1.000 | −3075.650 | 0.37 | — | 0.63 |
Abbreviations: −2LL, −2 log likelihood; χ2, by substracting the −2LL of two different (nested) models yields a statistic that is asymptotically distributed as χ2distribution; AIC, Akaike's Information Criterion: a lower value of AIC indicates a more parsimonious model; df, degrees of freedom; Δdf, difference in degrees of freedom between two models; P, P-value; vs, tested versus model.
Cervical smear results were categorized in no abnormalities, mild abnormalities and severe abnormalities. Thresholds (=prevalences) were allowed to vary with age group (<40 or ≥40 years) and year of diagnosis (<1996 or ≥1996). a2, additive genetic influences; c2, common environmental influences; e2, unique environmental factors.