Conference draft: do not cite without permission of the authors



Download 0.55 Mb.
Page6/6
Date conversion19.02.2016
Size0.55 Mb.
1   2   3   4   5   6

Notes: *** indicates p<.001, ** indicates p<.01, and * indicates p<.05, two-tailed tests.

Table A-2. Event History Coefficients, Standard Errors, and Risk Ratios for Selected Characteristics on the Risk of Having a Second Child (vs. Remaining Singleton)



WHITE


BLACK

 

Coefficient




Risk Ratio

Risk Ratio

Coefficient

Risk Ratio

Risk Ratio

 

Std. Err.




2nd Child

Singleton

Std. Err.




2nd Child

Singleton

Family Structure

























Married

0.024




1.024

0.976

0.221

**

1.248

0.802

 

(0.132)










(0.093)










Cohabiting

-0.204










0.382

***

1.466

0.682

 

(0.177)










(0.146)










Reproductive History

























Duration (from age at 1st child)

-0.037

***

0.964

1.038

-0.071

***

0.932

1.073

 

(0.005)










(0.010)










Previous Abortion

-0.196

***

0.822

1.216

-0.322

***

0.724

1.380

 

(0.053)










(0.084)










Previous Miscarriage

-0.161

***

0.851

1.175

-0.340

***

0.712

1.405

 

(0.032)










(0.080)










Infecundity Issues

-0.096

*

0.908

1.101

-0.123










 

(0.052)










(0.098)










Sterilized

-0.958

***

0.384

2.606

-1.114

***

0.328

3.047

 

(0.079)










(0.139)










Human Capital

























High School Graduate

-0.557

***

0.573

1.745

-0.076










 

(0.187)










(0.104)










Married*High School Graduate

0.494

**

1.639

0.610

NS










 

(0.198)






















Some College

-0.814

***

0.443

2.257

-0.063










 

(0.211)










(0.125)










Married*Some College

0.871

***

2.390

0.418

NS










 

(0.221)






















College Degree

-0.817

***

0.442

2.263

-0.609

**

0.544

1.838

 

(0.297)










(0.247)










Married*College Degree

0.960

***

2.611

0.383

0.693

***

1.999

0.500

 

(0.302)










(0.268)










Part-time Employment

0.085










0.157










 

(0.063)










(0.137)










Full-time Employment

-0.146

***

0.864

1.158

-0.146










 

(0.053)










(0.098)










Socioeconomic Characteristics

























Mother College Graduate

0.101

*

1.106

0.904

0.026










 

(0.056)










(0.123)










Mother Full-time Worker

-0.042










-0.062










 

(0.048)










(0.080)










Below 150% Poverty

0.209

***

1.233

0.811

0.300

***

1.350

0.741

 

(0.063)










(0.095)










Church Attendance

0.046










0.021










 

(0.044)










(0.125)










Demographic Characteristics

























Latina

-0.016










-0.100










 

(0.073)










(0.254)










Immigrant

0.037










0.393

**

1.482

0.675

 

(0.072)










(0.160)










NSFG Cycle

-0.141

***

0.869

1.151

-0.100

*

0.905

1.106

 

(0.031)










(0.053)










Notes: *** indicates p<.001, ** indicates p<.01, and * indicates p<.05, two-tailed tests.

1 Authors’ names are listed alphabetically. Please direct correspondence to Michelle J. Budig and Jennifer Lundquist, Social and Demographic Research Institute, W34A Machmer Hall, 240 Hicks Way, University of Massachusetts, Amherst, MA 01003-9278. Email: budig@soc.umass.edu or lundquist@soc.umass.edu.


2 These rates compare to peak birth rates of 99.2 for white women aged 25 to 29 and 128.9 for black women aged 20 to 25 (Downs 2003).

3 Of course, a nontrivial number of women become mothers through adoption, perhaps in response to the rise of age-related infertility issues. A strength of our approach is that we include adoptive mothers in the analyses and do not count them as childless.

4 Although birth rates to women at older ages have risen in recent decades due to fertility delay and infertility treatments, the rates are still very low.  Of all first births in 2003, for example, only 1% were to women ages 40-44 (Martin et al 2005).  Higher parity births are slightly more common at these ages, but still relatively rare, at just under 2% (Martin et al 2005).  We acknowledge therefore that our childlessness sample and only child sample may be slightly overestimated (more so in the latter case than the former case), but believe the effect to be negligible.

There is an NSFG variable that collects intentions on future fertility plans, which we plan to incorporate into the next version of this paper. A drawback is that it is asked of all respondents for waves 4 and 6 but only of currently married respondents in wave 5.



5 Our dependent variable is defined in the first set of models as the date of first biological birth or formal adoption and in the second set of models as the date of second biological birth or formal adoption. We expand our definition from biological parenthood to social parenthood for the following reasons: for all intents and purposes a childless woman who adopts is no longer childless, and furthermore the presence of an adopted child may be just as likely as a biological child to influence individuals’ decisions to move from one parity to the next. Stepchildren or other nonbiological children who were not formally adopted are excluded because NSFG wave 4 collected information only on legal adoptions. Furthermore, without information regarding custodial arrangements it would be unclear whether stepchildren and others regularly resided with the respondent.

6 In this version of the paper, marital dissolution is taken into account only in so far as it relates to the time changing married =1 or =0 variable. Thus, it is conflated with never married status, and more rarely, widowhood. We acknowledge that marriage interrupted by divorce is a probable staller of fertility and should be entered into the analysis in its own right. The next version of this paper will incorporate a time changing variable for ever-divorced.

7 By reverse causality we mean it is just as likely that 1) the number of children a respondent has had influences her current employment situation, as it is likely that 2) her current employment situation indicates her past employment which, in turn, influenced fertility outcomes.

8 Current poverty may partially result from the number of children a woman has had, or current poverty may reflect historic poverty, which, in turn, influenced the number of children borne to a respondent.

9 We emphasize that this is an incomplete control in two ways—one, not all sterilization procedures are elected expressly to end fertility, and two, many, probably most, voluntarily childless individuals have neither abortions nor sterilization procedures in their lifetimes. Voluntary childlessness is a problematic definition to pin down because it is rarely a permanent “state” but rather a fluidly changing orientation that is highly dependent on context. It is difficult to discern between women who are “voluntarily childless” versus women who “choose” childlessness because they have delayed too long. Nor can we account for people who are naturally (nonmedically) sterile or subfecund but who never have cause to know this about themselves

10 While one might expect that sterilization would perfectly predict a non-event in the analyses that follow, there are some cases where sterilized individuals adopt.

11 Statistical differences in the bivariate data are subject to error because they are unweighted. We do not weight the data because we have combined three different waves, each with separate weighting parameters.

12 Significant covariance in all of the following survival functions is confirmed by both Wilcoxon and Log-Rank statistics tests.

1   2   3   4   5   6


The database is protected by copyright ©essaydocs.org 2016
send message

    Main page