Re: Mea Culpa by Rinaldo Del Gallo, III Esq.
By Letters To The Editor David Pittelli (03/30/07)
Dear Mr. Del Gallo,
A number of points, starting with the trivial:
1) The difference between my calculated 6.97 and your calculated 6.83 is
merely the result of rounding errors in your calculation. So we can ignore
that.
2) The 30% national dropout rate figure (your “ASSUMPTION 2”) is
irrelevant to the question at hand (i.e., to what extent the fatherless
have a higher dropout rate than do children living with a father). Your
calculations will give exactly the same differential with different
national dropout rate figures, for any number between 0% and about 37%.
(Your calculations are easier to explain, but actually more complex, with
more variables, not simpler.)
3) In addition to the three Assumptions that your three pieces of data are
accurate, you are also assuming that each piece of data measures something
identical to that factor you need to perform your calculations. As I will
demonstrate in the rest of this letter, this is not the case.
4) You state “ASSUMPTION 1: 26% of children comes from fatherless homes.
Source: divorcemag.com.” First, that source has 27% as the sum of the
“mother only” and “no parent” figures, not 26%.
5) More importantly, with ASSUMPTION 1, you are falsely assuming that
their sample represents the same universe of children as does your sample.
But their universe is of all children aged 0-17, while your universe of
relevant children is clustered around the age of 17 (e.g., 16, 17, 18).
6) I have not seen any statistics on fatherless rates for children at
different ages, but it is evident that they are not fixed over time. Let
us imagine that, at birth, 10% of children are fatherless. And that over
the next 18 years, breakups, divorce and death bring the fatherless rate
of 17-year-olds to 50%. Given that the group of all children under age 18
contains approximately equal numbers of children at each age,
fatherlessness is in a simple, approximately linear relationship with
children’s age, and the reported statistics in this case will show that
the fatherless rate of all children is about 30% (the average of 10% and
50%).
7) Of course, infants, 7-year-olds, and even 12-year-olds essentially
never drop out of school. Most dropouts are 16 or 17, or even 18. So if we
look at an average age of a dropout as 17, and 50% of 17-year-olds are
fatherless, then we would expect (in the absence of a fatherlessness
effect) that 50% of dropouts would be from fatherless homes, not 26% or
30% or anything like that.
7B) Perhaps 50% is too high (or too low), but the 26% figure is certainly
way too low. And 50% is not implausible, considering that some kids are
fatherless from birth, and according to
http://www.divorcemag.com/statistics/statsUS.shtml
Percentage of married people who reach their 5th, 10th, and 15th
anniversaries:
5th: 82%
10th: 65%
15th: 52%
8) I do not know the methodology of the National Association of Secondary
School Principals, but I would expect that a boy who became fatherless at
age 16 or 17, and then dropped out at 17, would be counted as an example
of a dropout from a fatherless family, not as a dropout from an intact
family. So comparing apples to apples (or 17-year-old dropouts to the
universe of 17-year-olds), we would expect, in the universe of children
who might drop out, a baseline of about 50% -- not 26% -- to be from
fatherless homes.
9) Admittedly, 50% is less than the 71% apparently reported by the
National Association of Secondary School Principals, but only so much less
as to imply a fatherlessness effect of 2.45-fold, as I have previously
shown, not 7- or 9-fold.
10) That brings us to other factors making the 26% (or 27%) quoted average
fatherless rate not comparable to the fatherless rate expected among high
school dropouts, if fatherlessness itself has no effect.
11) The basic point of your initial letter was that fatherlessness swamps
race and class in terms of its connection with dropping out of school.
While I am prepared to believe that fatherhood may be as important, or
even somewhat more important, than race or class, one would have to do a
multivariate analysis to show that.
12) Further, there are deeper, philosophical questions to ask about how
one should do such an analysis, most having to do with one’s assumptions
about first causes, free will, and what one means by “all else being
equal.”
13) I agree that, as some other fathers’ rights writers (such as Daniel
Amneus) have pointed out, it is erroneous to use multivariate analysis to
factor out things like a child’s current income, or the poverty rate of
dropouts versus graduates, if there is some chance that the poverty itself
is caused by the fatherlessness. This is unquestionably the case, and
indeed poverty is, to Amneus, much of the mechanism by which
fatherlessness hurts children. So on this, he is right.
14) But one can still consider those socioeconomic factors which predated
parental breakup, such as parental education and income prior to that
point. If low-educated and low-income parents are more likely to have
children out of wedlock (as they are) and are more likely to break up and
divorce (as they are), and low-income and low-educated parents have
children who are more likely to drop out of high school even if these
parents stay together (as they are), then we have a situation where the
baseline percent we’d expect among dropouts isn’t 50%, as tentatively
expressed in #7 and #8 above, but rather higher.
15) As I wrote to you, “if within the socioeconomic classes for which
dropping out of high school is at all likely, the fatherless rate of
children aged 16-17 is 65%, then the fatherless are “only” 1.32 times as
likely to drop out as are those living with a father.”
16) For that matter, if within the socioeconomic classes for which
dropping out of high school is at all likely, the fatherless rate of
children aged 16-17 is 71%, then the fatherless are no more likely to drop
out than are those living with a father.
17) As it happens, I do not expect it is the case that fatherlessness has
absolutely no effect on dropout rates, but such a situation fits the sum
of available data as well as does your claimed 7-fold or 9-fold
difference.
David Pittelli
(Printer friendly version) Email: Letters To The Editor
