a Cre _ D> Ire o fe pes ve wid AL.
jehey vt. 1967)
oO
312 KINSHIP AND CAREERS

the sibling variable and first job (W), namely, education (U). The
32.73 per cent of the total sum of squares in W accounted for by U
alone is increased only by 4 per cent to 33.25 by the addition of the
sibling variable. Thus most of the effect of family size and birth order
on first job is mediated by education. ,

Moving farther to the left in Figure 9.1 and taking education as the
dependent variable, we can ask how it is influenced by the sibling
classification. The relevant percentages of sum of squares are:

Combination of variables Percentage of total sum of squares (U)

Sibling 10.78
X (father’s occupation) 19.86
V (father’s education) 20.77
Xand V 27.94
X and Sibling 24.42
V and Sibling 25.24
X, V, and Sibling 30.71

The effect of the sibling variable is substantial, nearly AL per-cént
of the total sum of squares, and the causal diagram $tipulates that
this is a direct effect. Does the sibling classification entirely mediate
the effects of the two variables lying behind it? The answer is clearly
negative because each of these variables has a larger gross effect than
the sibling variable and because each produces a substantial incre-
ment over the sibling variable alone, as is shown by the combinations
“V and Sibling” and “X and Sibling.” Comparing the three single
variables with either of these two combinations makes it clear that
direct paths from both X and V to education are called for, notwith-
standing the sibling classification as an intervening variable. In other
words, social origins influence occupational achievement not only by
affecting family size and composition but also in ways that are inde-
pendent of these family characteristics. On the other hand, the sibling
variable clearly exerts an influence on education independent of that
exerted by the two background variables, accounting for an additional
2.77 per cent of the total sum of squares (30.71 — 27.94).

In sum an attempt has been made to place the multiple factor
referring to the family of orientation into the causal model depicting
the process of stratification that was developed earlier. Father's educa-
tion and occupation influence the kind of family he establishes, and
the characteristics of this family influence son’s educational attain-
ment. The impact of the parental family on occupational achievement
seems to be entirely a result of its effect on education. By the nature
of the sibling variable this model cannot distinguish between the in-
fluence of family size and of sibling position. It therefore cannot reveal

&

v,