516
Le logiciel R - Effectuer des analyses statistiques
> legend("bottomright",c("SMOKE=0","SMOKE=1"),
+
col=c("red","blue"),lty=1)
●
●
● ●
●
● ●
●
●
●
●
●
●
●
●
●
● ●
● ●
● ●
●
●
●
●
●
●
● ●
●
●
●
●
● ●
● ●
●
●
●
● ●
● ●
●
●
●
●
●
●
●
●
●
●
● ●
●
● ● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
15
20
25
30
35
40
45
1000
2000
3000
4000
5000
BWT = β 0 + β 1 AGE + β 2 SMOKE + ε
AGE en années
Poids de naissance (g)
SMOKE=0
SMOKE=1
Figure 12.7 – Effet de l’ˆ age sur BWT dans un mod` ele sans interaction.
Le graphique ci-contre pr´ esente deux pentes diff´ erentes illustrant la notion
d’interaction.
> co <- coef(modele6); a0 <- co[1] ; a1 <- co[1]+co[3]
> b0 <- co[2]
# Effet de l’âge chez les non-fumeuses.
> b1 <- co[2]+co[4] # Effet de l’âge chez les fumeuses.
> plot(BWT∼AGE,xlab="AGE en années",ylab="Poids de naissance (g)",
+ main=expression(BWT∼"="∼beta[0]+beta[1] * AGE+beta[2] * SMOKE+
+
beta[3] * AGE∼"x"∼SMOKE+epsilon))
> points(AGE[SMOKE==1],BWT[SMOKE==1],col="red",pch=18)
> points(AGE[SMOKE==0],BWT[SMOKE==0],col="blue")
> abline(a=a0,b0,col="blue") ; abline(a=a1,b1,col="red")