Book Review: How School Really Matters (and how Analytics Wins Every Argument)

How School Really Matters, Douglas B. Downey, University of Chicago Press, 2020

The sub title of this book is “Why our assumptions about school and inequality is mostly wrong”. This tells you the real focus of the book and it’s a short (159 pages with appendix and notes), concise study looking at educational attainment but using data from outside school age, that leads one to pause and question the widely accepted positions.

As a practitioner in data and analytics one learns quickly to take data with a BIG pinch of salt. More importantly one continuously tries to not forget the different between correlation and causation; and the definitions and meaning of confidence and bias. Armed with these weapons most of what one reads in the press becomes interesting at most; informative hardly ever.

Conventional wisdom is that schooling, differences in public and private access, and level of funding, are primary causes of resulting inequality when kids leave school.  Kids in low performing schools have fewer opportunities, and kids in rich schools with more investment get better options, so the story goes.  To reduce this gap it is generally accepted that we should increase funding in schools less advantaged, such as public schooling and those particularly in depressed inner cities. This has become a standard mantra for as long as I can remember.

Downey uses a little history and a new lens to look at disparity in educational levels of kids in school as well as before school. And it’s that additional lens that exposes what appears to create a new argument. By comparing capability and performance before kids even start school, it would seem that disparity is already established.  This is non obvious.

We know from various research that schooling can, in some places, increase inequality. There are studies that show children that go to rich, private or well funded schools appear to perform ‘better’ than children that go to ‘poor’, public or schools in depressed environments. Note of course there are also other studies that show the gaps closing. Your political leaning may be a good predictor for which study you reach for.  But that is not the point of the book.

But the author shows new data that demonstrates the root of the disparity may actually be established before children even start school. Moreover he demonstrates that even after children have started school, many studies only review performance differences during the school year.  By expanding the time period over which performance is analyzed, it can be shown that performance widens more so during the periods outside of the school year.  In other words, even if inequality is increased or decreased during the school year, when kids are at home and outside of school, the gap widens.  That is thought provoking, in my mind.

These two points are intriguing. The author then goes back into history and reviews some early studies that exposes these findings. Back in the ‘60s there were arguments that were developed that suggested family structure, social support, and parental care and interest in kids lives may have a greater impact on the disparities in inequality before the child even starts school,  Schooling simply neutralizes (or slightly increases) the gap but it might not be the primary cause.

These arguments did not fit with the popular views of the day and such studies were ignored and society as a whole preferred to focus on school funding. It was, as the author implies, an easier scape-goat. And so the wider conversation remains focused on school funding, and not on family and social support and wellbeing.  From the D&A viewpoint, changing the time-frame of any analysis can upend a considered “fact” or insight.  That’s a big point to make.

Whatever your political leanings the data and insight is thought provoking. As a data and analytics practitioner the data should be taken with that pinch of salt; much as the conventional wisdom should be (but often isn’t). As if to demonstrate the authors’ political persuasion, his policy response is to focus on increasing direct investing on pre-school and family promotion.  He does not suggest we stop or reduce funding in schools.  It’s not that school does not matter; it’s just that it’s matters less than we think. I am no educator but the argument is plausible. Maybe we need more studies, more data, and more arguments, and more salt.

Recommend 8 out of 10.
Source: Gartner Hybrid Cloud