Sunteți pe pagina 1din 14

Is Zero Preferred?

American Attitudes toward Childlessness in the 1970s


Author(s): Judith Blake
Source: Journal of Marriage and Family, Vol. 41, No. 2 (May, 1979), pp. 245-257
Published by: National Council on Family Relations
Stable URL: http://www.jstor.org/stable/351694
Accessed: 27-06-2016 04:24 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

Wiley, National Council on Family Relations are collaborating with JSTOR to digitize, preserve and
extend access to Journal of Marriage and Family

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
Is Zero Preferred? American Attitudes
Toward Childlessness in the 1970s*
JUDITH BLAKE**
University of California, Los Angeles

Is a large increase in voluntary childlessness likely to contribute to extremely low fer-


tility rates in the United States? To address this issue, this paper uses special ques-
tions on attitudes toward childlessness that were commissioned on a Gallup survey
of voting-age adults in February, 1977. A Likert-type scale of attitudes toward child-
lessness was constructed and major sociodemographic, plus some attitudinal, pre-
dictors of scale values were studied by means of Multiple Classification Analysis.
There is a high level of consensus that nonparenthood is not an advantaged status
and, although offspring are not regarded as economic investments, they are viewed
as being socially instrumental-they are not solely consumption goods. A desire for
some children is thus not as vulnerable to cost factors as one might think on the basis
of a consumer model of reproductive motivation. Less advantaged groups in the pop-
ulation are more likely than others to regard reproduction as a social investment,
thus helping to explain the frequently-discovered inverse relationship between socio-
economic status (including educational status) and either preferred or actual family
size. A major finding of our work is that men are significantly more likely to regard
childlessness as disadvantageous than are women.

Change in the proportion of childless married Depression and the baby-boom cohorts of
women has had an important influence on women.

American fertility rates during the past 50 Recently, the fertility decline following the
years. As Grabill et al., (1958:365-371) have baby boom has been accompanied by a sharp
shown, among demographic influences, large rise in nonparenthood among young married
variations in childlessness during the prime women. Despite a countervailing trend
reproductive ages contributed the most to the toward augmented rates of premarital preg-
difference between Depression and baby- nancy, marriage cohorts of the late 1960s
boom birth rates. The other factors consid- have shown marked increases in childlessness
ered were: larger numbers of women, more at durations of marriage as long as five years.
women marrying, and more births per For example, the marriage cohort of
mother. Even permanent childlessness (that 1965-1969 still numbered 23 percent childless
occurring to ever-married women by age after five years of wedlock compared to 13.6
40-44) has fluctuated widely, from a high of percent without children at the same duration
approximately 20 percent around 1950 to a of marriage among those wed during 1960-
low of only 7 percent 25 years later in 1975. 1964 (Spencer, 1979; U.S. Bureau of the
These percentages represent, respectively, the Census, 1974: Table 17).
Not only has actual childlessness among
*This research was supported by the National Institute young wives risen in recent years but, ac-
of Child Health and Human Development. The research cording to the surveys of the United States
was undertaken during 1976 and 1977 while the author
was on leave at International Population and Urban Re-
Bureau of the Census, there has been an in-
search Institute of International Studies, University of crease in the proportions who expect to re-
California, Berkeley. The author is grateful to Neil Berg main permanently without children. Whereas
for data processing and statistical assistance. in 1967 only 0.04 percent of wives aged 20-21
**School of Public Health and Department of Sociol- expected to be childless, by 1976 this figure
ogy, University of California, Los Angeles, California had risen to 4.0 percent. Comparable figures
90024. for wives aged 22-24 are 1.7 percent in 1967

May 1979 JOURNAL OF MARRIAGE AND THE FAMILY 245

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
and 5.3 percent in 1976 (U.S. Bureau of the prosperity, by providing improved occupa-
Census 1976:5, 1977:25). Among young tional slots for women, generates growing op-
single women in 1976 (those aged 18-24), 17 portunity costs for childbearing, while reces-
percent expected to be childless throughout sion reduces such costs. Although they think
their lives (U.S. Bureau of the Census, (1977:viii) that fertility rates "will not reach
1977:28). Unfortunately, no comparable data zero," they look for "a continuing secular fer-
for single women are available for earlier tility decline toward this asymptote punc-
years from the Current Population Survey. tuated by countercyclical fertility move-
Information concerning ideal family size ments."
commissioned on Gallup surveys shows, As a proponent of the thesis that raising
among both men and women under age 25, the costs of children, including opportunity
an increase in proportions saying that non- costs, will lower fertility, I would join in the
parenthood is ideal-from 1 percent for men expectation that augmented labor-force
and less than 1 percent for women in 1970 to participation will have this effect. The
3.7 percent among men and 2.5 percent question is how drastic will the effect be?
among women in early 1977. Here it is important to have some under-
Thus, although the proportions expecting standing of why one would expect a result as
or preferring to be permanently childless are extreme as is posited by Butz and Ward-a
still small, the increases in these proportions movement toward a zero fertility rate and a
have been large over the past decade. Such consequent major increase in childlessness.
increases, coupled with rising numbers of Such an expectation comes about, theoreti-
young women who remain childless well into cally, from the notion that children in
their married lives, lead us to question modern societies are primarily "consumption
whether further growth in voluntary child- goods." Consequently, the desire for them is
lessness is likely? Theoretically, have we highly sensitive to escalating costs. There is,
reason to believe that a rise in voluntary theoretically speaking, nothing to stop chil-
childlessness will be with us for some time to dren from being priced off the market like
come, contributing to ever lower levels of fer- any other commodity. This assumption that
tility in the United States? One argument children, in modern societies, are primarily
supporting such a thesis has been offered by (even solely) consumption goods is found
William Butz and Michael Ward (Butz and among all microeconomic analysts of repro-
Ward, 1977). According to Butz and Ward, duction in advanced societies (Cochrane,
in an extension of the "new home economics" 1975:373-390; Leibenstein, 1957:159-165;
theory of fertility, increases in female labor- Becker, 1960:209-240; Turchi, 1975:107-125;
force participation and wages have created Easterlin, 1966:131-153; Schultz, 1973,
unprecedented opportunity costs for women 1974). Along with the fact that control of
in having a family, and the more prosperous fertility is highly voluntary, this assumption
the economic period, the more this type of forms the basis for the application of con-
cost impinges. Thus, Butz and Ward believe, sumer economics to reproductive decision-
the United States is in a long-run condition of making. Yet the validity of the assumption is
counter-cyclical fertility trends in which never examined. Are children primarily con-
FIGURE 1. UTILITIES DERIVED FROM CHILDREN

1. Economic utilities-examples:
a. Child labor
b. Help in old age
2. Noneconomic utilities-examples:
a. Conferral of kinship statuses. One can attain the status of "parent" only by having children.
b. Creation of a customized personal group larger and more varied than a dyad. Family-size goals
may be affected by how long individuals want to preserve this family-type of group in their
lives.
c. Long-term extension of one's self, one's family, one's group-the immortality syndrome.
d. Performance of religious rituals after one's death that are believed necessary for salvation, fu-
ture incarnations, and the like.
e. "Psychic income" of various origins-children as household pets, children as objects of affec-
tion, children as givers of satisfaction, and so on.

246 JOURNAL OF MARRIAGE AND THE FAMILY May 1979

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
sumption goods in modern societies? What is sion-making regarding childlessness, because
the evidence? negative public views of the childless person
When we search through the microeco- may themselves constitute a pressure toward
nomic literature on this topic, we find that parenthood-a personal obstacle to be over-
children have been designated to be con- come in the decision to remain without
sumption goods by default. This is because children.
they have ceased to be economic investment
goods. The literature points out that couples DEVELOPING AN INSTRUMENT
would be ill-advised to expect economic gain TO MEASURE ATTITUDES
from children. Child labor is, by and large, il- TOWARD CHILDLESSNESS
legal, and help in old age (since it is almost Since, to our knowledge, no previous work
entirely voluntary) is apt to be sporadic and
had been done on measuring national atti-
slight. So, it is assumed that children, no tudes toward childlessness, it seemed wise to
longer being economic investments, have be-
begin with a familiar and widely understood
come "consumption goods."
scaling technique for attitudes and opinions.
However, not all investments are economic
These criteria are met by a Likert-type scale.
in nature. Indeed, a more systematic focus on
Likert scales are relatively simple to con-
the possible utilities derived from children,
struct, easy to use, and have reliability coeffi-
even in modern societies, makes us aware
cients that compare well with those of other
that children may not primarily be considered
types of instruments. In the case of scales
consumption goods. Among noneconomic
measuring attitudes toward childlessness, we
utilities, for instance, children may represent must wait for additional research on this sub-
social investments. A schematic representa-
tion of the theoretical utilities derived from
ject in order to measure reliability.
children is set out in Figure 1 (Blake, 1973: Sampling the Content Area of
279-288; see also Arnold et al., 1975; Hoff- Views About Childlessness
man and Hoffman, 1973:19-76). In searching for items for inclusion in a
Thus, barring the role of children as scale, it is naturally highly desirable to be as
economic investments (items la and ib), and wide-ranging as possible. Yet, in the early
even assuming little direct enjoyment and re- stages of scale construction, it seems almost
ward from offspring as pure consumer goods inevitable that a share of the spectrum of pos-
(items 2e), there are still, theoretically, cogent sible views will be missed. Nonetheless, a
reasons for parenthood that are uniquely in- major effort was made to be inclusive.
herent in the role of children as social invest- First of all, we distinguished between state-
ments; i. e., investments in parental status, in ments relating to the disadvantages of being
creating a family group of which one can be a childless and those pertaining to the advan-
member, in "immortality" or a long-term ex- tages. We surmised that there might be a def-
tension of oneself, in companionship or moral inite substantive difference between alleging
support in old age, or perhaps in some sort of that childlessness does not have disadvan-
posthumous remembrance of one's presence tages of a particular kind and claiming that it
on this earth (items 2a-2d). has positive advantages.
The present research reports on a survey Second, to tap the variety of alleged disad-
investigation of public views concerning vantages and advantages of childlessness, the
childlessness in the United States in Febru- growing popular literature on "nonparent-
ary, 1977. This investigation was conducted, hood" was studied carefully, since it contains
in large part, to determine whether or not a lively dialogue between two sets of propo-
children are, indeed, regarded as social in- nents: for parenthood and for nonparenthood
vestment goods in American society, and (Peck, 1971; Peck and Sendorowitz, 1974;
whether or not people believe that other types National Organization for Non-Parents,
of consumption goods are more attractive 1971-1976). This literature includes editorial
than children. This effort should help us to suggestions for handling a wide range of pro-
assess levels of social support for nonparent- natalist coercions that readers may experi-
hood as an option in the United States. Such ence, together with articles by nonparents de-
social support is important to individual deci- tailing the pressures they undergo and ex-

May 1979 JOURNAL OF MARRIAGE AND THE FAMILY 247

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
pressing their reasons for being childless. Ad- dents who are voluntarily childless on a
ditionally, rounds of letters by readers are in- permanent basis from others, this was not
cluded that elaborate on proddings by rela- possible in the present research. Such a dis-
tives and friends, and share the responses tinction would have involved us in detailed
that a nonparent can offer. We have also re- questioning concerning fecundity status, vol-
ferred to "advice" columnists (such as "Dear untary and involuntary causes of childless-
Abby" and Ann Landers), as well as to rele- ness, and so on, that would be inappropriate
vant articles in women's magazines. Finally, to a nationwide sample of voting-age adults.
we relied on theoretical expectations, already Since one purpose of the present study is to
discussed, concerning the utilities derived develop questions and scaling instruments
from and costs of children. that can be used for national fertility surveys
In all, five questions were asked concerning in the future, we are deferring comparisons of
the possible disadvantages of childlessness childless and nonchildless respondents to
and two concerning the possible advantages. such time as questioning on attitudes toward
The questions were prefaced by a statement childlessness can be included in a study de-
assuring respondents that there were neither voted exclusively to fertility.
right nor wrong answers to these items, but
that we were interested solely in respondents' Scoring
opinions and attitudes. Care was taken to al- All seven statements were coded so that a
ternate the "agree" and "disagree" responses value of one was assigned to the most strongly
so as to avoid a bias due to acquiescence. The antinatalist (prochildlessness) sentiment, and
questions are presented in Appendix 1. a value of five to the most pronatalist (anti-
childlessness) view. Inter-item correlations
Sampling the Population Used (Table 2) demonstrated that four of the ques-
,for Constructing the Scale tions on the disadvantages of childlessness
Because we want to build a widely applica- formed a homogeneous scale. Furthermore,
ble scale of attitudes toward childlessness, the questions on the advantages of childless-
our work has been developed on a random ness are intercorrelated. However, it is clear
sample of the American population of voting that these two dimensions (disadvantages and
age. The questions formed part of a battery of advantages) should not be combined into one
queries on childlessness and the one-child instrument-at least not one scored in the
family (26 questions in all) commissioned on manner of a Likert scale. We thus constructed
a Gallup survey of approximately 1,600 re- a four-item scale of attitudes toward child-
spondents in early February, 1977. Percen- lessness based on the disadvantages items.
tage distributions of the responses to the The final scale containing four items ranges
seven items for this sample are given in Table in value from 4 to 20. The higher values are
1. the most pronatalist (antichildlessness) re-
Although, for some purposes, it would be sponses. The mean value of the scale is 12.3
desirable to be able to distinguish respon- with a standard deviation of 3.1.

TABLE 1. ATTITUDES TOWARD CHILDLESSNESS: GALLUP SURVEY, FEBRUARY, 1977 (TOTAL SAM-
PLE)
% Strongly % Strongly % Don't
Agree % Agree % Undecided % Disagree Disagree Know Total
Disadvantages of
Childlessness

Empty lives 7 37 11 37 8 0 100


Unfulfilled 5 40 18 32 5 1 100
Divorce 4 30 17 44 6 0 100
Lonely 14 50 7 21 7 0 100
Financial 2 13 13 62 9 1 100
Advantages of
Childlessness

Best time 2 12 11 64 10 1 100


Intimacy 4 9 8 57 22 1 100

248 JOURNAL OF MARRIAGE AND THE FAMILY May 1979

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
Inter-Item Correlation THE ADVANTAGES AND
DISADVANTAGES OF
The inter-item Pearsonian correlation ma-
CHILDLESSNESS
trix for the seven items is shown in Table 2.
In judging the size of these coefficients, the Our inability, for methodological reasons,
reader should bear in mind that inter-item to combine questions on the advantages and
correlations seldom exceed 0.40. Indeed, Scot disadvantages of nonparenthood was based
(1968:258) speaks of coefficients of 0.33 as on a virtual unanimity among respondents
"unusually high," and notes that most psy- that childlessness is not advantageous. This
chological scales achieve coefficients of 0.2 or consensus regarding advantages stands in
less. marked contrast to the spread of responses
Turning first to the five items purporting to concerning the disadvantages of childless-
measure attitudes toward the disadvantages ness. We will now consider the separate re-
of childlessness, it is apparent that, except for sponses in more detail before looking at the
the item referring to financial stress in old scale results.
age, the inter-item coefficients are high. This Questions 1 and 5 in Appendix 1 related to
is particularly true since none of the questions the advantages of childlessness. The purpose
constituted a repetition (or rewording) of any of these questions was to tap whether or not
other question, and the sample was hetero- respondents believed that being childless of-
geneous. As can be seen in Table 2, the fered unique opportunities for highly de-
average correlation coefficient for "financial" sirable alternative forms of consumption
is the only one below 0.2. If "financial" is ("best times") and whether or not children
omitted from the matrix, then three of the were seen as interfering with the marital
inter-item correlation coefficients for the four bond. Clearly, despite the claims of the non-
other "disadvantages" items are more than parenthood literature, the childless are not
0.30, and one is almost 0.40. The lowest is seen as enjoying the most satisfying life-style,
then 0.27. nor are they seen as having the most satisfy-
By contrast, low inter-item correlations are ing marriages (Table 1). Being childless does
found between the questions relating to ad- not have a "glamour" image.
vantages and disadvantages (Table 2). How- As for the disadvantages of childlessness,
ever, the "advantages" items do appear to attitudes toward them were measured by
form the beginnings of a scale between them- questions 2-4, 6 and 7 in Appendix 1. It is
selves, as is evident from the fairly substantial worth noting that these questions are worded
inter-item coefficients found in Table 2. strongly so that agreement with them indi-
Again, the coefficients are rendered more im- cates fairly outspoken negative views of non-
portant by the fact that the questions are parenthood. By the same token, disagree-
nonrepetitive in subject matter. ment may reflect an unwillingness on the part

TABLE 2. INTER-ITEM PEARSONIAN CORRELATION MATRIX: SEVEN ITEMS FOR POSSIBLE INCLU-
SION IN LIKERT SCALE OF ATTITUDES TOWARD CHILDLESSNESS-GALLUP SURVEY, FEBRUARY,
1977

Disadvantages Advantages
Empty Lives Unfulfilled Divorce Lonely Financial Best Time Intimacy
Disadvantages
Empty lives .3968 .3838 .3771 .2537 .1475 .1388
Unfulfilled .3968 - .3276 .2277 .2604 .0338 .0591
Divorce .3838 .3276 - .2059 .2025 .0383 .0912
Lonely .3771 .2277 .2059 - .0770 .1502 .1489
Financial .2537 .2604 .2025 .0770 - .0244 -.0905
Mean Score .3528 .3031 .2800 .2219 .1984
Mean Score
Without Financial .3859 .3174 .3058 .2702
Advantages

Best
Intimacy - time- . .2962
.2962
Mean Score - .2793 .2530

May 1979 JOURNAL OF MARRIAGE AND THE FAMILY 249

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
TABLE 3. ATTITUDES TOWARD DISADVANTAGES AND ADVANTAGES OF CHILDLESSNESS BY AGE,
SEX, AND EDUCATIONAL LEVEL: GALLUP SURVEY, FEBRUARY, 1977

Total
% % %
U.S. Women Men

Agree Undecided Disagree Agree Undecided Disagree Agree Undecided Disagree


Empty Lives
44 11 45
Sex
Women 40 10 49
Men 48 11 40 - -
Age
Under 30 32 12 55 32 11 56 31 13 54
30-44 41 11 48 36 10 53 47 11 41
45+ 52 10 37 46 10 43 58 10 32
Education
Grade School 64 14 22 60 12 27 68 15 18
High School 45 11 44 41 11 48 50 10 39
College 34 10 55 29 9 61 38 11 50
Unfulfilled
45 18 37
Sex
Women 42 13 45
Men 48 22 29 - - -
Age
Under 30 38 21 40 35 14 50 40 28 31
30-44 42 16 42 38 12 49 47 18 34
45+ 51 17 32 48 13 39 53 21 24
Education
Grade School 60 21 18 62 18 21 60 24 16
High School 47 16 36 45 12 43 50 22 28
College 35 18 46 26 14 59 42 21 36
Divorce
33 17 50
Sex
Women 26 14 60 - -
Men 40 20 39 - - -
Age
Under 30 28 16 55 20 15 65 36 18 45
30-44 31 15 54 24 10 67 40 20 40
45+ 37 18 44 31 16 53 43 21 35
Education
Grade School 45 23 32 47 17 36 44 28 28
High School 32 15 52 26 10 63 40 21 38
College 30 17 52 16 20 62 41 14 45
Lonely
65 7 28
Sex
Women 64 6 30 - -
Men 66 7 26 - - -
Age
Under 30 60 9 30 63 7 30 59 11 29
30-44 59 7 34 56 6 37 62 7 30
45+ 71 6 23 69 6 24 72 5 22
Education
Grade School 72 7 21 69 8 23 74 5 20
High School 67 6 27 66 6 29 68 7 24
College 58 8 33 57 8 35 60 8 32
Best Time
14 11 74
Sex
Women 15 8 77 - -
Men 13 14 72

(Continued on next page.)

250 JOURNAL OF MARRIAGE AND THE FAMILY May 1979

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
TABLE 3 (CONTINUED)

Total
% % %
U.S. Women Men

Agree Undecided Disagree Agree Undecided Disagree Agree Undecided Disagree


Best Time (Continued)

Under 30 19 15 65 22 9 68 16 21 62
30-44 13 10 77 9 8 82 17 12 71
45+ 11 10 78 14 7 79 9 12 78
Education
Grade School 16 14 70 17 14 70 16 14 69
High School 14 11 74 15 8 77 14 15 70
College 12 10 77 13 7 79 11 12 76
Intimacy
13 8 79
Sex
Women 13 7 80
Men 13 9 77 - - - -
Age
Under 30 14 12 74 17 9 74 12 14 73
30-44 15 5 79 12 4 84 18 7 74
45+ 11 7 81 11 8 81 11 7 82
Education
Grade School 12 13 75 12 14 73 11 12 76
High School 13 8 79 13 7 80 12 9 77
College 14 6 80 13 4 82 15 6 78

of some respondents to subscribe to such cat- is particularly large between college-educated


egorical statements about other people, men and women, because there is a sharp ed-
rather than a feeling that childlessness is ucational gradient among women but not
without disadvantages. Insofar as this is true, among men. In fact, for all four questions,
our research underestimates negative views of the sex difference is greatest between the col-
nonparenthood. It seems most unlikely that lege-educated, indicating that "conscious-
we are overestimating popular disapproval on ness-raising" on this subject is greater among
this subject. college-educated women than among college-
As the marginals for "financial" make educated men. However, for both sexes, an
clear (Table 1), most respondents explicitly educational gradient exists. The educational
deny that progeny are a long-term economic spread is greatest for "empty lives" and
investment. Are they an important social in- "unfulfilled," indicating that, for lower-edu-
vestment? There is substantial agreement cated respondents particularly, children are
that children are a hedge against loneliness in instrumental in giving life meaning and pro-
old age. A very small proportion are unde- viding women with an important status.
cided on this point. There is a range of re- In effect, these data suggest that for a clear
sponses concerning whether or not offspring majority of respondents, regardless of age,
are of instrumental value (social investments) sex, or education, children are a social invest-
in giving life meaning, in providing fulfill- ment against loneliness in old age. For high
ment for women, and in cementing mar- proportions of respondents with less than a
riages. college education, children are seen as having
Who is most likely to think of childlessness additional social investment value-for pro-
as having the specific disadvantages studied viding meaning in life, for giving women a
here? On a bivariate level, we see from Table status without which they would be unful-
3 that men are more likely to have this view filled, and for cementing marriages.
than are women. The sex difference is mar- These examples thus lend credence to the
ginal with regard to three of the items, but is claim that the reproductive decision-making
fairly large when it comes to the belief that of modern couples is not entirely, or even pri-
marriages without children are more likely to marily, a "consumption-goods" decision,
end in divorce. The sex difference on this and, hence, that the consumer theory of re-
topic is maintained among all age groups and production is misspecified. Awareness of this

May 1979 JOURNAL OF MARRIAGE AND THE FAMILY 251

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
TABLE 4. RELATIONSHIP BETWEEN SOCIODEMOGRAPHIC VARIABLES AND THE SCALE OF ATTI-
TUDES TOWARD CHILDLESSNESS: GALLUP SURVEY, FEBRUARY, 1977
Class Deviations from Adjusted Deviations from
Predictor* N Mean** Grand Mean Mean Grand Mean

Age
Under 30 395 11.6 -0.7 11.9 -0.4
30-39 306 11.8 -0.6 12.0 -0.3
40-49 221 12.6 0.2 12.7 0.3
50-59 269 12.8 0.4 12.6 0.3
60+ 355 13.2 0.8 12.6 0.3
Sex
Men 758 12.8 0.5 12.9 0.5
Women 788 11.9 -0.4 11.8 -0.5
Race
White 1370 12.3 0.0 12.3 0.0
Black 166 12.7 0.4 12.6 0.3

Religion
Catholic 449 12.4 0.0 12.6 0.2
Protestant 950 12.5 0.2 12.3 0.0
Jewish 30 12.0 -0.3 12.7 0.4
Other 20 12.6 0.3 12.8 0.5
None 97 10.8 -1.5 11.3 -1.1
Education
0-8 202 13.8 1.5 13.2 0.8
H.S. Incomplete 244 12.9 0.5 12.8 0.5
H.S. Completed 609 12.3 -0.1 12.3 0.0
College Incomplete 262 11.8 -0.5 12.1 -0.2
College 4+ 229 11.3 -1.1 11.5 -0.8
Marital Status
Married and Separated 1158 12.5 0.1 12.4 0.1
Widowed 129 13.4 1.0 13.1 0.8
Single 193 11.4 -1.0 11.8 -0.5
Divorced 66 11.1 -1.2 11.3 -1.1

Family Income
Under $5,000 210 12.9 0.6 12.4 0.0
$5,000-$9,999 338 12.8 0.5 12.7 0.3
$10,000-$14,999 335 12.3 0.0 12.4 0.0
$15,000-$19,999 272 11.8 -0.6 12.0 -0.4
$20,000-$24,999 134 12.0 -0.3 12.3 0.0
$25,000+ 168 12.4 0.0 12.6 0.3
Community Size
Under 10,000 479 13.0 0.6 12.7 0.4
10,000-50,000 188 12.4 0.0 12.4 0.0
50,000-250,000 280 12.0 -0.3 12.0 -0.4
250,000+ 599 12.0 -0.3 12.2 -0.1
Region
East 410 12.1 -0.3 12.1 -0.2
Midwest and Rocky Mountain 492 12.2 -0.1 12.3 -0.1
South 426 12.9 0.6 12.7 0.3
Pacific 218 11.9 -0.4 12.3 -0.1
Grand Mean 12.3
Standard Deviation 3.1

*F-ratios are not shown in this table because, in the case of MCA, it is inappropriate to compute them for
each predictor based on the adjusted sum of squares.
**Figures in this table have been rounded independently.

commitment to children as a social invest- why poor people want any children, much less
ment helps to explain a number of anomalies why they typically want larger families than
that consumer theories of reproduction leave the well-to-do.
untouched. For example, the consumer Additionally, these results lead us to ques-
theory of fertility has no ready explanation for tion whether or not reproductive levels are as

252 JOURNAL OF MARRIAGE AND THE FAMILY May 1979

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
drastically vulnerable to the opportunity- iate relationships when controls are intro-
costs women are currently experiencing as duced. We may also ask how much of the var-
analysts like Butz imply. If children still rep- iation in scale scores is accounted for by
resent important and unique types of social sociodemographic and attitudinal variables
investment, especially for some sectors of the and, finally, which of the variables provides
population, fertility rates may resist sinking the most explanation.
to levels that would imply high rates of child- To answer these questions we have used
lessness. Multiple Classification Analysis (MCA)
The scale scores (unadjusted means) for a which is an elaboration of "dummy variable"
range of major sociodemographic variables multiple regression analysis (cf Andrews et
are shown in Table 4. Here we see that the less- al.: 1973). This method allows for categorical
educated, the middle-aged and old, the reli- variables as well as numerical values. The
gious, the married and widowed, those with method makes no assumptions about linear-
low incomes, Southerners, and those residing ity and is capable of demonstrating possible
in places of less than 100,000 population are nonlinearities as a normal feature of its oper-
the most inclined to view children as a social ation. However, MCA does assume a high de-
investment. Interestingly, men more than gree of additivity. Data sets in which there are
women are prone to see the disadvantages of major interactions are not appropriate for use
childlessness. Table 5 presents scores for two with MCA unless it is possible to construct in-
additional variables-belief in God and a teraction variables. To detect interactions
scale of social striving. The results according (and other anomalies with the data), AID
to religious orthodoxy are as was expected- (Automatic Interaction Detector), a compan-
the more orthodox the respondent, the more ion program to MCA, was first run on the
pronatalist he/she is. The social striving re- survey (cf Sonquist et al.: 1973; Sonquist,
sults might seem surprising since the strivers
1970). The AID results indicated very little
interaction. Test MCA runs demonstrated
are the most pronatalist. An interpretation of
this result is in the following section. that the relationships were effectively additive
and, therefore, that an MCA analysis could
THE PREDICTORS OF ATTITUDES be used without further adjustments.1
TOWARD CHILDLESSNESS:
A MULTIVARIATE ANALYSIS 'AID demonstrates the existence of major interactions
but it provides no objective basis for evaluating the seri-
We will now see what happens to our bivar- ousness of the interactions for MCA analysis. To test for

TABLE 5. ATTITUDES TOWARD CHILDLESSNESS: MEAN SCALE SCORES ACCORDING TO BELIEF IN


GOD AND ATTITUDES TOWARD STATUS STRIVING: GALLUP SURVEY, FEBRUARY, 1977
Mean Scale
Scores

Belief in God

Which one of the statements on this card comes closest to expressing what you believe
about God?

1. I know God really exists and I have no doubts about it. 12.6 (1066)
2. While I have doubts, I feel that I do believe in God. 12.1 ( 253)
3. I find myself believing in God some of the time, but not at other times. 12.1 ( 57)
4. I don't believe in a personal God, but I do believe in a higher power of some kind. 11.3 ( 99)
5. I don't know whether there is a God and I don't believe there is any way to find out. 11.2 ( 45)
6. I don't believe in God. 11.0 ( 25)
Social Striving
Would you agree or disagree that raising one's social position is one of the most important
goals in life?

1. Strongly agree 13.0 ( 335)


2. Agree
3. Undecided 12.9 ( 95)
4. Disagree 12.1 ( 904)
5. Strongly disagree 11.9 ( 229)
Total* 12.3 (1570)

*Categories do not add up to total because of "Don't Know" and "Other" responses.

May 1979 JOURNAL OF MARRIAGE AND THE FAMILY 253

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
TABLE 6. RELATIONSHIP BETWEEN SCALE OF ATTITUDES TOWARD CHILDLESSNESS AND BELIEF
IN GOD AND SOCIAL STRIVING, CONTROLLING FOR SOCIODEMOGRAPHIC VARIABLES: GALLUP
SURVEY, FEBRUARY, 1977
Class Deviations from Adjusted Deviations from
Predictor N Mean Grand Mean Mean* Grand Mean

Social Striving_
Strongly Agree and Agree 326 13.0 0.6 12.8 0.5
Undecided 94 12.9 0.6 12.6 0.3
Disagree 894 12.1 -0.2 12.2 -0.2
Strongly disagree 225 11.9 -0.4 12.2 -0.2
Belief in God
No doubt 1048 12.6 0.3 12.5 0.1
Have doubts 251 12.1 -0.2 12.3 -0.1
Some of time 56 12.0 -0.3 12.4 0.0
Higher power 99 11.3 -1.0 11.7 -0.6
Don't know 45 11.2 -1.1 11.7 -0.7
Don't believe 23 10.9 --1.4 11.7 -0.6
Grand Mean 12.3
Standard Deviation 3.1

*Age, sex, race, religion, education, marital status, family income, community size, and region are held con-
stant.

In addition to the unadjusted means, Table Some original bivariate relationships


4 gives adjusted means. The adjusted means change in interesting ways when other vari-
for any single variable control for the effects ables are introduced. The gap between the
of the other variables in the analysis. The ad- sexes becomes even wider-with men receiv-
justed deviations from the grand mean are ing somewhat higher adjusted scores and
presented, as well, in order to clarify the na- women somewhat lower ones. The enhance-
ture of the relationships. ment of these sex differences after other vari-
We see, from a comparison of the unad- ables are controlled may indicate a greater
justed and adjusted means (Table 4) that, in commitment to having a family on the part of
general, the bivariate relationships are men. On the other hand, it may simply
attenuated when other variables are intro- demonstrate that men are less likely to bridle
duced as controls. The differences between at categorically stated pronatalism than are
whites and blacks, the South and other women at this point in time. This is parti-
regions, the married/separated/widowed and cularly true since, on balance, the questions
the divorced/single, young and old, rural and relate more to women's status and bargaining
urban, those with more and less education, power than to men's. In the 1970s, many
and rich and poor are all lessened when the women may resent an overt pronouncement
effects of the other variables are taken into that they would be "unfulfilled" without chil-
account. However, in most cases, fairly sharp dren, or that their marriages will hold to-
differences remain, and the nature of the re- gether only if children provide the glue.
lationships does not change. For example, A further change which occurs when the
even when age and other possibly confound- data are adjusted is that the pattern of scores
ing variables are controlled, the more-edu- for religious groups shifts markedly. Control
cated are considerably less pronatalist than for sociodemographic variables shows Cath-
those with less schooling. And, conversely, olics and Jews to be more pronatalist and
even with education and other variables held Protestants to be less so than was indicated by
constant, the young are decidedly less pro- the bivariate distributions. In effect, the
natalist than the old.
sociodemographic characteristics associated
with each religious affiliation suppress and
the effects of interactions, the researcher must run two distort the differentials in scale scores. This
MCA analyses on the data set under the respective as- type of effect has been found before in
sumptions of additivity and nonadditivity. When this was American fertility research on religious dif-
done, we found that the adjusted R2 for the additive
model was actually higher than the adjusted R2 for the ferences (cf. Freedman et al., 1961:608-614).
interaction model, indicating that the latter assumption What of the two attitudinal variables-be-
was inappropriate. lief in God and social striving? Even when

254 JOURNAL OF MARRIAGE AND THE FAMILY May 1979

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
TABLE 7. MARGINAL EFFECT OF SOCIODEMOGRAPHIC AND ATTITUDINAL VARIABLES ON SCALE
OF ATTITUDES TOWARD CHILDLESSNESS

Adjusted R2 Percentage Gain in R2 from Adding Deleted Variables*


Full model** .130
Less: Sex .104 24
Education .114 14
Marital Status .119 9
Family Income .122 6
Age .123 6
Religion .123 5
Community Size .123 5
Region .127 2
Race .130 0
Full model plus Belief in God .131 1
Full model plus Social Striving .134 4
*The figures in this column indicate the percentage gain in R2 of the full model compared with the R2 of the
model with the designated variable deleted. Thus, the gain in R2 between the full model and the one with sex de-
leted is .026 or 24 percent.
**Age, sex, race, religion, education, marital status, family income, community size, and region.

other variables are held constant, believers sensus that nonparenthood is not an advan-
remain more pronatalist than doubters or un- taged status and, although offspring are not
believers. However, the relationship is regarded as economic investments, they are
weakened when other variables are con- viewed as being socially instrumental-not
trolled. As for social striving, Table 6 shows solely as consumption goods. A desire for
that where sociodemographic variables are children, thus, is not as vulnerable to cost
held constant, the relationship between factors as one might think on the basis of a
social-striving attitudes and pronatalism is consumer model of reproductive motivation.
attenuated but not changed. It is still true Indeed, our results suggest that the applica-
that those most committed to social striving tion of consumer theory to reproductive be-
are the most pronatalist. They do not view havior is inappropriate since children are im-
children as a hindrance in getting ahead. portant social investments, and the degree to
They want both: to get ahead and to have which this is the case varies among social
children. This result gives added weight to strata. Less advantaged groups in the popula-
our contention that children are regarded as tion are more likely than others to regard re-
social investments rather than as consump- production as a social investment, thus help-
tion goods to be readily sacrificed when the ing to explain the frequently-discovered in-
opportunity costs escalate. verse relationship between socioeconomic
Which of the sociodemographic variables status (including educational status) and
exerts the greatest effect on values of the either preferred or actual family size. In
childlessness scale when the other variables effect, persons of lesser socioeconomic status
are held constant? From Table 7 we see that desire and have more children (in spite of
sex, education, and marital status are the lesser income) not only because they are
most influential. Except for race, which has willing, on the average, to settle for lower
no effect, the remaining background vari- quality offspring, but because children are
ables have similar, relatively small amounts perceived as having more of a social invest-
of influence. We can also see that, after ment value than is the case at higher socio-
taking into account the sociodemographic economic levels.
influences, the marginal effect of the atti- A major finding of our work is that men are
tudinal variables is not great. significantly more likely to regard childless-
ness as disadvantageous than are women. If
CONCLUSION this result is not an artifact of questioning
techniques, it raises interesting issues. Since
In the future, is a large increase in volun- the opportunity costs in childbearing still im-
tary childlessness likely to contribute to ex- pinge more heavily on women than on men,
tremely low fertility rates in the United States? women are, as Butz and Ward suggest, vul-
To judge by the results reported on here, the nerable to an escalation of such costs.
answer is "no." There is a high level of con- However, one cannot conclude that the only

May 1979 JOURNAL OF MARRIAGE AND THE FAMILY 255

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
result of a change in women's job oppor- tials in fertility." American Sociological Review
tunities will be a mammoth increase in child- 26 (August):608-614.
lessness. The fact that men of all strata view Grabill, Wilson H., Clyde V. Kiser, and Pascal K.
Whelpton
nonparenthood as distinctly disadvantageous 1958 The Fertility of American Women. New York:
suggests that they may be willing to help Wiley.
women lower the opportunity costs of child- Hoffman, L. W., and M. L. Hoffman
bearing and rearing in order to have a family 1973 "The value of children to parents." Pp. 19-76
-a small family, perhaps, but a family none- in James T. Fawcett (Ed.) Psychological Per-
spectives on Population. New York:Basic
theless. Equal sharing of household and Books.
childrearing responsibilities between mates Leibenstein, Harvey
could have the effect of mitigating many op- 1957 Economic Backwardness and Economic
portunity costs for women and allowing Growth. New York:Wiley.
couples to experience parenthood if they wish National Organization for Non-Parents
1971- Newsletter. Baltimore:Author.
to do so. A greater egalitarianism between
1976
mates in the United States, as contrasted with
Peck, Ellen
many other low fertility countries, could thus 1971 The Baby Trap. New York:Bernard Geiss As-
prevent an extreme escalation of nonparent- sociates.
hood here. Peck, Ellen, and J. Sendorowitz (Eds.)
1974 Pronatalism: The Myth of Mom and Apple Pie.
New York:Crowell.
REFERENCES Schultz, Theodore (Ed.)
1973 Journal of Political Economy 81 (March/April):
Andrews, Frank M., James N. Morgan, John A. Son- Passim.
quist, and Laura Klem 1974 Journal of Political Economy 82 (March/April):
1973 Multiple Classification Analysis (2nd ed.). Ann Passim.
Arbor:University of Michigan. Scott, W. A.
Arnold, Fred, Rodolfo A. Bulatao, Chalio Buripakdi, 1968 "Attitude measurement." Pp. 204-273 in Gard-
Betty J. Chung, James T. Fawcett, Toshio Iritani, Sung ner Lindzey and Elliot Aronson (Eds.), Hand-
Jin Lee, and Tsong-Shein Wu book of Social Psychology (Vol. 2). Reading,
1975 The Value of Children: A Cross-National Study Massachusetts:Addison-Wesley.
(Vol. 1) Introduction and Comparative Analy- Sonquist, J. A.
sis. Honolulu:East-West Population Institute, 1970 Multivariate Model Building. Ann Arbor:Sur-
East-West Center. vey Research Center, University of Michigan.
Becker, G. Sonquist, J. A., Elizabeth L. Baker, and James N. Mor-
1960 "An economic analysis of fertility." Pp. 209- gan
231 in National Bureau of Economic Research, 1973 Searching for Structure. Ann Arbor:Survey Re-
Demographic and Economic Change in Devel- search Center, University of Michigan.
oped Countries. Princeton, New Jersey:Prince- Spencer, Gregory
ton University Press. 1979 "Childlessness and one-child fertility: A com-
Blake, J. parative analysis." Unpublished doctoral dis-
1973 "Fertility control and the problem of voluntar- sertation, University of California, Berkeley.
ism." Pp. 279-288 in Pugwash Committee Turchi, B.
(Eds.), Scientists and World Affairs. London: 1975 "Microeconomic theories of fertility: A cri-
Proceedings of the Twenty-Second Pugwash tique." Social Forces 54 (September): 107-125.
Conference on Science and World Affairs, Sep- United States Bureau of the Census
tember. 1974 "Fertility histories and birth expectations of
Butz, William P., and Michael Ward American women: June 1971." Current Popu-
1977 The Emergence of Countercyclical U.S. Fertil- lation Report, Series P-20, No. 263. Washing-
ity. Monograph No. R-1605-NIH. Santa Mon- ton, D.C.:U.S. Government Printing Office.
ica:The Rand Corporation. 1976a "Fertility history and prospects of American
Cochrane, S. H. women:June 1975." Current Population Re-
1975 "Children as by-products, investment goods port, Series P-20, No. 288. Washington, D.C.:
and consumer goods: A review of some micro- U.S. Government Printing Office.
economic models of fertility." Population Stud- 1976b "Fertility of American women:June 1975."
ies 29 (November):373-390. Current Population Report, Series P-20, No.
Easterlin, R. 301. Washington, D.C.:U.S. Government
1966 "On the relation of economic factors to recent Printing Office.
and projected fertility changes." Demography 1977 "Fertility of American women:June 1976."
3:131-153. Current Population Report, Series P-20, No.
Freedman, R., P. K. Whelpton, and J. W. Smit 308. Washington, D.C.:U.S. Government
1961 "Socioeconomic factors in religious differen- Printing Office.

256 JOURNAL OF MARRIAGE AND THE FAMILY May 1979

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms
APPENDIX 1. TEXT OF QUESTIONS ON ATTITUDES TOWARD CHILDLESSNESS
(The identifying words to the left are used in the tables in this paper.)
Intimacy 1. Having a child gets in the way of the closeness and intimacy of a couple's relationship.
a. Strongly agree
b. Agree
c. Undecided
d. Disagree
e. Strongly disagree
Lonely 2. People who are childless are more likely to be lonely in their older years than persons who
have had children.
a. Strongly disagree
b. Disagree
c. Undecided
d. Agree
e. Strongly agree
Empty Lives 3. Childless couples are more likely to lead empty lives than couples with children.
a. Strongly disagree
b. Disagree
c. Undecided
d. Agree
e. Strongly agree
Financial 4. If you have never had children, you are more likely to have a hard time financially when you
are older.
a. Strongly agree
b. Agree
c. Undecided
d. Disagree
e. Strongly disagree
Best Times 5. It seems to me that childless couples are the ones who are having the best time in life.
a. Strongly agree
b. Agree
c. Undecided
d. Disagree
e. Strongly disagree
Divorce 6. Childless marriages are more likely to end in divorce than are marriages where there are chil-
dren.
a. Strongly disagree
b. Disagree
c. Undecided
d. Agree
e. Strongly agree
Unfulfilled 7. A woman is likely to feel unfulfilled unless she becomes a mother.
a. Strongly agree
b. Agree
c. Undecided
d. Disagree
e. Strongly disagree

May 1979 JOURNAL OF MARRIAGE AND THE FAMILY 257

This content downloaded from 128.197.26.12 on Mon, 27 Jun 2016 04:24:03 UTC
All use subject to http://about.jstor.org/terms

S-ar putea să vă placă și