PAA Poster 2002: Women in the Great Migration

1
The Difference a State Makes Womens Allocation
of Unpaid Work in the 50 States Jennifer Hook,
Department of Sociology, University of Washington
Results
Research Context
Measures

• There is considerable variation in the amount of
  time women spend on unpaid work across the United
  States. Estimates, based on the 2003 American
  Time Use Survey, range from 22 hours a week in
  Idaho to over 35 hours in Nebraska. Some of this
  variation is explained by differences in
  demographic composition across the states, but
  some of this variation remains unexplained.

• This figure shows predicted values of womens
  unpaid work time for varying levels of child care
  costs - set at the minimum, mean, and maximum
  observed values, separately for mothers of
  infants, pre-schoolers, and school-aged children.
  All other variables are set at the mean.

• I model unpaid work time in minutes per day. The
  measure encompasses most forms of unpaid domestic
  work, including child care.

Table 1. Individual - level variables
Table 2. State - level variables
Figure 2. The effect of child care costs.
Figure 1. Womens mean unpaid work time by state

• In this portion of the project, I examine the
  influence of state-level conditions on womens
  unpaid work time. Theoretical and empirical work
  in cross-national contexts suggests that both
  levels of gender equality and specific policy
  configurations influence household labor.

• The positive effect of having an infant,
  pre-school, or school-aged child on womens
  unpaid work time is larger in states where the
  cost of child care is a larger percentage of the
  median household income.
• The negative effect of womens individual
  earnings is less negative where women have more
  political equality.
• The positive effect of womens work hours is less
  positive where women have more political
  equality.

Results
Table 3. HLM predicting womens unpaid work time
Hypotheses

• I hypothesize that women will do less unpaid work
  where child care is less expensive and gender
  equality is greater (Hook 2005).
• I expect the costs of child care to be relevant
  for mothers, particularly mothers of infants and
  pre-school aged children.
• I further expect levels of gender equality to
  influence the effect of womens earnings and work
  hours. Where gender equality is greater the
  negative effect of womens earnings and of
  womens work hours on unpaid work time should be
  greater (Fuwa 2004).

Conclusions

• Both state- level gender equality and specific
  policy conditions influence womens unpaid work
  time across the United States. The effects,
  however, are not direct they work through the
  effects of individual-level characteristics
  including parenthood, earnings and work hours.
• In contrast to cross-national findings, womens
  macro-level earning equality appears unrelated to
  their unpaid work time.
• Womens political equality is related, but not in
  the hypothesized way for womens individual
  earnings. It appears that in politically unequal
  states, the negative effect of womens earnings
  on their unpaid work time is actually greater,
  suggesting that where equality is low women are
  better able to use their increased earnings to
  reduce their unpaid workload.
• Future work on this project will explore the
  intersection of paid and unpaid work in the lives
  of American women.

Data and Sample

• I utilize the American Time Use Survey (2003).
• I restrict the sample to working age women (ages
  22 to 59) and to states with more than 30
  respondents in this category. I lose Alaska,
  Delaware, the District of Columbia, Hawaii,
  Montana, North Dakota, South Dakota, Vermont, and
  Wyoming. The state-level N ranges from 38
  respondents in Rhode Island to 762 in California
  for a total N of 7,741 at the individual-level
  and 42 at the state-level.

Methods

• I estimate hierarchical linear models.

Note All education variables, Other race, and
Hispanic are non-significant and omitted from the
table.