Category Archives: Research reports

Marriage promotion: That’s some fine print

In a (paywalled) article in the journal Family Relations, Alan Hawkins, Paul Amato, and Andrea Kinghorn, attempt to show that $600 million in marriage promotion money (taken from the welfare program!) has had beneficial effects at the population level. A couple quick comments on the article (see also previous posts on marriage promotion).

After a literature review that is a model of selective and skewed reading of previous research (worth reading just for that), they use state marriage promotion funding levels* in a year- and state-fixed effects model to predict the percentage of the population that is married, divorced, children living with two parents, one parent, nonmarital births, poverty and near-poverty, each in separate models with no control variables, for the years 2000-2010 using the American Community Survey.

To find beneficial effects — no easy task, apparently — they first arbitrarily divided the years into two periods. Here is the rationale for that:

We hypothesized that any HMI [Healthy Marriage Initiative] effects were weaker (or nonexistent) early in the decade (when funding levels were uniformly low) and stronger in the second half of the decade (when funding levels were at their peak).

This doesn’t make sense to me. If funding levels were low and there was no effect in the early period, and then funding levels rose and effects emerged in the later period, then the model for all years should show that funding had an effect. Correct me if I’m wrong, but I don’t think this passes the smell test.

Then they report their beneficial effects, which are significant if you allow them p<.10 as a cutoff, which is kosher under house rules because they had directional hypotheses.

However, then they admit their effects are only significant because they included Washington, DC. That city had per capita funding levels about 9-times the mean (“about $22″ versus “about $2.50″), and had an improving family well-being profile during the period (how much of an outlier DC is on the dependent variables they didn’t discuss, and I don’t have time to show it now, but I reckon it’s pretty extreme, too). To deal with this extreme outlier, they first cut the independent variable in half for DC, bringing it down to about 4.4-times the mean and a third higher then the next most-extreme state, Oklahoma (itself pretty extreme). That change alone cut the number of significant effects down from six to three.

coupdegrace

Then, in the tragic coup de grâce of their own paper, they remove DC from the analysis, and nothing is left. They don’t quite see it that way, however:

But with the District of Columbia excluded from the data (right panel of Table 3), all of the results were reduced to nonsignificance. Once again, most of the regression coefficients in this final analysis were comparable to those in Table 2 (right panel) in direction and magnitude, but they were rendered nonsignificant by a further increase in the size of the standard errors.

Really. What is “comparable in direction and magnitude” mean, exactly? I give you (for free!) the two tables. First, the full model:

tab2

Then, the models with DC rescaled or removed (they’re talking about the comparison between the right-hand panel in both tables):

tab3

Some of the coefficients actually grew in the direction they want with DC gone. But two moved drastically away from the direction of their preferred outcome: the two-parent coefficient is 44% smaller, the poor/near-poor coefficient fell 78%.

Some outlier! As they helpfully explain, “The lack of significance can be explained by the larger standard errors.” In the first adjustment, rescaling DC, all the standard errors at least doubled. And all of the standard errors are at least three-times larger with DC gone. I’m not a medical doctor, but I think it’s fair to say that when removing one case triples your standard errors, your regression model is not feeling well.

One other comment on DC. Any outlier that extreme is a serious problem for regression analysis, obviously. But there is a substantive issue here as well. They feebly attempt to turn the DC results in their favor, by talking about is unique conditions. But what they don’t do is consider the implications of DC’s unique change over this time for their analysis. And that’s what matters in a year- and state-fixed effects model. How did DC change independently of marriage promotion funds? Most importantly, 8% of the population during 2006-2010 was new to town each year. That’s four-times the national average of in-migration in that period. This churning is of course a problem for their analysis, which is trying to measure cumulative effects of program spending in that place — hard to do when so many people moved there after the spending occurred. But it’s also not random churning: the DC population went from 57% Black to 52% Black in just five years. DC is changing, and it’s not because of marriage promotion programs.

Finally, their own attempt at a self-serving conclusion is the most damning:

Despite the limitations, the current study is the most extensive and rigorous investigation to date of the implications of government-supported HMIs for family change at the population level.

Ouch. Oh well. Anyway, please keep giving the programs money, and us money for studying them**:

In sum, the evidence from a variety of studies with different approaches targeting different populations suggests a potential for positive demographic change resulting from funding of [Marriage and Relationship Education] programs, but considerable uncertainty still remains. Given this uncertainty, more research is needed to determine whether these programs are accomplishing their goals and worthy of continued support.

*The link to their data source is broken. They say they got other data by calling around.

**The lead author, Alan Hawkins, has received about $120,000 in funding from various marriage promotion sources.

23 Comments

Filed under Research reports

Brad Wilcox tries to save saving marriage for the marriage movement

Bradford Wilcox and the right-wing family policy community have found a way to make millions of dollars, taking from the welfare budget, to do battle on behalf of the institution of marriage. The premise of their boondoggle is twofold: that increasing the number of marriages will reduce poverty, and that the federal government can accomplish that if it just spends enough of poor single parents’ former money. They’ve gotten the project written into the welfare law. And they have the over-assetted conservative foundations convinced that this is a useful waste of their millions. So they are understandably defensive when social scientists point out that it’s a scam.

 In this guest post, Ohio State University sociologist Kristi Williams responds to Wilcox’s latest commentary.

hands-huckster-cross

By Kristi Williams

In a recent article for the American Enterprise Institute and an op-ed in the Deseret News, W. Bradford Wilcox, director of the National Marriage Project critiques my recent briefing report for the Council on Contemporary Families. My report, “Promoting Marriage among Single Mothers: An Ineffective Weapon in the War on Poverty” discusses the most rigorous experimental evidence available about the effectiveness of federally-funded relationship skills training programs to promote marriage among unmarried parents. The conclusion: They have failed spectacularly.

Wilcox points to one of the programs in Oklahoma as a success. He writes, “Indeed, the Oklahoma Marriage Initiative has succeeded in helping poor, unmarried couples with children enjoy more stable relationships.” Really? After 36 months, participation in the Oklahoma program failed to improve: (a) couples’ relationship quality or the probability of being married (b) the quality of the co-parenting relationship, (c) father involvement and parenting behavior or, most importantly, (c) child poverty and socioemotional development. From the “Building Strong Families” program report:

mathematica null effects

More concerning is the fact that across the 8 program sites included in the study, participation was associated with modest negative effects on father involvement, father financial support of children, and the likelihood that couple would be living together or romantically involved (although they were no more likely to be married). Although children whose parents were in the control group had slightly higher average scores (1.41) on an index of behavior problems and socioemotional development than children of participating parents (1.38), these benefits were only seen in the 4 sites that included home visits and parenting training. Therefore, the report concludes that the modest effect on behavior problems “is more likely due to the home visiting services offered in these 4 BSF sites than it is to the relationship skills education services that were offered in all BSF sites.”

mathematica negative effects

Why does Wilcox call the Oklahoma program a success? There is only one thing he can possibly be talking about: At the 3-year follow up, slightly more children whose parents participated had lived with both parents since birth (49% compared to 41% in the control group).  But what did this get the children? Not lower poverty, not fewer behavior problems and not more father involvement.  This underscores the point of my briefing report: Focusing on keeping low income single parents together at all costs is unlikely to solve the biggest problems facing single mothers and their children.

The only explanation for Wilcox pointing to Oklahoma as a success is that what he really cares about is keeping couples together and promoting marriage at all costs—regardless of whether doing so reduces poverty and helps children and single mothers live better lives.  It’s one thing if you want to preach publicly about the value of marriage from an ideological or religious perspective. But when you claim that you are doing so out of a desire to reduce poverty and you distort the research evidence in order to support your argument, it’s time to omit the Ph.D. from your byline.

The other central argument in Wilcox’s piece is that pointing to the failure of marriage promotion policies is a straw man because no one believes that marriage is a panacea for the problems facing single mothers and their children. But the public dialogue, much of it framed by Wilcox himself, suggests otherwise. One needs only about 5 seconds and a search engine to find Wilcox telling unmarried parents to “put a ring on it” in the New York Times and in public lectures. More troubling, Florida Republican Senator Marco Rubio recently said, “The truth is that the greatest tool to lift people, to lift children and families from poverty, is one that decreases the probability of child poverty by 82 percent. But it isn’t a government program. It’s called marriage.” We could quibble about the meaning of the word, “panacea,” but Wilcox is just wrong when he implies that no one thinks marriage is a central answer to poverty among single mothers. Incidentally, Rubio’s conclusion relies on a fundamental misunderstanding of causality, as described here. Maybe we should forgive Senator Rubio for misunderstanding the data because he is not a trained social scientist. But what is Brad Wilcox’s excuse?

5 Comments

Filed under In the news, Research reports

State of Utah falsely claims same-sex marriage ban makes married, man-woman parenting more likely

This hasn’t been peer-reviewed, but it’s pretty simple, and I will give the results, data, and code to anyone who wants it. Also, ask me about my low-low expert witness rates ($0 per hour + expenses for federal same-sex marriage cases). If you know the Utah lawyers and they’re looking for this kind of thing, pass it on!

The State of Utah’s “Application to Stay Judgment Pending Appeal,” to stop same-sex marriage from continuing while they appeal their most recent loss, has nothing new to offer, legally. And the social science claims they make are by now a familiar patter of discredited blather, featuring the writing of Regnerus, Wilcox, Blankenhorn, and Allen (follow the links for debunking).

But I either never noticed or never thought about one of their stranger claims, which I felt compelled to debunk. They wrote (excerpting):

A final reason to believe there is a strong likelihood this Court will ultimately invalidate the district court’s injunction is the large and growing body of social science research contradicting the central premise of the district court’s due process and equal protection holdings: i.e., its conclusion (Decision at 2) that there is “no rational reason”—much less any compelling reason—for restricting marriage to opposite-sex couples. That research … confirms … (b) that limiting the definition of marriage to man-woman unions, though it cannot guarantee that outcome, substantially increases the likelihood that children will be raised in such an arrangement. (p. 14)

And then again:

[B]y holding up and encouraging man-woman unions as the preferred arrangement in which to raise children, the State can increase the likelihood that any given child will in fact be raised in such an arrangement. … [T]he district court ignored this fundamental reality. … [p. 18] … By contrast, a State that allows same-gender marriage necessarily loses much of its ability to encourage gender complementarity as the preferred parenting arrangement. And it thereby substantially increases the likelihood that any given child will be raised without the everyday influence of his or her biological mother and father—indeed, without the everyday influence of a father or a mother at all. (p. 17)

Wait a minute. Are they claiming that banning same-sex marriage actually results in more children being raised by married, man-woman couples? Unless you make heterogamous marriage and childbearing compulsory, this doesn’t seem like a sure bet. In fact, now that we have so many people living under the same-sex marriage regime, we can start to investigate this.

Does banning gay marriage work to put kids under heterogamously-married roofs?

Seven states plus the District of Columbia permitted legal same-sex marriage by 2012: Washington, New York, New Hampshire, D.C., Iowa, Vermont, Connecticut, and Massachusetts, which led the way in 2004. And as of very recently we have the 2012 American Community Survey, with ample sample size to assess family structure for every state in every year since 2004.

This analysis is very simple and not a causal analysis of family structure. I am simply testing the assertion by the State of Utah that banning gay marriage “can increase the likelihood that any given child will in fact be raised in such an arrangement.” I do this in a very simple way, and then a pretty simple way.

First, just the raw trends. This shows very simply that children are more likely to live with married parents in states that permit same-sex marriage (red lines) than in states that don’t (blue lines):

ssm-married-kidsI did this both for age 0, to capture marital status at birth, and for all children ages 0-14, to get closer to the concept of “raised.” Here is a table showing the numbers, with the differences calculated, showing exactly how much more likely children are to live with married parents if their states permit same-sex marriage:

ssm-married-kids-table

Whatever the reason, then, children in states that permit same-sex marriage have been 2% – 10% more likely to live with married parents over the last decade. (The same-sex couples themselves do not contribute to this pattern, because the public-use ACS files do not yet count them as married.)

Two potential problems with that as the analysis. First, maybe those states were just more pro-marriage places in the first place (the obvious inference to draw from the fact that they permit same-sex marriage). And second, the declining tendency of children to live with married parents nation-wide might be driving this, as more states join the same-sex marriage pool over time.

To fix these problems, I conducted a simple fixed-effects logistic regression, entering dummy variables for every state and every year into a model predicting whether children live with married parents or not. The only other variable indicates whether the child lives in a state that permits same-sex marriage. By holding constant each state’s average rate, and the national trend over time, the model isolates the statistical association with same-sex marriage legal status. This asks, in essence, whether states that change from not-legal same-sex marriage to legal same-sex marriage have lower or higher odds of their children living with married parents after the change.

Here are the results:

ssm-married-kids-logit

The odds ratios for the same-sex marriage variable are above 1.0, indicating the children in same-sex marriage states are more likely to live with married parents. The effect is not statistically significant from zero at conventional levels for infants, but it is for all children ages 0-14. Again, for whatever reason — it’s not important for this — children are more likely to live with married parents if they live in states where same-sex marriage is legal. All that matters is that the State of Utah’s claim is refuted.

Summarizing all the experience we have data for so far — 34 state-years of data — there is no evidence that allowing same-sex marriage reduces the likelihood that children will be born to or live with married, man-woman parents. If that’s your goal, this policy doesn’t seem to work. (I don’t share that goal, and I especially don’t think it’s relevant to determining legal access to marriage, but they brought it up.)

I’m not the first one to think of this, of course. An earlier analysis in PLoS One found no evidence that same-sex marriage affects the rate of different-sex marriage. That analysis was of marriage, and its most recent data were from 2009. I haven’t seen anyone else do this for children’s living arrangements, and the 2012 only recently became available. If Gary Gates or someone else has done this, please let me know.

6 Comments

Filed under Research reports

Divorce recession drop rebound, with the 2012 rate

Note: Technical addendum added.

The Census Bureau’s American Community Survey is the best annual national data source for marital events. The 2012 data came out recently, and I don’t believe anyone else has published a divorce rate for 2012. The refined divorce rate – the number of divorces per 1,000 married people – was 19.0 in 2012. Here is the trend since the ACS starting counting divorces:

divrat08-12

What does this mean? It’s a shame the ACS didn’t start counting marital events till 2008, because it means we can’t put that year’s high rate in context. Was it (a) a spike up, suggesting divorce was a leading indicator for the recession; (b) part of a consistent decline, suggesting the the years since have been a pretty substantial increase from the historical trend; or, (c) a data anomaly.*

To put this in the context of the larger trend doesn’t really help answer the question, since we switched from vital records to a national survey, and had a decade with no national statistics in between:

divrate40-12

So, it’s a mystery. My preferred interpretation is still that the recession caused a decline in divorces because disgruntled people were tied up in other crises, couldn’t sell their houses, or couldn’t afford to move out, followed by a rebound of accumulated divorces to our current level.

I published a working paper suggesting this [now forthcoming in Population Research and Policy Review], in which I use 2008 predictors of divorce and estimate that 4% fewer divorces occurred through 2011 compared to what would have been expected had the determinants of divorce not changed in the subsequent years.

My blog series on divorce includes previous reports on rates, and attempts to predict divorce rates using Google searches.

Technical addendum

To replicate my rates for 2012, you start here at the FactFinder, then get the number of married people by sex (ACS Table B12001) and the number of people who got divorced in the 12 months before the survey (ACS Table S1251) — you can enter the table numbers into the search box. There is a slight problem with this, however. Some people who say they got divorced in the past 12 months also say they are currently married (presumably remarried already). Those people are counted twice in the denominator of the FactFinder-based divorce rate — once as divorced people and once as currently married. If you download the public-use file and count those people only once in the denominator, the divorce rate rises by .02 per 1,000 (or 2 people per 100,000) — but this would not change the figures above at the level of precision reported. However, the public-use files produce slightly different estimates than the FactFinder files anyway, because the latter are based on the Census Bureau’s complete file not a subsample, so I use those even though they produce this tiny under-estimate of the divorce rate.

Secondly, what about the difference in divorce rates between men and women? This is a survey, not a vital records count, and there is no way to verify with the now-missing spouses whether they also consider themselves divorced. Maybe they weren’t legally married, or they didn’t really get legally divorced. So there are several possibilities: (a) lots of lesbian divorces, which is unlikely given the small number of lesbian marriages (but note we don’t know the sex of the spouse who is no longer in the household so we can’t distinguish homogamous from heterogamous divorces); (b) women are more likely to describe a breakup as a divorce for reasons unknown; (c) something funky with the survey weights (unweighted divorce rates from the public-use file also show the disparity, but it’s 20% smaller), or; (d) something funky with the sampling.

Who knows! If you are reading this and considering a new career — or a new direction in your existing career — consider becoming a family demographer and helping us figure it out.

9 Comments

Filed under In the news, Research reports

Sociology citing Becker

Which comes first, the Nobel prize or the citations in sociology journals?

Neal Caren produced a list of the 52 works most cited in sociology journals in 2013, which included two Nobel prize winning economists:

  • Heckman, James J. “Sample selection bias as a specification error.” Econometrica: Journal of the econometric society (1979): 153-161.
  • Gary S. Becker. A Treatise on the Family. Harvard university press, 1981.

I assume those Heckman citations are the result of sadistic journal reviewers or dissertation committee members impressive their colleagues by requiring people to add selection corrections to their regressions.

The Becker citations were applauded by economists. I assumed they were usually cursory mentions in the literature review, representing neoclassical economics in the study of families. And that is basically right. In the 10 most recent citations to Treatise in top-three sociology journals, the book is always mentioned only once. See for yourself. Here are the passages out of context (citations at the end):

  1. A great deal of work in sociological theory addresses the determinants of marriage and the bases of divorce. Some of this work posits marriage as a form of social exchange, whereby internal benefits (sex) and costs (time) are calculated and weighed relative to external costs (money) and benefits (social approval) (Becker 1991).

  2. According to the negotiation framework known as intra-household bargaining (Agarwal 1997), rather than households behaving as cohesive units (Becker 1991), household members’ bargaining and decision making over the allocation of resources (e.g., income, health, education, time use) are conditioned by gender-based power differentials.

  3. In the classic economic and game theoretic models of partner matching and mate selection (Becker 1991; Gale and Shapley 1962), the relative value of every potential mate is assumed to be already known or can easily be determined (Todd and Miller 1999).

  4. Generally used to explain behavior during the waking hours, the time availability perspective suggests that because men spend more time in paid work, they have less time to do caregiving; the related specialization hypothesis suggests that women have the time and incentive to specialize in caregiving and unpaid work (Becker 1991[1981]).

  5. A second means by which household wealth is accrued is by means of family transfers. Economic assets, whether financial or real, are transferred from family members to others, both within and across generations (Becker 1991; Mulligan 1997; Wahl 2002).

  6. The compensating differentials argument suggests that mothers are more willing than non-mothers to trade wages for family-friendly employment. For example, Becker (1991) suggests that mothers may choose jobs that require less energy or that have parent-friendly characteristics, such as flexible hours, few demands for travel or nonstandard shifts, or on-site daycare.

  7. Differences in life course patterns between men and women may reflect the influences of traditional gender roles in the family and corresponding intermittent labor force attachment among women relative to men, particularly during childbearing years (Becker 1991; Bianchi 1995; Mincer and Polachek 1974).

  8. One of the primary ways in which education leads to lower fertility is by changing the calculation of the costs and benefits of childbearing and rearing (Becker 1991).

  9. As has long been recognized in both economics and sociology, an adequate explanation of gender inequality in the labor force therefore requires the researcher to go beyond discrimination and productivity-related attributes (i.e., human capital) and to consider the role of the family (Becker 1973, 1974, 1991; Mincer and Polachek 1974; many others). … First, it is assumed that economic resources are a family-level utility that is shared equally between the spouses (Becker 1973, 1974, 1991; Lundberg and Pollak 1993; Mincer and Polachek 1974).

  10. Fathers’ economic contributions are an important resource for children in all types of families (Becker 1991; Coleman 1988).

I noticed, incidentally, that we may have hit Peak Becker. The Web of Science citation count for his work in journals coded as Sociology peaked in 2011. Maybe the 2012 data just aren’t complete yet.

peak-becker

Out of curiosity, I also checked the citations in major economics journals to the most highly-cited sociology article on the household division of labor known for a theoretical argument, Julie Brines’s 1994 article in the American Journal of Sociology. Just kidding; there aren’t any.

No, that’s not true. The article has been cited once in the top 40 economics journals, in Transportation Research Part A: Policy and Practice:

The higher wage earner enjoys a superior bargaining position, and thus can use that power to demand less household responsibility – a proposition that has been the focus of substantial empirical research among sociologists (Heer, 1963, Brines, 1994, Greenstein, 2000, Bittman et al., 2003, Parkman, 2004 and Gupta, 2007).

References

  1. Rose McDermott. and James H. Fowler. and Nicholas A. Christakis. “Breaking Up Is Hard to Do, Unless Everyone Else Is Doing It Too: Social Network Effects on Divorce in a Longitudinal Sample.” Social Forces 92.2 (2013): 491-519.
  2. Greta Friedemann-Sánchez. and Rodrigo Lovatón. “Intimate Partner Violence in Colombia: Who Is at Risk?” Social Forces 91.2 (2012): 663-688
  3. Michael J. Rosenfeld and Reuben J. Thomas. 2012. Searching for a Mate: The Rise of the Internet as a Social Intermediary. American Sociological Review August 2012 77: 523-547. doi:10.1177/0003122412448050
  4. Sarah A. Burgard. “The Needs of Others: Gender and Sleep Interruptions for Caregivers.” Social Forces 89.4 (2011): 1189-1215.
  5. Moshe Semyonov. and Noah Lewin-Epstein. “Wealth Inequality: Ethnic Disparities in Israeli Society.” Social Forces 89.3 (2011): 935-959.
  6. Michelle J. Budig and Melissa J. Hodges. 2010. Differences in Disadvantage: Variation in the Motherhood Penalty across White Women’s Earnings Distribution. American Sociological Review October 2010 75: 705-728, doi:10.1177/0003122410381593.
  7. Jennie E. Brand and Yu Xie. 2010. Who Benefits Most from College?: Evidence for Negative Selection in Heterogeneous Economic Returns to Higher Education. American Sociological Review April 2010 75: 273-302, doi:10.1177/0003122410363567.
  8. Brienna Perelli-Harris. “Family Formation in Post-Soviet Ukraine: Changing Effects of Education in a Period of Rapid Social Change.” Social Forces 87.2 (2008): 767-794.
  9. Emily Greenman. and Yu Xie. “Double Jeopardy?: The Interaction of Gender and Race on Earnings in the United States.” Social Forces 86.3 (2008): 1217-1244.
  10. Daniel N. Hawkins, Paul R. Amato, and Valarie King. 2007. Nonresident Father Involvement and Adolescent Well-Being: Father Effects or Child Effects? American Sociological Review December 72: 990-1010, doi:10.1177/000312240707200607.
  11. Sirui Liu, Pamela Murray-Tuite, Lisa Schweitzer. 2012. Analysis of child pick-up during daily routines and for daytime no-notice evacuations, Transportation Research Part A: Policy and Practice, Volume 46, Issue 1, Pages 48-67.

4 Comments

Filed under Research reports

Inequality, mobility, single mothers, and race: comment

I have no idea whether inequality increases intergenerational immobility. But I do know that lots of people would like to pin bad social trends on single motherhood, meaning — in their view — the bad decisions of people who already poor. And that has bad implications.

In a blog post by Scott Winship and Donald Schneider at the Manhattan Institute, they argue that the liberal argument that inequality blocks mobility is not well supported. To do that, they show simple bivariate correlations between single motherhood rates and immobility across U.S. labor markets. Their point is that, if you want to use that simple bivariate standard, you can just as well — but better — argue that immobility is caused by single motherhood rather than by income inequality, because the correlation is very strong. For their exercise they use data from the Equality of Opportunity Project, which is freely available here.

In a series of tweets, Winship clarified his point:

point wasn’t to highlight single parenthood—point was to show where low evidentiary standards on left can take you … look, single motherhood may very well be a big problem for mobility. Inequality might too…. but the left has to be held accountable when they make bad arguments skewing policy debates…  I clearly wrote that correlations shouldn’t constitute reason for getting worked up about single moms

I take him at his word on his intentions, but those with well-documented patterns of less scrupulous behavior are not so scrupulous, and so the post was bad. Despite a disclaimer about not reading causation from correlation, they also wrote:

In other words, a [labor market’s] prevalence of single motherhood predicts its relative mobility quite well all by itself. … the relationship between single motherhood and mobility holds up in all of these analyses. … On the basis of these charts, rather than a new Washington Center on Equitable Growth housed at CAP and devoted to discovering the damages that income inequality inflicts, the left should have started a Washington Center on Single Motherhood.

Again, my only dog in the fight is fighting against the easy right-wing causal association of single motherhood with bad outcomes. The Heritage Foundation, Scheider’s employer, is particularly egregious in this, as I’ve occasionally documented (here and here, e.g.)

So here’s a quick debunk on that. A simple glance at the map from the Equal Opportunity Project will tell you that race is involved here, but it didn’t come up in Winship and Schneider’s post:

immobilitymap

So let’s just look at the relationship between immobility, single motherhood and race. (Immobility here is measured by the effect of family income on children’s incomes. Higher scores are bad.)

So first, here is the relationship between population percent Black and immobility for the 100 largest metro areas, with the larger ones shown as bigger dots:

pb-immobThat relationship is quite strong: the higher Black population proportions are strongly associated with immobility. But so is the single motherhood relationship, as Winship and Schneider reported. So, we turn to the obvious tool, a multivariate regression. Here are two models, the first with just single motherhood — in effect, the Winship and Schneider result — and then a model with proportion Black added. Both are weighted by population size.

pb-immob-reg

This shows that the association between single motherhood rates and immobility is reduced by two-thirds, and is no longer significant at conventional levels, when percent Black is added to the model. That is: Percent Black statistically explains the relationship between single motherhood and intergenerational immobility across U.S. labor markets.

This is not a rigorous examination of the cause of intergenerational immobility. It is just debunking one bivariate story that is too easily picked up by the forces of bad.

 

 

7 Comments

Filed under Research reports

Gender devaluation, in one comparison

You can divide the reasons women earn less money than men do, on average, into three categories, in declining order or importance:

  1. Working fewer years, weeks, and hours
  2. Working in different occupations
  3. Being paid less in the same occupations

The first has to do with families and children. That has a large voluntary, or at least kind of voluntary, component (or it reflects hiring discrimination, which is hard to prove, prevent or punish under our legal regime). The third is illegal and sometimes actionable, as in the Lilly Ledbetter situation.

The second — occupational segregation — is a difficult hybrid. Segregation reflects both discrimination in hiring and promotions, and socialization-related choices, including in education. And it is wrapped up with divisions that may even be relatively harmless in a separate-but-equal kind of way — that is, not directly harmful, but contributing to the categorical divisions that make gender inequality more intractable. But the different pay in female- versus male-dominated occupations is a problem, well documented (see here and here) but virtually impossible to address under current law.

nurse-truck

Today’s example: nursing assistants versus light truck drivers

The government’s O*Net job classification system provides detailed descriptions of the qualifications, skills, and conditions of hundreds of occupations. The comparison between nursing assistants (1.5 million workers) and light truck or delivery services drivers (.9 million) is instructive for the question of gender composition. Using the 2009-2011 American Community Survey, I figure nursing assistants are 88% female, compared with 6% female for the light truck drivers. Here are some other facts:

  • The nursing assistants are better educated on average, with only 50% having no education beyond high school, compared with 67% of the light truck drivers.
  • But in terms of job skills, they are both in the O*Net “Job Zone Two,” with 3 months to 1 year of training “required by a typical worker to learn the techniques, acquire the information, and develop the facility needed for average performance in a specific job-worker situation.”
  • The O*Net reported median wage for 2012 was $11.74 for nursing assistants, compared with $14.13 for light truck drivers, so nursing assistants earn 83% of light truck drivers’ hourly earnings.

To make a stricter apples-to-apples comparison, I took those workers from the two occupations who fit these narrow criteria in 2009-2011:

  • Age 20-29
  • High school graduate with no further education
  • Employed 50-52 weeks in the previous year, with usual hours of exactly 40 per week
  • Never married, no children

This gave me 748 light truck drivers and 693 nursing assistants, with median annual earnings of $22,564 and $20,000, respectively — the light truck drivers earn 13% more. Why?

The typical argument for heavy truck drivers’ higher pay is that they spend a lot of time on the road away from home. But that’s not the case with the light truck drivers. They are more likely to work longer hours, but I restricted this comparison to 40-hour workers only. Here are comparisons of the O*Net database scores for abilities and conditions of the two jobs. For each I calculated score differences, so the qualities with bars above zero have higher scores for nursing assistants and those with bars below zero have higher scores for light truck drivers. See what you think (click to enlarge the figures). My comments are below.

abilities

context

You can stare at these lists and see which skills should be rewarded more, or which conditions compensated more. Or you could derive some formula based on the pay of the hundreds of occupations, to see which skills or conditions “the market” values more. But you will not be able to divine a fair market value for these differences that doesn’t have gender composition already baked into it. And “the market” doesn’t make this comparison directly, because nursing assistants and light truck drivers generally don’t work for the same employers or hire from the same labor pools. You might see reasons in these lists for why women choose one occupation and men choose the other, but I don’t see how that fairly leads to a pay difference.

The only solution I know of to the problem of unequal pay according to gender composition is government wage scales according to a “comparable worth” scheme (the subject of old books by Joan Acker and Paula England, but not high on the current political agenda). Under our current legal regime no one woman, or class of women, can successfully bring a suit to challenge this disparity.* That means occupational integration might be the best way to break this down.

*One exception to this is the public sector in Minnesota, in which local jurisdictions have their pay structures reviewed at regular intervals for evidence of gender bias, based on the required conditions and abilities of their jobs (as reported by me by Patricia Tanji of the Pay Equity Coalition of Minnesota).

19 Comments

Filed under Research reports