Sei sulla pagina 1di 15

Nonfamily Living and the Erosion of Traditional Family Orientations Among Young

Adults
Author(s): Linda J. Waite, Frances Kobrin Goldscheider and Christina Witsberger
Source: American Sociological Review, Vol. 51, No. 4 (Aug., 1986), pp. 541-554
Published by: American Sociological Association
Stable URL: http://www.jstor.org/stable/2095586
Accessed: 15-04-2016 17:54 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.

Sage Publications, Inc., American Sociological Association are collaborating with JSTOR to digitize,
preserve and extend access to American Sociological Review

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

NONFAMILY LIVING AND THE EROSION OF


TRADITIONAL FAMILY ORIENTATIONS
AMONG YOUNG ADULTS*
LINDA J. WAITE FRANCES KOBRIN GOLDSCHEIDER
The Rand Corporation Brown University

CHRISTINA WITSBERGER
The Rand Corporation

Young adults in recent cohorts have been leaving the parental home earlier and marrying
later now than they did several decades ago, resulting in an increased period of independent
living. This paper explores the consequences of time spent in non-family living, using data
from the National Longitudinal Surveys of Young Men and Young Women. We expect that
experience in living away from home prior to marriage will cause young adults to change
their attitudes, values, plans, and expectations, and move them away from a traditional
family orientation. We find strong support for this hypothesis for young women; those who
lived independently became more likely to plan for employment, lowered their expected
family size, became more accepting of employment of mothers, and more non-traditional on
sex roles in the family than those who lived with their parents. Non-family living had much
weaker effects on young men in the few tests that we could perform for them. The paper also
addresses the conditions under which living away increases individualism, and it discusses

the implications of these findings.

Beginning with the 1950s, the living arrangements

With these changes have come substantial

of American adults have been undergoing major

changes in attitudes toward family life. For

changes. More and more unmarried adults are

example, recent surveys show increasing accep-

living apart from their immediate families, forming

tance of divorce, permanent singleness and child-

non-family households' at a rate that far exceeds

lessness (Thornton and Freedman, 1982). Changes

the growth in the adult population. During the


1970s, non-family households increased 73 percent
while family households increased only 13 percent
(U.S. Bureau of the Census, 1980). This development can be seen as a further extension of the

in attitudes are normally explained by changes in


the social structure. However, it is also possible

that these effects are reinforced for individuals by

their own experiences, causing further shifts in


their family-related attitudes and behavior. Life

changes accompanying modernization that have


resulted in a decline in the centrality of the family.

course theory predicts that this would be particu-

Many have argued that this process may have


accelerated over the past several decades, as

childhood or early adulthood. Other things equal,

indicated by a series of demographic changes,


including rising ages at marriage, falling propor-

larly likely for experiences early in life-in

parental divorce increases the likelihood of divorce


for children (Pope and Mueller, 1976), teenage

tions of the population ever marrying, declining

parenthood is more common among daughters of

fertility and particularly marital fertility, the


sizeable proportion of marriages expected to end in

number of siblings positively affects own actual

teenage mothers (Baldwin and Cain, 1980),

divorce, and the large numbers of children living


with only one parent for at least part of their

and expected family size (Terhune, 1974; Waite and

childhood (Westoff, 1978; Davis, 1982; Espenshade,

intact families are less likely to marry at all but

1985).

very young ages (Kobrin and Waite, 1984).

Stolzenberg, 1976) and children raised in non-

In this context, perhaps the most critical


characteristic of the rapid increase in non-family
* Please address all correspondence to Linda J. Waite,
Senior Sociologist, The Rand Corporation, 1700 Main
Street, P.O. Box 2138, Santa Monica, CA 90406-2138.
The research reported here was supported by Contract

households is its concentration among very young


adults. While non-family households increased
their share of U. S. households from 15 to 22

No. NOI-HD-12814 from the Center for Population

percent between 1960 and 1975 over all ages, the

Research, National Institute of Child Health and Human

comparable increase for households with heads age

Development. We would like to thank Lee Lillard, Julie

25 or less was from 13 to 30 percent (Frey and

DaVanzo, and the members of the Rand Population


Research Center Workshop for helpful comments.

l Non-family households consist of an individual


living alone or with unrelated persons.

Kobrin, 1982). Young adults are leaving the

parental home substantially earlier now than they


did in the past. In Rhode Island, for example, the

American Sociological Review, 1986, Vol. 51 (August:541-554) 541

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

542 AMERICAN SOCIOLOGICAL REVIEW


percentage of sons still at home at age 26 dropped
from 50 percent for those who had reached that age
between 1930 and 1950, to 20 percent for those

outside the family can induce changes in these


family-derived values and tastes. Mason (1974)
has argued that a "role hiatus" between the

comparable decline for daughters was from 37 to

traditional role of daughter and the traditional roles


of wife and mother gives young women an

13 percent (Goldscheider and LeBourdais, 1986).

opportunity to develop tastes for roles alternative

reaching that age between 1976 and 1979. The

When the trend toward earlier nestleaving

to motherhood-especially employment-that as a

began, it was associated with declining marriage


age. However, it now is increasingly combined
with delays in marriage, with the result that more

result alter family formation. Rubin (1976) cites


impressionistic evidence that the rigidity of sex
roles in working-class families may result from
early marriage and men's lack of experience in
living independently of women-either mother or
wife-whose role it was to care for their physical
needs. This suggests that a role hiatus may also
affect young men's sex-role traditionality.
Spitze (1978) tested the role hiatus hypothesis
with longitudinal data, and found that those
women who attended college increased their
preference for employment, whereas those who
worked increased their preference for full-time
work in the home. However, a great many who
work or attend college after high school remain in

and more young people are spending an important


part of early adulthood in a context in which family

roles may be much less salient, and are developing


tastes and skills that are likely to reduce their
orientation to family roles. The transition to

adulthood for recent cohorts is following a


distinctly different pattern from that of their
parents, and could well have a substantial

influence on their later lives. We have already


documented that the experience of non-family
living leads to a postponement of marriage, at least
for women (Goldscheider and Waite, 1985). But
does it have an influence on the kinds of marriages
eventually formed through its effects on expectations about the balance of work and family roles? It
is these influences that we wish to explore.

BACKGROUND AND HYPOTHESES

The centrality of family formation in all societies


results in well-defined norms, rooted in an

underlying sexual division of labor, norms that


traditionally lead women to orient themselves more
toward family roles than do men. In the United
States the proper ages at which to marry and bear a
first child are normatively well defined, and clearly
differ for men and women, allowing men more
time to acquire adult levels of independence and
autonomy (Ryder and Westoff, 1971; Rindfuss and
Bumpass, 1978; Modell, 1979). Scanzoni (1975)
found that marriage, work, and fertility are
strongly associated with sex-role norms: women
who defined themselves through traditional female
roles married earlier, were less likely to be
employed full-time or to use effective contraception, and expected to bear more children than
women with a more non-familial orientation. He
sees sex-role modernity primarily as an emphasis
on individualism (although he makes this equation
reluctantly), and summarizes his results by concluding that "The greater the individualism, the less
the familism" (p. 187).
Family-related norms clearly affect the behavior

of individuals; they may also respond to that


behavior, as Marini (1984) argues. In either case,
it is important to establish how such attitudes are
formed and altered by experience. Although some
researchers postulate that family-related norms are

laid down early in life (Blake, 1972; Gustavus and


Nam, 1970; Hoffman and Wyatt, 1960), other
researchers have emphasized how experiences

the parental home (Goldscheider and DaVanzo,


1985). Non-family living prior to marriage may
provide a more complete "hiatus," and lead to
changes in the family-related attitudes, plans and
expectations of young men and women.
We test this proposition by examining young
women's plans for employment over the long run,
their total expected family size, the number of
children that they feel is an ideal family, men's and
women's attitudes toward the employment of
mothers under various conditions, and two dimensions of global sex-role attitudes. Together these
measures reflect many aspects of young adults'

orientation to family versus individual concerns. In


these analyses, we hold constant their earlier plans'
and attitudes, measured-to the extent possibleprior to any exposure to non-family living.
However, we show as well that omitting such
controls does not bias the results observed, since
the likelihood of non-family living in early
adulthood is not strongly related to underlying
work and family attitudes.

Specifically, we test the following propositions:


-The experience of non-family living during

early adulthood affects the attitudes, values,


and plans of young adults, compared with the
levels measured before the experience, and
moves them away from a traditional family
orientation.

-Living away from home has stronger effects


on personal plans and expectations than on

more global ideals, although the latter may be

affected as well, presaging normative change.


-Non-family living will have stronger effects
on the attitudes, expectations, and plans of
women than of men, since more of a role

hiatus is expected for men, even under

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

NONFAMILY

LIVING

543

traditional sex-role regimes, but non-family

variable. Most models incorporate two measures of

living may affect men's views as well.

the relevant attitudes, plans, or expectations: one

-Nonfamily living interacts with education past

measured at the beginning of the period and one

high school to alter attitudes, expectations and

measured later and used as the dependent variable.

plans. Both attending college and living away

Our sample is restricted to those age 14 to 17 in the


first year of the survey to allow us to observe and
measure a complete history of non-family living

from parents influence young adults, but those


who go away to college change their views
more than those who go to college while
living at home.

prior to marriage.4 Thus, we have data on Young


Women from 1968 through 1978, and on Young

-Those with more family-oriented attitudes are

Men from 1966 through 1976; each group was 24

not less likely to live away from home during

to 27 years old at the last interview analyzed here.

early adulthood than others, so that effects of

The outcome measures include a wide range of

non-family living will not change greatly with

young adults' family- and sex-role-related atti-

the addition or omission of controls for

tudes, plans, and expectations. First, young

attitudes, expectations or plans at time one

women were asked in each survey year about their

(prior to the risk period for nonfamily

plans to work at age 35, as an indicator of long-run

living).2

plans for employment. We divided this measure


into (a) those who planned to work and (b) those
who planned not to or were undecided.5 Second,

DATA AND METHODS

Data for this analysis come from the National


Longitudinal Surveys of Young Women and

Young Men. Conducted by the Ohio State


University Center for Human Resource Research,
these surveys include information over a recent
15-year period on more than 10,000 young men
and women. Personal interviews were conducted

with national probability samples of the non-

institutionalized population of females age 14 to 24


in 1968, and males age 14 to 24 in 1966. Those
included responded to lengthy interviews in many
of the succeeding years through the early 1980s.3
Attrition from the sample over the panel period has
been relatively low; three-fourths of the original
Young Women's sample was re-interviewed in

1978, and seventy percent of the original Young


Men's sample was re-interviewed in 1976. These
are the last years used here.

The analysis reported in this paper uses as


control variables measures of stable respondent

characteristics, such as race and characteristics of

the parental family, taken at the initial survey


(1966 for the Young Men, 1968 for the Young
Women). Independent variables measuring the
respondent's current situation-for example, employment, education or marital status-were measured contemporaneously with the dependent

we include two measures of fertility asked in 1971,

1973, and 1978: young women's expected family


size and the number of children that they think is
ideal for a family. Expected family size was
measured from questions on children already born
and additional children expected in the future.
Taken together, these questions cover the primary
roles-mother and worker-that young women

must balance during their adult lives.


Third, we examine three measures of respondents'
views about the proper ways that mothers in

general should balance work and family roles;


these are all scales. One of these, called attitude
toward mothers' working, consists of the sum of
responses to three questions (coded as five-point

Likert scales) about the conditions under which the


respondent thinks that married mothers of preschool children may work, given that a trusted
relative is available for child care: (1) if she needs
the money; (2) if she wants to and her husband
agrees; and, (3) if she wants to but her husband
doesn't particularly like it (see Appendix). The
points from these three items were summed into a

single scale which ranges from 3 (traditional) to 15


(liberal).6 Young Women answered these questions

'Preliminary examination of the available data


showed that none of the young women, and only a few
young men, lived away from home before age 18. Since
the initial survey did not ask about prior years' living

2 If early attitudes and non-family living are not

arrangements, we could not know if those older than age

strongly related, then eliminating early attitudes from the

17 at the first survey had already experienced some years

equation for later attitudes does not introduce specifica-

of non-family living.

tion bias; hence, where early and later attitudes are

5 Specifically, the young women were asked what kind

strongly related, lack of change in the coefficient for

of work they would like to be doing when they are 35.

non-family living when early attitudes are added or

Coded responses were: a) married, keeping home, raising

dropped indicates that young adults' likelihood of living

a family; b) same as present job; c) don't know; and d) an

away from home does not vary with their early attitudes.
3 The NLS Young Women were interviewed annually

occupation other than their current one.


6 For each of the three questions, five points were

from 1968 through 1973, again in 1975, 1977, 1978, the

assigned the response "definitely all right," and one

Young Men annually from 1966 through 1971, and again

point to "definitely not all right." The mid-point

in 1973, 1975, 1976, the last years that we use in the

response was "undecided." Although these three items

analysis reported here.

comprise a formal Guttman scale, we report results for a

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

544 AMERICAN SOCIOLOGICAL REVIEW


in 1968, 1972, and 1978; the Young Men did so in

1971 and 1976.

include years spent in a college dormitory, military


barracks, living with roommates of either sex and

Finally, two more general indices of sex-role

living alone. This measure uses the proportion of

attitudes were included in the 1972 and 1978

their families, and include value judgements about

potential years of non-family living actually away


from home, rather than the number of years, to
control for the differences between individuals in
the potential number of years. Since all respondents in our sample were living at home and
unmarried as of the first interview, only those who
married at age 18 had an opportunity to live away

appropriate roles for men and women. The

from home prior to marriage.

Young Women's interviews (see Appendix). These


were Likert-type questions, also coded as fivepoint scales. The questions overlap to a substantial
extent in the two years but are not identical. They

deal with the impact of women's employment on

variables used in our analyses are the composite

One disadvantage of the use of proportion rather

indices (or factor scores) derived from the

than number of years is that the smaller the number


of potential years of non-family living, the easier it

coefficient matrix of a factor analysis done with

sex-role attitudes (Jobs), measures views of the


importance of women's employment to their

is to have spent either all or none of it actually


away from home. If we assume that individuals
have an underlying propensity to live away from
home, then we have fewer observations on this
propensity for individuals with a small number of
potential years away. This gives us inefficient but
unbiased estimates of these underlying propensities

self-esteem and to the economic well-being of their

for these individuals.

oblique rotation. Our factor analysis of each of the


series produced two distinct dimensions. One,

which we called sex-role attitudes (Family),

reflects the importance of women's time at home to


their children and families. The second dimension,

families. A higher score indicates greater acceptance of non-familial roles for wives and mothers.

Finding two distinct dimensions of sex-role


attitudes in this series of questions replicates
results from a number of other studies. Mason and
Bumpass (1975) and Mason et al. (1976) found,
using questions overlapping to some extent with
those that we use, that respondents answered

relatively consistently all questions on the appropriateness of the traditional division of labor within
the family, and those on rights of the sexes in the
labor market, but apparently saw little need to
connect their views across these two dimensions.

Our key indicator of non-family living is the


proportion of the years that the respondent lived
outside the parental family between age 17 and
first marriage or the terminal date for the analysis,
whichever occurs first.7 Non-family living could

summated rating scale of the items, as described above,


to preserve maximum information about the respondents'
views toward employment of mothers. However, we
repeated all analyses after coding the items into a
Guttman scale. We created this scale with two alternative
coding schemes, either including or eliminating the
undecided category as a pass. Both scales performed
comparably. This Guttman scale has a coefficient of

reproducibility ranging from 0.97 to 0.99 for the Young


Women in 1968, 1972, and 1978, and a coefficient of
scalability varying between 0.81 to 0.94 for the same
years. For the Young Men, the Guttman scale on the
same items has a coefficient of reproducibility of 0.97 in
both years, and a coefficient of scalability of 0.85 and
0.87 in 1971 and 1976, respectively. Thus, the scales
appear to have identical-and excellent-measurement
properties for Boys and Girls.
7 This measure reflects the respondent's living arrange-

ments at the survey date, the only time for which we have
information about this variable. Living arrangements for
young adults have never been measured-to our

The use of proportion rather than number of


years away implies that the first year of non-family

living has a larger effect than subsequent years,


with the effect of each year declining at a rate of
(years - 1)/years, a slow to moderate decay
function. This function makes sense theoretically,
as we expect that the new experiences a young
adult undergoes by living away, as well as the new
skills acquired, are largest in the first year, with
additional skills and experiences added at a
declining rate with more years of non-family
living.

One could argue, alternatively, that family living

should include both time living with parents and


time living with a spouse. Just as young people
may decide to leave the family to live away, so
they may decide to form a new family of their
own. This argument implies a specification of
proportion of potential years of non-family living
spent actually away, that includes in the denominator all the years from age 17 until the measure of
the dependent variable, including years after
marriage, since the person could have lived away
those years if he or she had not married. We
created such a measure and include a discussion of
it in our results.

We measure living arrangements for respondents


using the household listing available in each survey
year. This gives the relationship to the respondent
of all individuals living in the household, their age,
sex, and certain other characteristics. Thus, we can

knowledge-at less than one-year intervals. We assume


relatively little change between survey dates in family vs.
non-family living. We weighted survey years to reflect
the actual amount of time between surveys.

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

NONFAMILY
identify those respondents who reside with parents
and those who live away at the survey date.8
We have also included as controls other

variables that are likely to affect the family- and


sex-role-related plans and attitudes of young
adults. These include family background character-

istics (socioeconomic status and family structure)


measured at the initial survey, and current
socioeconomic and family characteristics, measured in the same year as the dependent variable.
Table 1 presents definitions, means, and standard
deviations of all independent and dependent
variables.

Our analytic strategies vary depending on how


often and in which years measures of the
dependent variables were obtained. One measure,
work plans, was asked of the young women in
every survey year. The first response to this
question, asked when the young women were 17
years old, provides a Time 1 baseline before any

LIVING

545

For clarity, we have included a brief table that


gives the ages of the young women and the survey
years included for various outcome measures. The
young men were ages 24 to 27 when we measured
their attitudes toward mothers' work in 1976.
Equations with continuous dependent variables,
including both measures of fertility and all the
attitude scales, are estimated with ordinary least
squares. Equations for work plans, a dichotomy,
use logistic regression, estimated with maximumlikelihood techniques (Goodman, 1976). To permit
comparison of the effects of the independent
variables across equations, we transformed the
logit coefficients to yield measures analogous to
unstandardized ordinarily least squares (OLS)
regression coefficients (Hanushek and Jackson,
1977). The transformed logit coefficients reflect
the estimated effect of a unit change in the
independent variable on the probability of planning
to work at age 35, evaluated at the sample means.

non-family living could occur from which we

measure change. Plans for work at age 35 stated at


later ages are the dependent variable. We arbi-

trarily selected these later ages to be 20, 22, and 24


years old, at periods of 3, 5, and 7 years after the
first measure, based on questions answered for

RESULTS

ments available from these data see Goldscheider and

We discuss the results of the analyses described


above as they provide evidence on each of the
hypotheses presented earlier. We begin with the
first and most general hypothesis-that nonfamily
living changes young adults' plans, expectations
and attitudes about family and work roles away
from a traditional family orientation. Before we
discuss the individual models, however, we must
note the strikingly consistent picture that emerges
across the wide range of measures analyzed for
females: living away from parents tends to alter
young women's attitudes, expectations and plans
toward non-familial roles. The more experience
that young women have with independent living,
the more they change their plans for later work,
their family size expectations, their attitude toward
work by mothers, and their family-related sex-role
attitudes away from a traditional family orientation. In contrast, no significant effects appear in
the one analysis possible for young men, although
the coefficients have the expected sign.
Turning to the detailed results, we see in Table 3
that after controlling for young women's plans for
employment as reported when they were 17 years
old, the higher the proportion of years lived away
from parents between age 17 and the age in
question (or first marriage, whichever came first),
the greater the likelihood that the young women
planned to work at age 35. This effect is large and
significant for age 20, although it attenuates in size
with increasing age, suggesting that the effects of
non-family living may be short-lived. The coefficients imply that a young woman who lived away
all of the time between age 17 and 20 or her
marriage (whichever came first) was about 10
percentage points more likely to plan for later

Waite (1985) and Goldscheider (1985).

employment, net of her plans stated only 3 years

20-year-olds between 1971 and 1973; for 22-yearolds between 1973 and 1975; and for 24-year-olds
between 1975 and 1978. This portion of the
analysis provides the strongest test of our hypotheses, in that work plans prior to any experience of
non-family living can be fully controlled. Comparing the effects of non-family living with and
without controlling work plans at age 17 tests the
importance of controlling them.
The second set of dependent variables includes
those measured only occasionally. These are
expected and ideal fertility, attitudes toward
mothers who work under certain specific conditions, and the more global set of questions on
sex-role attitudes (Family and Jobs). For these

dependent variables we modified our analytic


strategy to use the later measure as the dependent

variable, with the earlier measure included as an


independent variable to provide an approximate
baseline. Thus, in our analysis of fertility expectations and ideal fertility, we predict 1973 expectations and ideals (when the young women were
aged 19 to 22) from 1971 expectations and ideals
(when they were 17 to 20), plus characteristics and

experiences of the individual. This latter analytic


strategy does not always allow us to measure the

attitude in question before any experience with


non-family living. We test the importance for our
conclusions of this limitation.

8 For further details, on measures of living arrange-

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

546

AMERICAN

SOCIOLOGICAL

REVIEW

Table 1. Description of Independent Variables Used in the Analysis*


Means (standard

deviations for

continuous variables)
Nonfamily living Proportion of years spent in non-family living between .199
age 17 and first marriage or age in question (.309)
Work 35 at 17 (D) Plans to work when age 35 as measured at age 17 .416
Black (D) 0=White and other races .299
Non-Intact (D) Did not live with 2 natural parents at age 14 .186
Parents' Education Average years of school completed 10.397
by
parents
(3.225)
South (D) Census region .396
Size of Labor Scale ranging from 1 (rural) to 8 3.431
Force in Area (urbanized areas of 3 million or more) (2.356)
Year Year became age in question 74.019
(1.001)

Education Years of schooling completed 12.581


(2.189)

Enrolled (D) Enrolled in school full-time .119


Employed (D) Current full-time or part-time employment .573
Husband's total income In 1967 constant dollars (thousands) 2.504
(3.664)
Married (D) Currently married for the first time .466
Divorced (D) Currently divorced, separated, or widowed .062
Remarried (D) Currently remarried .014
Kids (D) Living with own children .368
Employed * kids Interaction term .132
Standard deviation
(for continuous

Dependent

Variables

Plans to hold a job at age 35


(I=yes, 0=no):
age
20
age
22
age
24
Attitude toward mothers working
(3 to 15, high = more liberal):
Girls

1968
1972

Means

variables)

.551

.554
.652

9.733
10.682

2.577
2.400

Boys

1971
10.672
2.926
1976
11.785
2.780
Sex-role attitudes (Family)
(-2.9 to 1.6, high = more liberal):
1972
.005
.875
1978
.025
.923
Sex-role attitudes (Jobs)
(-2.3 to 1.8, high = more liberal):
1972
-.008
.711
1978
-.003
.715
Expected family size
(in children, 0 to 13):
1971
2.509
1.561
1973
1.949
1.253
Ideal family size
(in children, 0 to 14):

1971
1973

2.944
2.620

1.251
1.130

(D) Dummy variable.

* The statistics for the independent variables come from the Young Women's sample at age 22.

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

NONFAMILY

LIVING

547

Table 2. Ages and Years Measured for Young Women


Survey

Years

Ages

Work Plans Attitude toward

at Ages: Mother's Working Fertility Sex Roles


Age

in

14
15
16
17

1968

1973

20

22

24

1972

1973

1978

1978
1975

1977

18-21

19-22

24-27

1972
1971

1973

1975

No surveys were done in 1974 or 1976.

earlier, compared with a woman who did not live

either always lived with parents or moved straight

away during this period.9

from parents to husband. '0

The analysis of young women's attitudes toward

Ideal family size (Table 4, column 4), in

mothers' working, shown in the first column of

contrast, shows no effect of non-family living.

Table 4, also shows strong, statistically significant

This result reinforces the theoretical and analytic

positive effects of non-family living on acceptance

differences between these measures. Ideals seem

of employment of married mothers with young


home all of the possible years between age 17 and

quite stable, as reflected by the somewhat larger


effect of the prior measure for this than for birth
expectations, and by the lack of effect from all the

marriage, or Time 2, increased her score on this

new experiences except school enrollment.

children. A young woman who lived away from

scale by about 1 point on a 13-point scale

The right-most columns of Table 4 present the

compared with someone who never lived away.

two sex-role attitudes scores (Family and Jobs),

We find similar patterns in our analysis of the total


number of children expected by young women.

measured in 1978. The results show that non-

Living away from home reduces fertility expecta-

acceptance of non-family roles for women with

tions, with a decline for those who experience

family responsibilities (Family) and their view of

non-family living throughout the interval of about

the benefits of women's employment to their

a quarter of a child, compared with those who

family living increases both young women's

self-esteem and to their families' economic position (Jobs). When we analyze these two scales as

9 We were concerned with the possibility that the

effects we observed for non-family living might be


spurious, resulting from the unmeasured effects of work

measured in 1972, without a Time 1 control, we


find an even stronger effect of non-family living on
Family, but no effect on Jobs (results not shown).

experience, since our models include only current

These results for 1978, an outcome measured at the

employment status. We thought this unlikely, since

latest survey year and at the oldest ages, suggest

Spitze (1978) found that holding a job rather than going

that the effects of non-family living are not

to college during early adulthood caused a decrease in

necessarily short-lived, despite the attenuation

young women's taste for employment at age 35. We also


knew that research on the relationship of employment and

non-family living early in adulthood has shown them to


be only weakly related. Many young adults leave home

for non-work related reasons, and many others remain at

apparent on active work plans; its influence

persists up to a decade for fundamental attitudes


about women's work and family roles.
We performed several checks on the susceptibil-

home and work after finishing school. McElroy (1983)

ity of our findings to changes in the measurement

found a weak positive relationship between employment

of non-family living. First, we re-estimated all the

and non-family living for a sample of high school


dropouts and Goldscheider and DaVanzo (1985) found a

coefficients for our measure of non-family living were

weak negative relationship over the seven years after

larger in each of the alternate specifications than in that

completing high school in a sample that was restricted to

with current employment. We conclude that the use of

those who reached twelfth grade.

current employment does not mask effects of employ-

Nevertheless, reasoning that the problem should be

ment history in the model that theoretically should have

most acute for employment-related attitudes, we re-

the strongest effects of employment on the outcome

estimated the models of plans for work at age 35 (shown

measure, and that our results with current employment

in Table 3) three times, substituting various detailed

are probably somewhat conservative. To keep the

measures of employment history for current employment.

complexity of the reported results within reasonable

These measures were: (1) the proportion of years since

bounds we report only the analyses with current

age 17 that the respondent worked at all at any time

employment, but the alternative specifications are

during the year; (2) the proportion of years that the

available from the authors on request.

respondent worked at least 1000 hours (approximately half

10 As we noted earlier, we repeated all these analyses

time); and (3) the proportion of years that the respondent

with this scale coded into a Guttman scale, coded from 0,

worked at least 26 weeks (approximately a half-year).

for those who agreed that mothers should work under

Each of these models showed exactly the same pattern of

none of the stated conditions, to 3, for those who agreed

effects of non-family living on changes in work plans,

that mothers could work under all the conditions. The

with the same pattern of significance. In fact, the

substantive conclusions were identical.

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

548

AMERICAN

SOCIOLOGICAL

REVIEW

Table 3. Effects of Non-family Living and Other Variables on Young Women's Plans to Work at Age 35
Age

Explanatory

Variables

20

22

24

Non-family Living 0.101** 0.083 0.042


Work plans for 35 measured at age 17 0.279** 0.213** 0.179**
Black
0.138**
0.128**
0.021
Non-Intact 0.075** -0.051 0.116**
Parents' Education -0.004 0.006 -0.003
Employed
0.042
-0.033
0.158**
Enrolled
0.284**
0.186**
0.132*
Size
0.002
0.010
-0.002
South
0.053*
0.051
0.057*
Year
-0.037**
-0.007
0.021*
Married
-0.003
0.031
-0.066
Divorced/Separated 0.023 0.049 -0.069
Remarried
-0.031
0.111
0.065
Husband's Total Income -0.006 -0.011 0.001
Kids
0.080*
0.057
0.161**
Employed * Kid 0.091 0.068 -0.079
Educational Attainment 0.037** 0.059** 0.039**
N
1227
740
928
2 - log-likelihood ratio (df= 18) 240.43 113.93 103.01
One-tailed significance:

* p. 05<p<.10.
** p<.05.

models for the young women using the alternative results (not shown but available on request from

specification of proportion of years at risk of the authors) show somewhat weaker effects, but in
non-family living actually way. This specification only one case (the 1978 Family attitude score) does
includes years after marriage as part of the years at a previously significant effect become even
risk, because the individual could have lived marginally insignificant under the new specificaindependently had she not chosen to marry. These tion.

Table 4. Unstandardized Coefficients for Models of Young Women's and Men's Family Plans and Attitudes
Attitude Toward
Mothers' Working Expected Ideal Sex-role Attitudes

Variables

Girls Boys Family Size Family Size Family Jobs


1972 1976 1973 1973 1978 1978

Non-family Living 0.979** 0.261 -0.225** -0.066 0.161** 0.112*


Initial Measure 0.130** 0.211** 0.327** 0.407** 0.338** 0.223**
Black 0.385** 0.517** 0.128** 0.143** -0.040 0.114**
Non-Intact 0.264 0.271 -0.133** -0.105* -0.076 0.042
Parents' Education 0.024 0.052** -0.020** -0.010 0.001 -0.007
Employed 0.338** -0.126 -0.152** -0.075 0.401** 0.275**
Enrolled 0.048 -0.163 - .0.260** -0.170** 0.610** 0.159**
Size 0.041 0.037 -0.008 0.009 0.015* 0.009
South 0.432** 0.045 -0.216** -0.128** 0.064 0.024*
Age at
first survey 0.081 0.063 -0.104** -0.017 -0.015 0.024*
Married 0.073 -0.153 -0.065 -0.027 0.024 0.146**
Divorced/Separated 0.399 0.333 -0.340** -0.118 0.171** 0.029
Remarried -0.888 0.246 0.115 -0.210 0.179* 0.164**
Husband's Total

Income 0.004 -0.004 -0.012 0.005* -0.007**


Kids 0.407* 0.318** 0.001 -0.088* -0.259**
Employed * Kid -0.352 - -0.177* -0.031 0.014 0.217**
Education 0.053 0.116** 0.033* 0.015 0.033** -0.014*
N 1472 1153 1344 1365 1223 1223
R 2 0.064 0.091 0.230 0.229 0.233 .182

One-tailed significance:

*= .05 < p < .10.


** = p < .05.

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

NONFAMILY
Second, we re-estimated these models using a

LIVING

549

of the dependent variable." In no case did this

find no significant effect of non-family living on


ideal family size. However, we see significant
effects of living away on attitudes toward mothers'
working and both the Family and Jobs dimensions
of the Sex-role attitudes measure. All these
represent global, rather than personal views. We
conclude that non-family living affects some but

fully saturated model significantly improve on the

not all dimensions of global attitudes.

fit of the restricted model that we estimated, as

Our third hypothesis deals with sex differences


in the effects of independent living. The NLS data
contain very few measures of young men's

specification of non-family living that allows for


separate effects for each number of years of

independent living, given years at risk. It also


controls for number of years at risk, that is, the
number of years between age 17 and the measure

measured by F-statistics for the regressions and


Chi-squared statistics for the logits (all p > .08).
Among the control variables, a result worth
noting is that for NONINTACT, which also

measures non-traditional living arrangements.


NONINTACT shows a positive effect on plans for
work at age 35 at two of the three ages presented in

Table 3; in the equations for the other measures


NONINTACT tends to operate in the same

direction as non-family living. Where it has any


effect, it decreases family orientation, suggesting
that living in a non-traditional family during
adolescence moves young women away from plans
for family roles later in life. This finding is
consistent with our proposition that experiences in
non-family living during young adulthood erode an

orientation toward traditional family roles.


We also find that education is less associated

with family-oriented attitudes and plans, as is


current employment, but we see effects of marital
status only for sex-role attitudes and expected

attitudes toward family life or appropriate roles for


men and women. The one measure available (for
1971 and 1976) corresponds exactly to one of the
scales analyzed for young women: attitude toward
mothers' working. This is shown in the second
column of Table 4. We see no significant effect of
non-family living on men's views, in contrast to
those we found for women. Note, however, that
the coefficient for young men has the expected sign
and is nearly significant at the p<.10 level for a
one-tailed test. Moreover, we also saw a small but
statistically significant (p<. 10) effect of nonfamily living on young men's 1971 responses to
these same questions. This pattern suggests weak
effects of non-family living on young men's
family-related attitudes. The sex difference lends
support to our reasoning that independent living
has stronger effects for females than for males. 12

In Table 5 we test the proposition that college

family size. Not surprisingly, women who experi-

education and independent living reinforce each

ence marital disruption have decreased their

other to change young women's attitudes, plans,


and expectations. The first variable, COLLEGE
AWAY, identifies those individuals who lived
away from home at any time while attending
college. The second, COLLEGE HOME, are those
who attended college but never lived away from
home before marriage or the analysis age. Each of
these groups accounts for about one-fifth of the
sample. The third, COLLEGE OTHER, identifies
those who lived at home while attending college
but lived independently either before or after (3
percent of the sample). The fourth, NONCOLLEGE
AWAY, identifies those who lived away from
home but never attended college ( 1 percent of the
women). The omitted category comprises nearly
half the sample: those who never went to college
and never lived away from home before marriage or
the age in question, whichever comes first.
The first two panels of Table 5 show the effects
of college attendance and non-family living for
young women's plans for work at age 35. They

expected number of children. Married women and


working mothers hold more liberal views on the
Jobs measure, whereas divorced and remarried
women are more liberal on the Family scale. Black

women seem to hold more positive attitudes


toward work, however this is measured, and to
have a somewhat larger expected and ideal family

size, but do not differ from whites on familyrelated sex-role attitudes.

Our second proposition states that non-family


living not only affects personal plans and expectations, but also influences global ideals, although
less strongly, since individuals are likely to use the

extenuating circumstances in their particular case


to justify deviations from the ideal that they hold,
at least for a while. The results presented in Tables

3 and 4 offer mixed support for this reasoning. We

I The complete specification of this equation for a

differ onlv in the inclusion. in the scnnd set of

model with a maximum of 3 years at risk is:

Y = ao(l) + 02D2 + 03D3 + 04DI*I1 + 05D2*I1 +


036D3*11 + 037D2*12 + 038D3*I2 + PgD3*13 + OX

12 Part of the difference may be because the males'


analysis (1976) is 10 years after the initial interview,

where:

while that for females is only 4 years. Thus, the young


women are ages 18 to 21 when they report their attitudes

Dj = dummy for exactly j years at risk


1, = dummy for exactly m years away

towards mothers' working, whereas the young men are


ages 24 to 27.

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

550

AMERICAN

SOCIOLOGICAL

REVIEW

Table 5. Transformed Logit Coefficients for Models of Young Women's Plans, Attitudes and Expectations

(Controlling for All Variables in Table 2)


Plans for Work at Age 35
20

22

24

Model 1

Nonfamily

Living

0.101**

0.083

0.042

Model 2

College

away

0.173**

0.215**

0.148**

College * home 0.118** 0.107** 0.046


College other away 0.161* 0.131 0.139*
Non-college away 0.000 0.007 -0.014
Model 3

College

College

away

home

0.140**

0.088**

0.078

0.003

0.026

-0.047

College other away 0.141 0.052 0.079


Non-college away 0.002 0.023 0.003

Educational

attainment

0.029**

Attitude

0.056**

toward

0.041*

Sex-role

Mothers' Working Expected Ideal Attitudes


Girls Boys Family Size Family Sizes Family Jobs
1972 1976 1973 1973 1978 1978

Model 1
Nonfamily Living 0.979** 0.261 -.225** -.066 0.161** 0.112*
Model 2

College away 0.328* 0.400* 0.026 0.036 0.241** 0.027


College home -0.145 0.016 0.038 0.076 0.115* -0.061
College other away -0.382 0.230 -0.307* -0.282* 0.382** 0.099
Non-college away 0.551** -0.077 -0.199** -0.083 0.055 0.043
Model 3

College away 0.204 -0.342 -0.009 0.027 0.166** 0.071


College home -0.235 -0.567* 0.012 0.069 0.061 -0.030
College other away -0.427 -0.308 -0.322* -0.285* 0.344** 0.121
Non-college away 0.581** -0.089 - 0.193** -0.081 0.067 0.036
Educational attainment 0.094* 0.204** 0.020 0.005 0.024* -0.014

models, of a measure of educational attainment.

to the socializing effect of time spent with other

The two sets of analyses point to the same

students outside the classroom, away from parental

conclusion: young women who live away from

control and influence.

home while attending college are more likely to

The other panels of Table 5 show the joint


measures of college-going and non-family living
described above for the other outcome measures.
These coefficients suggest that both college
attendance and non-family living affect the magnitude of attitude change, whether or not these occur
together. These patterns are not always fully
consistent; in general, however, both those who
lived at home while attending college and then
lived away (a relatively small group), and those
who lived away at college showed more change
than others on the Family factor analysis scores.
As before, adding educational attainment attenuates but does not eliminate these relationships.
The final hypothesis is related to the importance
of including a Time 1 control measuring a given
attitude prior to any experience of non-family
living. One could argue that the relationship we
observe between attitudes and plans and nonfamily living arises because those who grew up
with the most non-familial views may be more

change their work plans than those who attend


college while living at home. These women, in
turn, are significantly more likely to change their
work plans toward employment than those who

remain at home and do not go to college. Those


who live at home while in college, but live

independently at some other time, also exhibit an


increased preference for working at age 35,

although the effect is not significant. Those who


live away from home without attending college
show no difference from the omitted category.
Adding educational attainment to the model
controls for the effect of years of school completed
by the age in question. This differentiates between
young women who attended college only briefly
and those who completed more years of higher
education. This control attenuates but does not
eliminate the effects of the joint events. The
coefficients suggest that living independently
reinforces the lessons learned at college, pointing

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

NONFAMILY

LIVING

551

Table 6. Effects of Non-Family Living on Attitudes and Expectations About Family Roles at Time 2 With
and Without the Inclusion of a Baseline Measure
Work Plans for Age 35

Age

Without

With

Time

Time

20

Age

.110**

.101**

22

Age

.120

.083

24

.068

.042

Attitudes toward Family Size Sex-Role Attitudes

Mothers' Working Expected Ideal Family Jobs


Without Time 1 .995** -.211 - .084 .192** .114
With Time 1 .979** -.225** -.066 .161** .112*

likely than others to live away from home. This

live away from home.'3 These results, in Table 7,

potential problem arises for all analyses where the

show that only the initial work plans variable has a


significant effect on the subsequent propensity for
non-family living-and only for age 22 or later.
The other attitude scales had no effect on
non-family living. These results corroborate those
in Table 6 that show that the percentage change in
the non-family living coefficient is relatively
small, except for the models for work plans at age
22 and age 24, where the coefficients increase 45

initial attitude measure may have occurred after


some experience with non-family living, as is the
case with sex-role scores. If so, the relationship
between non-family living and non-familial attitudes that we observe could be spurious and only
reflect this relationship with the omitted variable.

Our controls for early attitudes, measured prior to


any non-family living, eliminate this possibility for
those outcomes for which we have this baseline;
these are work plans for age 35 and attitudes

toward mothers' working for young women. Both

of these models show the hypothesized effect,


providing strong tests of our central proposition.
These models also allow us to determine the extent

of the strength of the association between nonfamily living and early attitudes, and hence, the
likely extent of specification bias in the analyses
without Time 1 controls.

To test our reasoning, we re-estimated all the


young women's models in Tables 3 and 4,

eliminating the initial measure of the dependent


variable. The resulting coefficients for non-family

living are presented in Table 6. Comparison with

to 62 percent when the Time 1 baseline measure is


omitted.

These results provide strong support for our


hypothesis that living away during young adulthood is not highly selective with respect to initial
attitudes regarding values or belief systems. Those
with the least family-oriented views to begin with
are not substantially more or less likely to live
away from their parents during the transition to
adulthood, so that these and other analysis of the

effects of non-family living on attitudes and


behavior should not be unduly concerned about
issues of selectivity. However, there is indication
that some selectivity may exist with respect to

quite small. In only one case, expected family size,

expressions of intentions or plans, if these plans


may produce actions that affect living arrangements. In our analysis, specific work plans for age
35 might well be expected to interact with living
arrangements more than judgments regarding
appropriate familial sex roles. In this case, it is
prudent to include a baseline measure, which we

does adding or deleting the initial measure change

have done in this analysis.

the coefficients for non-family living in Tables 3


and 4 shows that adding or deleting the initial
measure makes little difference in the conclusions;
generally the results for non-family living without
controls for the initial measure are slightly stronger
than those with controls, but these differences are

the significance of non-family living. Given the


strong relationships between the initial measure
and the later one shown in Tables 3 and 4, it is

clear that including the initial measure improves


the overall fit of the model; however, it does not

eliminate the effects of non-family living.


As a second test of this hypothesis, regression
models were estimated with the non-family living
measure at Time 2 as the dependent variable and
the Time 1 attitudes or work plans variable
included as an explanatory variable. All other
explanatory variables were retained as measured at
Time 2, which allowed us to isolate the effect the
attitudinal measure alone has on the propensity to

DISCUSSION

We argued earlier in this paper that living away


from home prior to marriage changes young adults'
attitudes, expectations, and plans away from
13 For example, measuring marital status at age 20 in
the work plans model adjusts for the effect marriage had
on the proportion of years lived away from home between
age 17 and 20. If explanatory variables were measured at

Time 1 when all women were single, the effects of the


attitudinal measure would be confounded with the effects

of marriage on the non-family living variable.

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

552

AMERICAN

SOCIOLOGICAL

Table 7. Effects of Time 1 Measure on Non-Family


Living Variable at Time 2
Time 1 Measure Coefficient Time 2
Work Plans-age 17 .017 age 20
Work Plans-age 17 .052** age 22
Work Plans-age 17 .090** age 24

Attitudes, Mothers
Working, 1968 .001 1972
Sex-role Attitudes,
Family, 1972 .008 1978
Sex-role Attitudes,
Jobs, 1972 .001 1978

REVIEW

possible mechanisms directly, but our results


suggest that such tests might prove a fruitful
avenue for further research.

Previous research on the development and


revision of sex-role attitudes has focused on
education, employment (Spitze, 1978; Dambrot et
al., 1983; Morgan and Walker, 1983), marriage,
becoming a parent, and divorce (Spitze and Waite,
1981). We find that education and employment are
associated with less family-oriented attitudes and
plans, but we see effects of current marital status
primarily for sex-role attitudes. Married women
hold less traditional views on the economic

family and toward individual concerns. Our results

benefits to the family of women's employment,

support this hypothesis consistently, indicating that

and divorced women and working mothers are

young women who lived independently became

more liberal on the appropriate division of roles

more likely to plan for employment, lowered their

within the family. Non-family living, by contrast,

expected family size, became more accepting of

seems to have broader effects than marital status,

employment of mothers, and became more non-

with coefficients sometimes larger than those of

traditional on sex roles in the family than those

employment and education. Non-family living is a

who lived with their parents. We found much

previously ignored but potentially important expe-

weaker effects of non-family living in the tests we

could perform for young men. Our results suggest


that personal plans and expectations respond in the

expected way to experiences in non-family living,


but that some dimensions of ideals and global

attitudes change as a result of this experience, as


well. We also find evidence that living away from

home reinforces the effects of attending college on

young women's attitudes, expectations and plans.


Given these results, it is important to learn more
about why non-family living should have such an

effect on attitudes about new family formation and

rience in the transition to adulthood, especially as

it influences young adults' attitudes and plans for


their own lives.

One particularly significant result of this research is the finding that the omission of Time 1
controls has little effect on measuring the impact of
non-family living on later orientations toward work
and family life. This provides increased confidence
that results obtained previously that show the

impact of non-family living on the likelihood of


marriage were not biased by the omission of some

the kinds of families likely to be formed. We

underlying anti-family attitude among those who

suggest the following possibilities. First, living


away from home lessens parental control over their
children's activities and may weaken the link
between parents' values, attitudes, and behavior
and those of their children. Young adults in their
own apartment are freed from parental curfews and
supervision of their friends and behavior, can
manage their own household in the way that they

both live away from parents early in adulthood and

choose, and may have a sexual freedom impossible


in their parents' house. Second, experience in

independent non-family living may equip young


adults with new social and domestic skills (such as
housekeeping, maintenance of a dwelling unit or
car, household finance, health care, and entertaining). Living away from home may give both young
men and women the self-confidence that they can
get along without a family and may enable them to

acquire the skills that they need to maintain their


independence as long as they wish. Third,
non-family living potentially exposes young adults
to a wider variety of experiences and influences

are also unlikely to marry. It also opens the way to

research on other possible sequelae of non-family


residence in early adulthood, such as marital
dissolution, or less traditional assignments of tasks
within marriage, allowing greater range in the

analysis of the forces leading to change in marriage


and the family in the late twentieth century.

These changes in the nature of the family, many


of which we discussed earlier, have altered some
of the bases on which traditional, gender-based
divisions of labor rested. These divisions may have

served some important functions under earlier

conditions (Marwell, 1975). But as Marini (1984)


has argued, norms about the transition to adulthood
tend to reflect behavior, as well as affect it, and
thus may change as behavior changes. The shifts in
individual attitudes, plans, and expectations that

our findings link with experiences with non-family

living may be a first step in the process of

than those encountered while they live with their

alteration of norms. Thus, our finding that

parents. These may cause individuals to change

non-family living influenced all our measures of

their plans and expectations for themselves, and

behavioral plans and evaluations, but did not affect

their views of appropriate behavior for men and

some indicators of global ideals, should not be

women, in general. We do not test any of these

interpreted as showing that non-family living may

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

NONFAMILY
only have a transient effect, one with impacts only
on those experiencing it. For if Marini is correct,
the increase in non-family living, by changing
plans relating to male and female roles in the

family, may presage larger shifts than we have


already observed in norms about family life.

LIVING

553

REFERENCES

Baldwin, Wendy and Virginia S. Cain. 1980. "The

Children of Teenage Parents." Family Planning


Perspectives 12:34-43.

Blake, Judith. 1972. "Coercive Pronatalism and American Population Policy." Pp. 84-108 in Aspects of
Population Growth Policy, edited by Robert Parke, Jr.

and Charles Westoff. Washington, D.C.: The CommisAPPENDIX

Attitudes towards mothers' working

"Now I'd like you to think about a family where there is


a mother, a father who works full-time, and several
children under school age. A trusted relative who can
care for the children lives nearby. In this family situation,

how do you feel about the mother taking a full-time job


outside the home-

a. If it is absolutely necessary to make ends meet?


b. If she wants to and her husband agrees?
c. If she prefers to work, but her husband doesn't
particularly like it?

Sex-role attitudes scores #1 and #2, 1972

a. Modern conveniences permit a wife to work


without neglecting her family.
b. A woman's place is in the home, not in the office
or shop.

c. A job provides a wife with interesting outside


contacts.

d. wife who carries out her full family responsibilities


doesn't have time for outside employment.
e. A working wife feels more useful than one who
doesn't hold a job.

f. The employment of wives leads to more juvenile


delinquency.

g. Working wives help to raise the general standard of


living.

h. Working wives lose interest in their home and


families.

i. Employment of both parents is necessary to keep up


with the high cost of living.

Sex-role attitudes scores #1 and #2, 1978

The 1978 series includes items a, b, d, e, f, and i. It also


includes the questions listed below.

j. It is much better for everyone concerned if the man


is the achiever outside the home and the woman
takes care of the home and family.
k. Men should share the work around the house with
women, such as doing dishes, cleaning, and so
forth.

1. A working mother can establish just as warm and


secure a relationship with her child as a woman who
doesn't work.

m. Women are much happier if they stay at home and


take care of their children.

n. A woman should not let bearing and rearing


children stand in the way of a career if she wants it.

Items a and c had no relationship with either dimension in


the 1972 scales, and were dropped; items a and k were
deleted from the 1978 scores for the same reason. Items
e, g and i loaded highly on score #2 (jobs), and the
remaining items loaded highly on score #1 (family).

sion on Population Growth and the American Future


Research Reports, U.S. Government Printing Office,
Volume 6.

Davis, Kingsley. 1982. "Changes in Marriage since


World War II." Paper presented at the conference on
"Contemporary Marriage: Comparative Perspectives
on a Changing Institution." Center for Advanced
Study in the Behavioral Sciences, Stanford, California.
Dambrot, Faye H., Mary E. Papp, and Cheryl Whitmore.

1983. "The Sex-Role Attitudes of Three Generations


of Women." Personality and Social Psychology
Bulletin 10:469-73.

Espenshade, Thomas J. 1985. "Marriage Trends in

America: Estimates, Implications, and Underlying


Causes." Population and Development Review
11:193-245.

Frey, William H. and Frances Kobrin. 1982. "Changing


Families and Changing Mobility: Their Impact on the
Central City." Demography 19:261-77.

Goldscheider, Frances Kobrin. 1985. Crossvalidating


Data from National Longitudinal Surveys Measuring
the Transition to Adulthood N:9999:NICHD. Santa
Monica, CA: The Rand Corporation.
Goldscheider, Frances Kobrin and Julie DaVanzo. 1985.
"Living Arrangements and the Transition to Adulthood." Demography 22:545-63.

Goldscheider, Frances Kobrin and Celine LeBourdais.

1986. "The Decline in Age at Leaving Home:


1920-1979. Sociology and Social Research 70:99-102.
Goldscheider, Frances Kobrin and Linda J. Waite. 1985.

"Effects of Nestleaving Patterns on the Transition to


Marriage. " Paper presented at the 1985 annual

meetings of the Population Association of America,


Boston.

Goodman, Leo A. 1976. "The Relationships Between


Modified and Usual Multiple-Regression Approaches
to the Analysis of Dichotomous Variables." Pp.
83-110 in Sociological Methodology, edited by David
Heise.

Gustavus, Susan 0. and Charles B. Nam. 1970. "The


Formation and Stability of Ideal Family Size Among

Young People." Demography 7:43-52.


Hanushek, Erik A. and John E. Jackson. 1977. Statistical
Methods for Social Scientist. New York: Academic.
Hoffman, Lois W. and Frederick Wyatt. 1960. "Social
Change and Motivations for Having Larger Families:
Some Theoretical Considerations." Merrill-Palmer

Quarterly 6:235-44.
Kobrin, Fances E. and Linda J. Waite. 1984. "Effects of

Family Stability on the Transition to Marriage."


Journal of Marriage and the Family 46:807-16.

Marini, Margaret Mooney. 1984. "Age and Sequencing


Norms in the Transition to Adulthood." Social Forces
63:229-44.

Marwell, Gerald. 1975. "Why Ascription? Parts of a


More or Less Formal Theory of Functions and

Dysfunctions of Sex Roles." American Sociological


Review 40:445-55.

Mason, Karen 0. 1974. Women's Labor Force Partici-

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

554

AMERICAN

SOCIOLOGICAL

nation and Fertility. Research Triangle Park, NC:


Research Triangle Institute.

Mason, Karen Oppenheim, and Larry L. Bumpass.


"U.S. Women's Sex-Role Ideology." American
Journal of Sociology 80:1212-19.
Mason, Karen Oppenheim, John L. Czajka, and Sara

REVIEW

Ryder, Norman and Charles F. Westoff. 1971. Reproduction in the United States, 1965. Princeton, NJ:
Princeton University Press.

Scanzoni, John H. 1975. Sex Roles, Life Styles and


Childbearing. New York: Free Press.
Spitze, Glenna D. 1978. "Role Experiences of Young

Arber. 1976. "Change in U.S. Women's Sex-Role

Women: A Longitudinal Test of the Role Hiatus

Attitudes, 1964-1974." American Sociological Re-

Hypothesis." Journal of Marriage and the Family

view 41:573-96.

40:471-80.

McElroy, M. 1983. "The Joint Determination of

Spitze, Glenna D. and Linda J. Waite. 1981. "Young

Household Membership and Market Work: The Case

Women's Preferences for Market Work: Resources to

of Young Men." Discussion Paper 83-1, Economics

Marital Events." Research in Population Economics

Research Center, NORC, Chicago.

3: 147-66.

Modell, John. 1979. "The Marriage Transition: Norma-

Terhune, Kenneth W. 1974. A Review of the Actual and

tive Aspects of Change Since World War II." Paper

Expected Consequences of Family Size. Washington


D.C.: U.S. Government Printing Office.

presented at the meetings of the Population Association


of America, Philadelphia.
Morgan, Carolyn Stout, and Alexis J. Walker. 1983.
"Predicting Sex Role Attitudes." Social Psychology
Quarterly 46:148-51.

Pope, Hallowell, and Charles W. Mueller. 1976. "The

Intergenerational Transmission of Marital Instability:


Comparisons by Race and Sex." Journal of Social
Issues 32:49-66.

Thornton, Arland, and Deborah Freedman. 1982.

"Changing Attitudes Toward Marriage and Single


Life." Family Planning Perspectives 14:297-303.
U.S. Bureau of the Census. 1980. "Household and

Family Characteristics: March 1980." Current Population Reports, Series P-20, No. 366, Washington,
D.C., U.S. Government Printing Office.

Rindfuss, Ronald R. and Larry Bumpass. 1978. "Age

Waite, Linda J. and Ross M. Stolzenberg. 1976.

and the Sociology of Fertility: How Old is Too Old?"


Pp. 43-56 in Social Demography, edited by Karl
Taeuber, Larry Bumpass and James Sweet. New York:

of Young Women: Insights from Nonrecursive Mod-

"Intended Childbearing and Labor Force Participation


els." American Sociological Review 41:235-52.
Westoff, Charles F. 1978. "Some Speculations on the

Academic.

Rubin, Lillian Breslow. 1976. Worlds of Pain: Life in the


Working-Class Family. New York: Basic Books.

Future of Marriage and the Family." Family Planning


Perspectives 10:79-83.

This content downloaded from 195.229.156.148 on Fri, 15 Apr 2016 17:54:17 UTC
All use subject to http://about.jstor.org/terms

Potrebbero piacerti anche