Samantha Kleinberg - Why: A Guide to Finding and Using Causes

See http://oreilly.com/catalog/errata.csp?isbn=9781491949641 for release details.

While the publisher and the author have used good faith efforts to ensure that the information and instructions contained in this work are accurate, the publisher and the author disclaim all responsibility for errors or omissions, including without limitation responsibility for damages resulting from the use of or reliance on this work. Use of the information and instructions contained in this work is at your own risk. If any code samples or other technology this work contains or describes is subject to open source licenses or the intellectual property rights of others, it is your responsibility to ensure that your use thereof complies with such licenses and/or rights. PROD: Uncomment the following sentence if appropriate and add it to the above para: This book is not intended as [legal/medical/financial; use the appropriate reference] advice. Please consult a qualified professional if you require [legal/medical/financial] advice.

978-1-4919-4964-1

[LSI]

Will drinking coffee help you live longer? Who gave you the flu? What makes a stocks price increase? Whether youre making dietary decisions, blaming someone for ruining your weekend, or choosing investments, you constantly need to understand why things happen. Causal knowledge is what helps us predict the future, explain the past, and intervene to effect change. Knowing that exposure to someone with the flu leads to illness in a particular period of time tells you when youll experience symptoms. Understanding that highly targeted solicitations can lead to political campaign donations allows you to pinpoint these as a likely cause of improvements in fundraising. Realizing that intense exercise causes hyperglycemia helps people with diabetes manage their blood glucose.

Despite this skill being so essential, its unlikely you ever took a class on how to infer causes. In fact, you may never have stopped to think about what makes something a cause. While theres much more to the story, causes basically increase the chances of an event occurring, are needed to produce an effect, or are strategies for making something happen. Yet, just because a medication can cause heart attacks does not mean that it must be responsible for a particular individuals heart attack, and just because reducing class sizes improved student outcomes in one place does not mean the same intervention will always work in other areas. This book focuses on not just what inferences are possible when everything goes right, but shows why seeming successes can be hard to replicate. We also examine practical questions that are often ignored in theoretical discussions.

There are many ways of thinking about causality (some complementary, some competing), and it touches on many fields (philosophy, computer science, psychology, economics, and medicine, among others). Without taking sides in these debates, I aim to present a wide range of views, making it clear where consensus exists and where it does not. Among other topics, we will explore the psychology of causality (how do people learn about causes?), how experiments to establish causality are conducted (and what are their limits?), and how to develop policies from causal knowledge (should we reduce sodium in food to prevent hypertension?).

We start with what causes are and why we are often wrong when we think we have found them (Chapters ).

Large datasets make it possible to discover causes, rather than simply testing our hypotheses, but it is important to realize that not all data are suitable for causal inference. In ). This book will give you an appreciation for why finding causality is difficult (and more nuanced and complex than news articles may lead you to believe) and why, even though it is hard, it is an important and widely applicable problem.

Though there are challenges, you will also see that its not hopeless. Youll develop a set of tools for thinking causally: questions to ask, red flags that should arouse suspicion, and ways of supporting causal claims. In addition to identifying causes, this book will also help you use them to make decisions based on causal information, enact policies, and verify the causes through further tests.

This book does not assume any background knowledge and is written for a general audience. I assume only a curiosity about causes, and aim to make the complex landscape of causality widely accessible. To that end, well focus more on intuitions and how to understand causality conceptually than mathematical details (actually, there wont be any mathematical details). If you have a PhD in computer science or statistics you may pick up some new tools and enjoy the tour of work in other fields, but you may also yearn for more methodological detail. Our focus here, though, is causality for everyone.