Archive for the ‘Statistics’ Category

You read it here first

Thursday, October 1st, 2009

You knew this would pull me out of my blog hiatus, right?  The Washington Post has a front page story today that highlights the new Census findings that "stay-at-home moms" are typically younger and in less affluent families.  Thus, the media obsession with highly compensated professional women who drop out of the workforce to be full-time parents is "largely beside the point."

Except that, this isn't such a new point.  I pointed it out — based on Census data — oh, about four years ago.  It wasn't so neatly packaged then, but the underlying data has been available for years and doesn't look like it's changed very much.  (I'll be interested in seeing what the figures look like in a couple of years, when they're available for the recession period, but this is 2007 data, pre-recession.

There is some new info in the report from the American Community Survey, mostly about the geographic distribution of families — previous tables were only based on the Current Population Survey, which doesn't have representative samples of small stats.  Check out Figure 8, on page 15.  Overall, Northern and Eastern states have a higher percentage of two-income married parent families that average, Western states have a lower percentage.  But New York is lower than average.  I'd love to see the NY-specific data on income — New York is  a big enough state that it's probably possible to do the run without having huge margins of error.

WBR: Intelligence and How to Get It

Wednesday, April 29th, 2009

As promised, here's a review of Richard Nisbett's Intelligence and How to Get It: Why Schools and Cultures Count.  it's the book that Nicholas Kristof's column a couple of weeks ago was based on.  The book jacket describes this book as "the authoritative anti-Bell Curve" and indeed, much of the book is  a full-out attack on the claim that intelligence is primarily determined by genetics and that any attempts to improve outcomes for members of disadvantaged groups are doomed.

To be honest, the "how to get it" part was the least interesting part of the book for me, because it covered ground that I already know about — Perry Preschool, KIPP, Carol Dweck's work on the "mindset" that effort matters more than inherent ability.  That said, Nisbett does a good job of writing about these issues in a non-technical manner, and I'm hopeful that he will influence public opinion.

The "intelligence" part of the book was far more interesting, because Nisbett is implicitly arguing with both the strong hereditarians who believe that intelligence is overwhelmingly genetic and that environment (including parenting) doesn't matter much, and with the liberals who aren't sure exactly what is meant by "intelligence," and are pretty skeptical that intelligence tests are picking up underlying ability rather than leaning.  The first two chapters (and a more technical appendix) are aimed squarely at these issues, and should be mandatory reading for anyone who wants to talk about intelligence.

Nisbett argues that the high estimates for the genetic component of intelligence are overwhelmingly based on twin studies, and especially adoptive studies, and these don't haver nearly as much variation in environments as there exists between families overall.  He also notes that overall IQ levels have risen steadily over time, far too quickly to be accounted for by natural selection (if you look at the raw scores, rather than the normed ones which are forced to have a constant mean of 100).  Addressing the question of racial differences in IQ specifically, he points out that the black-white gap has also decreased significantly in the past decades, and that African-Americans with a higher percentage of European genes do not have higher IQs than African-Americans with fewer European genes.

I'm going to end this review where Nisbett begins the book, on the question of what is intelligence.  Even after reading the book, I find it hard to define.  Nisbett is clear that he believes that schooling does increase intelligence, and that scores on even the most abstract and supposedly culture-free components of the IQ test (such as the Raven progressive matrices*) improve markedly with practice.  So he doesn't agree with the opening quote from Cyril Burt that intelligence is "inborn, all-around intellectual ability.. inherited, not due to teaching or training… uninfluenced by industry or zeal."  But he also thinks it's a real characteristic, distinct from specific knowledge of a subject.  In some ways, he almost seems to define intelligence as that which is measured by IQ tests, which is a strong predictor of academic and career sucess although not the only factor in either (with effort, emotional skills, self-discipline, and motivation being the strongest non-intelligence factors in these).

* For what it's worth, I would have chosen a different answer than the "correct" one on the sample problem given in the book, and still think that my answer is equally plausible.

graphing the tax plans

Wednesday, September 17th, 2008

Via Bitch PhD and Yglesias, this terrific graph showing the Obama and McCain tax plans and how much they’ll affect different income brackets’ taxes, with the bands scaled to reflect the number of people affected:

This is from a site called chartjunk, which attempts to use Edward Tufte‘s principles in designing charts.  Definitely lots to learn from.

The Freakonomics blog at the NY Times picked up on this too.

While we’re on the topic of taxes and distribution, I’ll point out that the big tax cut bill coming out of the Senate, which includes both the Alternative Minimum Tax patch and a bunch of business tax extensions, does include one provision that is really important for low-income families: allowing families to start to receive the child tax credit starting at an income of $8,500, down from the $12,050 under current law.  This would help 13 million low-income children.  It’s not at all guaranteed that the House bill will also include this provision, so it’s worth dropping a line or calling your representative.

I’m going to use the blogger’s prerogative to add this to the post, rather than risking having it get buried in the comments with all the back and forth about child support.

Maria commented on the stat that’s shown in the third chart on the Freakonomics post — that the top .1 percent of the country pays 20 percent of the income tax.  I haven’t seen that elsewhere, but it seems plausible.  There are great statistics on US income and wealth inequality here.

It’s worth noting that for all conservatives in the US mutter about European socialism, the tax system in almost all European countries is far less progressive than the US system, because they collect a large portion of their government funding through a Value Added Tax (VAT) which is if anything, somewhat regressive.


Composition of the US Labor Force by Marriage and Parenting Status

Friday, September 5th, 2008

Here’s what I’ve been working on this week:


This is pretty different from the usual way these numbers are presented, which is based on families rather than workers.  (Remember, if half of the families with children have an at-home spouse and the other half is dual income, only 1/3 of the workers will have an at-home spouse.)

For what it’s worth, the furthest back I was able to come up with
roughly comparable numbers for is 1975, when 41.5 percent of the
workforce were parents, and 35 percent of the working parents had an
at-home spouse.


I’d love some feedback on these graphs — what interests you?  Surprises you?  Is the second one too many slices to be easily interpreted?

Update:  I’m responding in the comments. But I also want to register my fury that Microsoft in Excel 2007 has made it impossible to apply patterns to different slices on a pie chart so that you can tell them apart when you print them in black and white.

Update 2: Ok, here’s one that shows part-time vs. full-time.


Homeownership rates

Monday, May 19th, 2008

When I posted about whether young people are "falling behind" their parents, almost all of the commenters agreed that a big part of the reason that even relatively affluent young adults *feel* poor is that homeownership seems so out of reach (even with the declining market).  This made a lot of intuitive sense to me.

But my dad then sent me a ton of Census data on homeownership rates by age, going back to 1982.* (Yes, I come by my geekery honestly.)  And his point is that households under age 35 were just about as likely to own homes in 2008 (41.7 percent) as in 1982 (41.2 percent).   Homeownership rates for this group hit a low of 37.3 percent in 1993-1994, and then rose to 43.1 percent in 2004, before falling off slightly.

So how is it possible that homeownership can feel so out of reach to almost everyone I know, even as the homeownership rate didn’t decline at all?  Well, part of the answer is that I live in an expensive housing market, so the "everyone" I know is a biased sample.  (The readers of this blog are more diverse, but I think are still disproportionately living in large urban areas, compared to the country as a whole.)   Also, a whole lot of condos were built in the 1990s, so if by "homeownership" you mean "owning a single family detached home," the homeownership rate probably did decline somewhat.

But it’s also true that a lot of people — at all age groups — bought homes only by extending themselves to their limits.  There was this credit bubble that you might have heard about… (Supposedly in 2005, half of all loans made in DC were interest-only.)  And there was this dreadful fear that if you didn’t jump in right away, even if you couldn’t really afford it, you’d be priced out forever.  So, the people who didn’t buy houses felt like they were falling behind because they couldn’t afford a home, and the people who did felt like they were falling behind because they couldn’t afford anything else.

*The Census table is only online as a text file — if you want my Dad’s Excel spreadsheet, I’m happy to send it on.

My father is right

Wednesday, April 9th, 2008

After reading yesterday’s post, my dad emailed me to say that he thought the question of whether young adults are better off than their parents depends mostly on what level of education each generation has attained.  Specifically, he argued that a young adult with a college degree is likely to be better off than her parents if she’s a first generation college student, but not if her parents also went to college.

Let’s look at the possibilities. 

  • If your parents went to college, and you went to college, they are probably earning more than you are.  (Obviously, there are exceptions when the parent suffers from a disability, or chose to be a starving artist, or got laid off, or when the kid joined Google or Microsoft at just the right time, but on average, 55 year old college grads earn a lot more than 25 year old college grads.  To be precise, in 2006, the average 25-34 year old with a bachelor’s degree in 2006 earned $40,276 and the average 55-64 year old with bachelor’s degree earned $50,397. 
  • I wasn’t convinced that young college graduates were necessarily earning more than their non-graduate parents but I looked up the numbers, and my father is right.  The average 55-64 year old with a high school degree and no college education earned  $29,283 in 2006.  While there are some plumbers and union mechanics who earn good money with just a high school degree, there’s not enough of them to affect the median.
  • Young high school graduates are also earning less than their HS-grad parents — the average 25-35 year old with a high school degree and no college earned just $25,0354.
  • And, to fill out the options, the HS grad child of college-graduate parents is clearly downwardly mobile.

[Sources PINC-03-part 37 and PINC-03-part 91.  All figures cited are medians.]

My dad’s point was that because the fraction of the population going to college has increased so much, a significant portion of college graduates are from families where their parents didn’t go to college. And they’re doing better than their parents.  At least in terms of income — they also have more college debt. And, as lots of people commented yesterday, their parents probably own a home that has appreciated significantly since they bought it, while in a lot of the country, homeownership is still out of reach for most young people, even those with good incomes.

Also, check out Figure 4 in this report.

Mothers labor force participation

Monday, August 13th, 2007

Here’s something that I pulled together at work, and then wound up cutting from the document I did it for.  So I thought I’d share it here. 

This chart (from the new Indicators of Welfare Dependence report, issued by my old friends at HHS) shows the trends in labor force participation of married vs. divorced/separated/widowed vs. never-married mothers over the past 30 years.

labor force participation of mothers by marital status

I think it’s pretty remarkable how sharply the line for the never married mothers goes up in the 1990s.  So, what’s going on here?

Before turning to the question of why never married mothers labor force participation (LFP) rose so much during the 1990s, it’s first necessary to consider why it didn’t rise before the 1990s.  Another way to think of this question is to ask why did the labor force participation of married mothers rise during this period, and why didn’t the same factors increase the labor force participation rates of never married mothers (at least until the 1990s).

  • One reason that labor force participation rates increased for married women is that women now have greater potential wages, which make paid labor more attractive.  Women are both more educated and more experienced than they used to be, and blatant labor market discrimination is far less common, opening many lucrative career options to women.
  • However, these economic explanations only go so far; a key part of the story is changing societal norms that have made continued employment by married mothers, regardless of economic need, far more common.  As Blank and Shierholz comment, the effect of marriage itself on women’s labor force participation “virtually disappeared over time.”

What about divorced mothers?

  • Same arguments as for married mothers, plus:
  • Lack of alternative resources makes for lower reservation wages.
  • Among more skilled women, single parenting has a positive effect on labor supply – true in both 1979 and 2003 (Blank and Shierholz)

So why didn’t the LFP for never married mothers rise in the 1980s?

  • On average, younger, less educated than divorced mothers, so potential wages are much lower – may not equal the cost of child care or other lost benefits.
  • In addition, the “child penalty” on LFP rate is higher for younger mothers, and less educated mothers, even when children are the same age (Boushey)
  • Welfare provided a meager alternative to low-wage work – not a great living standard, but possible to eek by.  Kathy Edin’s work showed that low-wage work often didn’t provide any more disposable income.
  • Welfare policies provided large incentive to keep all earnings off the books.
  • In 1979, but not 2003, less skilled single moms were less likely than comparable childless women to work (Blank and Shierholz) – may be capturing the effects of welfare policy

What happened in the 1990s?

  • Strong economy led to employment expansions for most low-income workers – male and female, parents and non-parents.
  • EITC expansion greatly increased the returns to work in the formal sector for low-income parents – studies have shown that the effect was concentrated on single mothers.
  • Time limits and work requirements largely removed the alternative of choosing full-time parenting over low-wage work for welfare recipients, even for parents of young children. Just between 1996 and 1999, the employment rate for single mothers under 200 percent of poverty with a child under the age of 6 increased from 44.4 percent to 58.5 percent.  (TANF 7th annual report, page IV-33).
  • Work supports reduced the cost of going to work – child care, SCHIP, expanded earnings disregards.
  • Rate of increase in LFP did increase for divorced women, but not as sharply as never-married women.
  • Some of the increase in employment among never married mothers is likely due to composition effects – with declining teen birth rates, increases of overall non-marital birth rates, never married mothers are more likely to be older, more educated.

And in the 2000s?

  • Weaker economy reduced both employment and LFP for all types of workers
  • But married women’s LFP peaked in 1997, when economy was still booming – suggests that recession isn’t the whole story .  May be due to substitution with husbands’ earnings (although women’s LFP has become far less affected by husband’s earnings over time.)  (Blau and Kahn)
  • Divorced women’s LFP peaked in 2001; never-married women’s LFP in 2002; single mothers under 200 percent of poverty in 2000.

Data geek heaven

Friday, July 13th, 2007

Via Kameron at Brutal Woman, I found this presentation on how much of what we think we know about "the third world" is wrong.  Specifically, it points out how much Asia and, to a lesser degree, South America have converged with Europe and Northern America with respect to factors like infant mortality and life expectancy, leaving Africa behind.  But the use of graphics is the attention getter.

So, I was quite pleased to discover that the data tool, gapminder, that Rosling uses is now available for anyone to play with.  You can pick from a bunch of pre-loaded data sets, and the site promises that there will eventually be the ability to add your own data.  Data geek heaven.

On a somewhat related note, does anyone know if there’s a way to condition the color of a bar in an Excel bar graph on the value of a boolean value of another data series?  I wound up doing it by hand for the presentation I was working on today, but there ought to be a way to it automatically?

Stats on parenting and class

Monday, March 19th, 2007

Poking around the Census web page today, I ran across this report, issued earlier this year, on A Child’s Day, 2003 (Selected Indicators of Child Well-Being).

It’s full of all sorts of odd and interesting statistic, like 6.7 percent of parents living with a child 12-17 said that they talked to or played with their child for 5 minutes "never" to "once a week."  What really jumped out at me is the ability to see what parental characteristics are associated with different parenting behaviors.  Affluent parents are more likely to report  reading to their preschool aged children than poor parents (although 40 percent of poor parents still said that they read to their kids 7 or more times in the last week).  The association with parental education is even stronger than with income.

I was quite struck by the correlation they found between "television rules" imposed on children (restricting the type of programs, the time of day, or the number of hours watched) and the frequency with which parents read to their kids.   This suggests at least the  possibility that the supposed negative effects of television on young children is a spurious correlation with parenting behaviors.

Consistent with Lareau’s description of concerted cultivation vs. accomplishment of natural growth, more affluent and more educated parents were far more likely to report that their school-age children participated in extra-curricular activities, including sports, clubs, and classes.  (There was no "egghead effect" — children of parents with post-baccalaureate degrees were still more likely to play sports than any other kind of activity.)  And the higher level of education the parents have, the more likely their children are to participate in gifted classes, and the less likely the children are to have been suspended or to repeat a grade.

The endless to-do list

Friday, March 10th, 2006

I’ve been thinking about that NYTimes article on mother’s labor force participation.  The article suggests that the slight recent drop-off in women’s labor force participation in recent years is because we’ve pushed unpaid work — housework and child care — about to its lower limit, and there are only so many hours in the day and something has to give. 

Bitch, PhD thinks that makes sense.  She wrote:

if, broadly speaking, we’ve wrung about all we can out of the 24 hours in a day, then it makes sense both that some women would step back from the grueling regime in favor of a more balanced personal life, regardless of the possible risks they run in doing so: when you’ve reached the limit of your energy, you can’t keep going and that’s all there is to it. It also makes sense that women who are still trying to hang onto the stressful balancing act of career, children, and coupledom would feel that they’re singlehandedly carrying the world on their shoulders. And given the pressures on all of us, of course we’re all defensive and insistent and argumentative about our choices.

But one of her commenters, Steve Horwitz, points to this Economist article (based on this paper by Aguilar and Hurst) which uses the same underlying data as the Times article and comes to the conclusion that total leisure time for all groups — including working moms — has increased significantly over the past 40 years.  Is this possible?  And if it’s true, why do we all feel so tired?

I think there’s a bunch of different things going on.

If I’m reading the papers accurately, the biggest issue is whether you consider time spent with children doing generally recreational activities — reading to them, taking them to parties, watching school plays, even going to the park — as leisure.  Aguilar and Hurst do, while I think Bianchi (whose data the NYTimes uses) counts them as child care.  Conceptually, I think these activities somewhere between true leisure and work.  They’re not in the same category as changing diapers or attending parent-teacher conferences, which you do because they’re important, but no one really considers fun.  But they’re also at least semi-obligatory —  you feel guilty if you don’t do them enough, and you often have to do them even if you’d really rather be doing something else.  So they add to the modern parent’s endless to-do list.

While the time-use studies clearly show that the amount of time spent on housework has dropped significantly, they don’t account for the fact that people’s expectations  haven’t fallen as much.  So even if we only vaccuum once a month, we feel like we ought to do it more often, and it stays on our to-do list, even if we know that we’re never going to get to it.

Aguilar and Hurst also point out that there’s been an increase in inequality in leisure time, with more of the gain in leisure concentrated among less educated individuals.  If you believe Annette Lareau, the parents with more education are also spending more of their "free" time in intensive parenting activities.  And if you’re reading this blog, or Dr B’s, the chances are high that you’re in that group.

As the Economist article acknowledges, the blurring of the lines between work and free time are also a factor in our perception of overwork.  If you have to carry a blackberry to your kid’s soccer game, and check your voice mail over the weekend, it’s hard to leave the office behind.  And I don’t think it’s coincidence that Dr. B and Sandy Piderit are academics.  It’s not just that professors work long hours, but that their hours of work are unbounded — there’s almost always something else that they could/should be working on.

Overall, I think it’s that sense of things left undone, rather than the total number of hours worked, that makes people feel overwhelmed.  When I started work after getting my masters, I remember how excited I was at the concept of the weekend.  Look, it’s Friday, and I get to go home!  And I don’t have to think about work, or feel guilty about not doing it, until Monday morning!  What a concept.

But at this point in my life, my personal to-do list is a lot longer than my work one.  Some days are busier than others at work, but I generally leave the office having accomplished most of what I need to do.  At home, I almost always feel like I’m running behind.   Therefore, I need to make a conscious choice at times to let go of the endless to-list.

Or, as Abraham Joshua Heschel wrote:

"The meaning of the Sabbath is to celebrate time rather than space. Six days a week we live under the tyranny of things of space; on the Sabbath we try to become attuned to holiness in time. It is a day on which we are called upon to share in what is eternal in time, to turn from the results of creation to the mystery of creation, from the world of creation to the creation of the world."

Shabbat Shalom.