Moral judgments influence assessments of objective risk

Well this is an interesting example of the tangle of evidence, media communication and motivated reasoning.

I first came across If you leave your kids alone, it’s not predatory strangers who are a risk by Virginia Postrel which was published in 2016. Postrel is reporting the results of some research but does not link to the research study. I went in search of the original study.

In doing so, I came across NPR's report on the same research, Why Do We Judge Parents For Putting Kids At Perceived — But Unreal — Risk? by Tania Lombrozo. Lombrozo does both a report on the research as well as an interview with the researchers. From the NPR article, I was able to find the original research, No Child Left Alone: Moral Judgments about Parents Affect Estimates of Risk to Children by Ashley J. Thomas, P. Kyle Stanford, and Barbara W. Sarnecka.

Triangulating these three sources, you suddenly get a very interesting alternate set of answers to questions about:

The empirically determinable risk to children of being left alone unsupervised.

The hard to measure impact of unintended consequences.

The even harder to discern risk arising from over-protection of children which might be even larger than the risk of being unsupervised.

The gap between empirical risk and perceived risk.

The question of how assessments of morality affect perception of risk.

The question of the role media plays (intentionally or not) in distorting perceived risks.

The question of the role media plays (intentionally or not) in distorting the facts of the research.

The question of the net risks of tactical childhood actions and longterm life outcomes.

At the heart of these articles is an unstated question: What is the appropriate level of freedom parents should accord to their children to develop self-sufficiency, autonomy, and skills while remaining cognizant of the relative probabilities of risks associated with such freedom (stranger danger and accidents) and the relative benefits (self-sufficiency, autonomy, transition to adulthood responsibilities, etc.)?

Tangled up with this question are all sorts of observable trends of recent years such as helicopter parenting, infantilization of young adults (or adult responsibility postponement), continued expansions of the definitions and legislation of what constitutes child abuse, the increase of the surveillance state (both state and citizen), and hyperbolic media communication.

To expand on the five issues which are much in the news but rarely treated with rigor or objectivity.

Helicopter parenting is the constant presence of parents in their children's lives, socially, educationally, athletically, employment, etc. It is the attempt on the part of parents, not only to keep children safe but also to ensure that they do not face great difficulties and that they are successful (however defined). The classic trope of a helicopter parent is that of the parent doing the sixth grade child's science project for them.

Infantilization of young adults is the postponement of normal societal points of transition from dependent childhood to autonomously responsible adulthood. Independent living, employment, marriage, family formation, etc. are all being shifted out to later and later ages. The classic trope is of the 27 year-old child still living at home.

Expansion of the definition of child abuse is the process by which social fears are legislated into being and then enforced through the legal system for activities which were the norm in recent years. Parental legal obligations have shifted away from shielding children from clear and present dangers (don't let them play with rabid wolves) to an obligation to avoid long term health conditions (example, secondary smoking, light drinking while pregnant, etc.) to an obligation to avoid conceivable risks. It is the precautionary principle forced onto parenthood. The classic trope is the parent being arrested for allowing their six year old to walk unassisted three blocks to the child's friend's home, crossing two streets.)

Rise of the universal surveillance state is the increasing intrusion into private life of state surveillance via CCTV cameras, sound recording, financial tracking, email capture, etc. It is not just the state. With the near ubiquity of smartphones, it is a matter of a minute for a well-intended busybody to take a picture, video, or otherwise report to police an action which they deem to be illegal, reckless, or otherwise dangerous, even when the action might not be. The classic trope is the current efforts in Britain and Germany (and elsewhere, such as American universities) to criminalize free speech.

Hyperbolic media communication - This ranges from simple communication techniques 'You won't believe what happened next', 'It was literally a million times worse', 'Hate speech is the equivalent of a physical attack', etc.; to fake news/unreliable news, the inaccurate reporting of events; to the well documented replication crisis arising from media reporting the most extreme research findings despite counter studies which undermine the extreme finding (publication bias), to straightforward click bait strategies to attract attention rather than to inform.

These five trends are rife with unintended consequences.

With all that as background, what were the researchers actually investigating? It's there in their title: Moral Judgments about Parents Affect Estimates of Risk to Children. The normative moral judgments people make of parents affect the observer's opinion about the nature of the risk children are exposed to. That's an interesting insight. From the abstract.

In recent decades, Americans have adopted a parenting norm in which every child is expected to be under constant direct adult supervision. Parents who violate this norm by allowing their children to be alone, even for short periods of time, often face harsh criticism and even legal action. This is true despite the fact that children are much more likely to be hurt, for example, in car accidents. Why then do bystanders call 911 when they see children playing in parks, but not when they see children riding in cars? Here, we present results from six studies indicating that moral judgments play a role: The less morally acceptable a parent’s reason for leaving a child alone, the more danger people think the child is in. This suggests that people’s estimates of danger to unsupervised children are affected by an intuition that parents who leave their children alone have done something morally wrong.

To make it very concrete, an observer's assessment of risk to a child changes depending on the observer's understanding of the parent's motivation. Observer's assess the real danger to a child left alone playing in a public park for thirty minutes as being different dependent on whether the child was left alone so that the parent could go take a nap, could go do some work, or had to return a lost dog to an owner. The more the action was taken for perceived self-indulgent reasons, the more danger the child was perceived to be in. The more the action was taken for perceived noble reasons or compelled by unavoidable circumstances, the less danger the child was perceived to be in.

The objective danger to the child does not change regardless of the reason for the parent's absence and yet people let their moral judgment affect their estimation of risk.

The interview portion of Why Do We Judge Parents For Putting Kids At Perceived — But Unreal — Risk? is revealing. Clearly the researchers are motivated by more than academic interest. They are motivated by a sense that the world was not the way it should be. They were concerned about social justice and gender equity. Nothing wrong with these as motivations but it sheds some light on the reporting confusion.

I thought, here's a single mother who works for low wages for a corporation that doesn't provide child care, and she was treated as a criminal for letting her daughter do something that is relatively safe. It seemed like people were angry at this woman for not being a full-time mom — for not fulfilling the unrealistic expectation that mothers should be with their children at all times. Those are moral judgments, but people weren't talking about it in moral terms. Instead, they were using the language of risk and danger — saying that Harrell was criminally negligent because she had left her daughter in a dangerous situation.

[snip]

The increasing frequency of legal action against parents of the kind that both Ashley and Barbara describe was certainly our primary motivation.

[snip]

I guess what surprised me the most was the difference in responses that we got from different groups of participants — men vs. women and parents vs. nonparents. I expected that mothers would be less likely than other people to buy into this paranoid view that any child, of any age, left alone for any length of time is going to drop dead the minute the mother looks away. I guess I figured that I'm a mother, and I'm not hysterical about this, so I assumed that other mothers weren't either. But I was wrong — mothers rated all the situations as more dangerous than fathers did; followed by childless women and finally childless men. The people with presumably the most child care experience (mothers) actually expressed the most exaggerated overestimates of risk. I was genuinely surprised by that. But I guess that's because I was expecting people to be rational, and people are just not rational about this subject.

[snip]

It seems to be socially acceptable to harass parents (particularly mothers) who are "caught" leaving their child unattended for any time at all.

[snip]

If you think about Debra Harrell's situation, she's raising a child while working a minimum-wage job. Suddenly, we as a society have decided (without any rational basis) that she is negligent for allowing her 9-year-old to play in a public park. This is very, very disturbing to me. It is basically criminalizing poverty and single parenthood.

I am sounding overly critical. I reiterate that their finding that moral judgments influence estimations of risk is an important and valuable finding (subject to replication).

But they haven't actually answered any of the issues raised in the reporting and interviews.

I am deeply sympathetic to the idea that we need to quit over-legislating manners and risks. If freedom is to mean anything, we need to allow people to make their own decisions.

But there remains an empirical question unanswered - Is it riskier for children to be closely supervised or to be left on their own to play? The researchers don't tackle that question. In the interviews and the reporting, multiple issues are run together without distinguishing them. Everyone is absolutely correct that stranger danger remains a real but vanishingly small danger to children. But the question as to whether children are safer supervised or unsupervised is different. Is a child more likely to suffer death and injury on their own or in the company of their parents. I don't have access to a robust treatment of that question but I would feel fairly confident that, whether in absolute terms or by a normalized rate, children are almost certainly in greater jeopardy when unsupervised.

From their interviews, the researchers appear to want to reduce the frequency with which parents, and mothers in particular, are harassed for leaving their children unsupervised. I agree. Absent clear and present danger, the state has no business intruding in parental risk assessment. I agree with the researchers goal but their work does not address the relevant question and unfortunately were such work to be done, it likely would support the nanny-state proposition and enshrinement of the precautionary principle (which is fundamentally wrong) that nothing should be done unless it can be shown to be risk free.

I think the reporters and researchers are adopting the wrong framing of the question.

Everything entails risk. Some number of children will die each year from avoidable accidents, whether supervised or unsupervised.

The better framing of the question is: What is the optimal balance of autonomy and supervision over a childhood in order to prepare them with the best assortment of knowledge, skills, values, and behaviors for a successful adulthood?

We are seeing more and more evidence of the harm that arises from children with too sheltered a childhood. Adulthood is postponed, critical skills are learned later when there are substantive consequences, behaviors and values are not acquired at all or too late.

If you practice free range child rearing, you commit to a life strategy where children build skills, knowledge, values and behaviors early on which will ensure that they have greater probabilities of success later on in life. You accept a tiny increase in risk of mortality, and a (probably) significant increase risk of injury in childhood when they are most resilient in order for the child to have long term strategic benefits (including, because of the better knowledge, values, skills, and behaviors, better morbidity and reduced adult mortality.)

If you commit to a helicopter parenting strategy, you commit to a life strategy where children are more protected and sheltered in childhood but at the risk that they have a poor mix of knowledge, skills, values, and behaviors when they emerge as young adults.

Right now, I am unaware of any research that allows us to answer those questions. And indeed, I think, based on our constitution, that it would anyway be moot. The state should not be impinging on the freedom of parents in that fashion in order to adjudicate complex and contingent risks. It would be interesting to know the answer, and I am inclined towards the free-range end of the spectrum, but it would provide no moral basis for law.

Thomas, Stanford, and Sarnecka are doing interesting research from a social justice, class, and maybe ideological position. They arrive at an interesting finding which has nothing to do with the core question. When they publish their findings, reporters then write articles based on the reporter's preferred narrative rather than on the limited scope of the research. It is the same point Fivethirtyeight is making here, the media see all information filtered through their pre-existing assumptions.