Published on Development Impact

We know socio-emotional skills are important – How do we develop them? A review of Paul Tough’s Helping Children Succeed

This page in:
ImageIn 2012, Paul Tough published How Children Succeed: Grit, Curiosity, and the Hidden Power of Character, in which he highlighted how children who develop “noncognitive capacities” like self-control and perseverance do much better in adulthood. But as he spoke on that book, audiences repeatedly confronted him with the question: “Ok, now that we know this, what do we do?” So he wrote another book to answer that question – Helping Children Succeed: What Works and Why.

While Tough follows a familiar pattern for popular social science books – accessibly summarize social science research and accompany it with stories – he does it well, and he explicitly seeks not to highlight great programs as specific models to scale up, but rather “as expressions of certain underlying ideas and strategies.” To an economist, this evokes Ludwig et al.’s work on Mechanism Experiments and Policy Evaluations: Don’t just evaluate programs; evaluate mechanisms that we can apply to a wider range of programs. Tough relies on a wide range of studies, including lots of randomized trials, and the study dates consistently extend well into 2016. 

So “what works and why”?  

It’s not about teaching; it’s about creating an environment. Tough makes the argument that many of the best examples we have of individuals who seem to help youth develop noncognitive skills “never said a word about these skills in the classroom.” Add to that the stress literature from early childhood ( here and here, for example), demonstrating how children exposed to more traumatic events have a wide array of adverse physical and psychological outcomes later in life. And for older children, bring in the evidence that high rates of suspensions from school adversely affect not only the children suspended but also those left behind. He ultimately concludes that noncognitive capacities are “products of a child’s environment,” which means we can’t teach them the same way we teach math and reading.

What classes of programs seem to create this environment? I know that mechanisms were promised earlier. Tough stops short of giving cut-and-dried conclusions, which I suspect is fair given the state of the research. But he does a great job of highlighting an array of programs in a category, rather than a single example to take to scale.

1. Home visits with parent coaching. It’s not enough to provide information, but coaching parents has shown good results. Coaches watch parents interacting with their children, praising and building on positive interactions. Two examples are Attachment and Biobehavioral Catch-up (with randomized trial results here and here) and Filming Interactions to Nurture Development

2. Early childhood programs outside the home that provide a “plenty of responsive, warm, serve-and-return interaction with caring adults.” Examples are Educare – which has an ongoing randomized trial but positive preliminary results, All Our Kin – which also rates well in a recent evaluation (March 2016!), and the Chicago School Readiness Project, with cluster randomized trial evidence of success in child development outcomes.

3. Not financial incentives. Here, Tough relies principally on Roland Fryer’s randomized trials. On teacher incentives in New York: “I find no evidence that teacher incentives increase student performance, attendance, or graduation, nor do I find any evidence that the incentives change student or teacher behavior.” And on student incentives in New York, Chicago, and Dallas: “The impact of financial incentives on student achievement is statistically 0 in each city.” Two studies of student incentives showed positive results, but only for the highest achieving students at baseline, so not really the target population.

Here, I should note that all the evidence on financial incentives is that they don’t improve test scores, when in fact Tough has made a strong case that noncognitive abilities aren’t developed in the same way as reading and math. That said, my intuition is that noncognitive abilities might be even less responsive to extrinsic incentives.

4. Classrooms that create a “positive emotional climate,” show “sensitivity to student needs for autonomy,” and – in addition to that environment – “a pedagogical approach that promoted student engagement in the learning process.”

How do you measure success in developing these noncognitive abilities? In David McKenzie’s review of the recent book Grit, he highlights the measurement challenges self-reported or teacher-reported noncognitive abilities. Tough instead turns to Kirabo Jackson’s work (May 2016!), where four pieces of school administrative data stand in for noncognitive ability: school attendance, suspensions, timely grade progression, and grade point average. One can argue about whether these capture noncognitive abilities, but Jackson shows – in North Carolina – that these measures predict college attendance, future wages, and future arrests better than test scores. He’s also able to construct a value-added measure, showing that some teachers are consistently better at improving these “noncognitive outcomes.” As Tough concludes, “Instead of laboring to come up with a perfectly calibrated new measurement tool for grit or self-control or self-efficacy, we can measure noncognitive capacities by measuring the positive outcomes that we know those capacities contribute to.”

Reactions as a Development Economist
Overall, I found this book a pleasure to listen to and then (partly) read. Here are a few reactions:

1. Almost all of this evidence is U.S.-based. How relevant is it for low- and middle-income countries (LAMICs)? A common critique in development is that rich-country consultants bring their rich-country solutions that fail dramatically (sometimes comically; sometimes tragically) in local contexts. That’s an important principle. At the same time, as I read review after review of what works to improve learning in low- and middle-income countries, I wonder if we’re missing something by not drawing on evidence from rich countries. This is where I think Tough’s emphasis on mechanisms is useful. It doesn’t make sense to drop a program from Chicago into Lomé, but I suspect that when we lack good, local evidence, rich-country evidence can give good suggestions on where to start.

There is a growing literature on parent coaching in LAMICs, but my sense is that the evidence on the emotional climate in classrooms and on promoting student engagement still has much more room to grow. Interestingly, while the US evidence on incentives has been dismal, there is positive evidence on teacher incentives in India (and not-so-positive evidence in Kenya), and positive evidence on student incentives in Benin. That’s not to say that financial incentives are the answer to our problems, just that the global evidence is more nuanced.

2. Where does a book like this fall in the synthesis literature? I’ve written and spoken before about different kinds of synthesis, ranging from narrative reviews to meta-analysis. Of course, most non-specialists won’t read technical reviews. Instead, they’ll read books like this. This is fundamentally a narrative review, with the addition of rich qualitative work. The qualitative work, providing a clear picture of how programs actually work, is an improvement on many quantitative reviews.
Here’s what troubled me: Most of the evidence in the book fed into a single narrative. I found myself thinking: Are there studies that muddy the narrative that weren’t included? In my own work, I’ve seen that the variation in effectiveness within classes of programs can often be at least as high as the variation across classes of programs.

Tough has really innovated here, with a slim, accessible book accompanied by a host of on-line resources. I don’t object to a narrative, but I’ll propose an addition for future work. Technical readers would benefit from an online appendix that characterizes the universe of studies that Tough reviewed, which studies don’t fit the narrative, and why they didn’t make it into the book.

How should you consume this book? The hardcover book – coming in at a clean 114 pages – is beautiful but has no references and no figures (just, you know, words); so it’s great for getting the argument but insufficient for a reader who wants to know exactly which studies are driving the conclusions. The audiobook – which I listened to, and is narrated well by Tough – has the same advantage (easy!) and disadvantage (sources!).
Tough has also provided a free version on-line, complete with blog-style links to sources, relevant charts, figures, and videos. (It’s pretty cool.) I’ve drawn on that heavily for this post. You can download a PDF.

Read more book reviews on Development Impact.


David Evans

Senior Fellow, Center for Global Development

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000