Teen Births and Violent Crime — a Weak Connection

I’ll be the first to admit, giving me an Excel spreadsheet is the intellectual equivalent of handing a chimp a machine gun. What I don’t know about statistics would, well… it would fill a statistics textbook. But I abuse statistics less than most journalists, commentators, and politicians, who, to paraphrase renowned economist Ronald H. Coase, routinely torture the data until it confesses. I count on readers to call B.S. when they see it and modify my findings accordingly. In the spirit of exploration and with all due humility, I present the following:

In a previous post, I disputed the conventional wisdom that “poverty” is a “root cause” of violent crime. The lack of income and material resources is undoubtedly a contributing factor, playing into feedback loops of tremendous complexity, but overall the correlation between the poverty rate and the crime rate across Virginia’s 100+ localities is weak — an R² of .1802, which is considered a small effect size. There is a much stronger correlation — an R² of .4007 across Virginia localities, a moderate effect size — between the percentage of single-parent households and violent crime.

If the percentage of single-family households in a population has a moderate influence on crime, I wondered about the percentage of teen births.

Is it possible that young mothers in contemporary society lack the parenting skills and resources to keep their children out of trouble? I crunched the numbers using data from the County Health Rankings and Roadmaps, and the results can be seen above. The verdict: There is an effect, but it is a weak one.

Conservative views of crime place greater emphasis on behavioral issues unrelated to income. In theory, the responsibility of raising a child places a huge stress on a teenager’s life, making it far more difficult to complete high school and find a job. Deferring childbirth until later in life would allow young women to mature, become more employable, and enjoy greater resources, thus better equipping them to control their children’s behavior and/or insulate them from harmful influences.

There may be something to this line of thinking — the R² is higher than that for poverty — but the effect is weak. There is not much to be gained by dwelling on this variable at this stage. Tomorrow, I will address the relationship between a different sign of social dysfunction: sexually transmitted diseases.

There are currently no comments highlighted.

7 responses to “Teen Births and Violent Crime — a Weak Connection

  1. So if mom has a terrible education and works all day, maybe at night to try to make a living what happens to the kids?

    A mom who has a good job – can afford to send her kids to day care or other places or hires someone to clean house and look after them.

    The thing is if you do not have a good education, “waiting” to have kids doesn’t have any better outcome.

    I think sometimes when people without knowledge of how to do statistics properly – start using the tools – and primarily looking for correlations, it’s a fools errand.

    And yes, I do think others abuse , often some right wing think tanks and yes some journalists but most journalists don’t actually try to generate or manipulate the data, they just latch on to something someone else did -and they don’t really vett it, they just publish it.

    But you don’t need to be (or should need to be) a rocket scientist to take a practical look at what happens to folks who live in poverty who have kids. Most love their kids just as wealthier folks do but they simply lack the economic means to provide for a nurturing life for a child. We actually do have criteria in the K-12 realm for kids of this type – they are called “at-risk” and there are several criteria that are used to designate. For instance, schools will provide free or reduced lunch. Many schools will also provide breakfast and during summer -some form of a daily meal if the kid can get to where they are served.

    People who fail to get a decent education don’t suddenly get better when they graduate and get a job. A good paying job sufficient to pay for rent, food and spend time with the kid to bring them up right is not a sure thing.

    We kind of miss the practical issues at ground level when we play around with broad brush data at 10,000 feet.

    And heckfire, if we don’t believe real scientists data modelling, …….

    In Rock Springs, this morning – headed north and west…

  2. “I’ll be the first to admit, giving me an Excel spreadsheet is the intellectual equivalent of handing a chimp a machine gun …”

    Well, Jim, at least you can spell, unlike those who see and hear through a fog of dyslexia compounded by uncorrected auto correction malfunctions (AI) and ADHD with the result that “Swedes are dower.” My mastery of statistics and charts are equally confused save for big pictures. One soldiers on.

    Hence, I would note that unwed teen pregnancy has fallen dramatically in this country. Here I suspect contraception education and abortion rule that decline. On results and remedies here, of course, we shout at one another across battlements formidable as the Maginot line. We all know how in the end that worked out for the French, and everyone else who has tried it.

    As to poverty, I’d suggest there is no poverty in America except that induced by a collapse of our culture, a collapse that has recently spread like wildfire throughout all classes of people in America save for the elite. Money is not our problem, except that we have far too much of it, hence we have wasted Tens of $Trillions in two decades, at a rate of wastage that mounts every year.

    In short, poverty in America today got nothing to with lack of money. But it has got everything to do with bad habits on the part of all involved, including those getting so rich off of their claims of poverty, a fraud and a sham, that they refuse to deal with the real underlying problems that fester within our horribly broken culture. Here Donald Trump has come to America’s rescue. He’s done far more to solve the problem than anyone in America since Ronald Reagan.

  3. Do you know how many of those single parents were teen parents? Do you know why they are single parents? What is their income? Education? Where do they live? Ages of children? What supports in the community do they use? We need more information – starting with more demographics. How is violence defined?? A lot more factors are involved and such a simplistic report/analysis doesn’t reveal dependable results.

    We don’t do regression to compare just two factors – we’d use correlation – so the analysis you’re taking this from has to contain more or the statistical procedure isn’t properly used. I suspect you’re trying to use part of a larger analysis to focus on just one piece. Maybe you should consult a statistician to ensure you know what’s what.

  4. Folks who do not receive a good education have lots of problems ranging from finding employment to being able to make informed/good choices and prioritizing what’s imperative versus not when your income is insufficient to pay for all needs.

    Many of us went through this when we were young and starting out and our entry level salaries just simply could not provide all that we felt was needed.

    Now take that entry level salary combined with a stunted k-12 education and watch it last a lifetime -there is no future when you finally start to get more salary and move up. Throw in a couple of kids and things get worse…

    and the thing is – look at those kids and ask yourself if they are going to grow up and follow in their parents footsteps in terms of a job, salary, and kids of their own.

    You should not need skills in graphing data to see how things are going to turn out.

    And yes.. it is so much easier to find blame on someone’s part – the parents, the advocacy groups, the media, etc etc… but the bottom line is what do we do those kids to keep them from repeating their parents failures?

    Yeah I know… it’s not “our” fault so why do we have to do anything at all.. it’s their problem and their fault for not doing better at school and graduating to a good job.

    Finally, we have a lot of focus on males – the school-to-prison pipeline .. keep in mind 1/2 of the problem is girls/young women – having kids and not getting enough education to pay for caring for those kids.

    That’s here right now – ongoing – and we’re doing what about it? drawing graphs that show correlations… I suppose so that we’ll then know what we should be doing instead of what we are right now.

    Not in favor of throwing good money after bad and not in favor of accepting the lack of personal responsibility as an excuse but at the same time if all WE DO is work at affixing blame… geeze.

    • Larry, we spend billions nationally on social programs (with a great deal of the money going to people providing services). In most locations, we spend a lot more on education for students coming from low-income families than we do on general education students because of Title 1 and related programs. We’ve got federal refundable tax credits for low-income workers with children. We spend money on food stamps. We spend money on housing assistance and day care assistance. Obama made birth control free. We’ve expanded Medicaid.

      When can we say “We’ve spend enough on you. You’re now responsible for your own decisions”?

      When can society expect a person to be responsible for the decisions that they make? When do we recognize that people have free will and, as such, are responsible for how they exercise their free will? Can society demand people answer for their own actions?

  5. Jim, at the risk of getting over my head myself into a statistical morass, I think I can say with confidence that your interesting and worthy story misuses some statistical terms.

    It states: “an R² of .1802, which is considered a small effect size. There is a much stronger correlation — an R² of .4007 across Virginia localities, a moderate effect size”.

    This wording conflates statistical significance with importance of the effect. “R-squared is a statistical measure of how close the data are to the fitted regression line.” It says nothing about the size of the effect, which is determined by the size of the slope of the line and one’s normative assessment about that size.

    The importance of the effect is the “oomph” popularized by Deiredre McCloskey in a famous article that has changed the field: “The Cult of Statistical Significance,” (with coauthor Stephen T. Ziliak). Here is from the abstract: “Fit is not the same thing as importance. Statistical significance is not the same thing as scientific importance or economic sense.” Link to article: https://www.deirdremccloskey.com/docs/jsm.pdf.

    Here’s a quick example: suppose I tell you with 98% statistical certainty (!) that if we spend an extra $1 million per patient, we can extend each person’s life by 1 hour. You’d rightfully say—“Jon, I don’t give a *$%^ if it is 100% guaranteed, it isn’t a large enough impact to make any economic sense funding!”

    And you’d be right. “Importance” (i.e., whether this is a strong or weak effect) is not determined by statistical significance. This may sound confusing, but McCloskey’s article lays it out, and I’d be willing to share a beer to talk it out or explore what the data do show.

    BTW, in the olden days (as when we went to school), many economists made this mistake of conflating R-squared with importance, so don’t feel bad! It’s a common problem. Over the past decades McCloskey has pretty much won this debate and econ students today are warned about this issue.

    • Thanks for the education, Jon. The difference between “importance” (which measures impact) and “statistical significance” (which measures variability) makes total intuitive sense. I’ll try to be more careful in the future.

      One of these days, I’ll have to take an actual course in statistics.

Leave a Reply