Every chapter in this volume is built identically. There are four sections, each covering one part of the chapter's subject. Each section carries: a beginner tier, an intermediate tier and an advanced tier, and each tier holds two to four titles. A chapter therefore contains thirty-three books. Twelve of them, one per tier per section, are marked with a star as the recommended place to start. Each section closes with a short recap identifying its strongest titles, and each chapter closes with a longer recap, drawing the chapter together. Across the volume this produces 660 entries and 240 starred recommendations.
The uniformity is deliberate and serves the reader. A person who learns the format in Chapter Two does not relearn it in Chapter Fifteen, and two chapters can be compared directly because they are shaped the same way. It also imposes discipline on the writing: a subject with a vast literature and a subject with a thin one are given the same space, which forces the thin subject to be honest about its thinness rather than over-expanding.
One thing the uniformity does not mean, and it should be said here rather than left for a reader to work out. Equal length is an editorial convenience, not a claim that the twenty subjects rest on comparable evidence. They do not. Chapters Nine through Eleven and Chapter Fourteen stand on the firmest ground in the volume, with decades of randomized trials and large cohort studies behind them. Chapters Twelve, Fifteen and Sixteen contain the widest gaps between the confidence of popular literature and the evidence supporting it. A reader who treats the format as a signal of equivalence will overrate the thin subjects and underrate the solid ones. The individual entries state the strength of the evidence; the architecture cannot.
Each entry follows the same order. The title and author come first, then the author's credentials, stated plainly and including the absence of a relevant credential where that is the case. The body of the entry then covers what the book contains, what distinguishes it from comparable titles, why it earns its place in the collection, and finally its limitations, which may be dating, scope, an argument that outruns its evidence, or a commercial interest held by the author. Entries run to roughly a hundred and fifty words, which is enough to be useful and short enough that a reader can work through a section in a sitting.
A word on the recaps, which do more work than their placement suggests. The recap closing each section names the recommended entry point and the recommended next step, and explains the reasoning for both, so a reader who reads nothing but the four recaps in a chapter still leaves with a defensible reading plan. The recap closing each chapter goes further: it states what the chapter concluded, which claims within the subject are contested and which are settled, and what a reader should carry forward. In several chapters the recap is the most useful thing in it.
Cross-references between chapters are given by chapter numbers and are always deliberate. Where a subject genuinely belongs somewhere else, the entry says so rather than duplicating the coverage. Where two chapters approach the same problem from different directions, both say so. And where a reader would be badly served by starting in the wrong place, particularly with existing pain or a diagnosis, the chapter that would mislead them says which chapter to read first.
The volume can be read in any order. Chapters are self-contained, cross-references between them are explicit, and nothing later depends on having read anything earlier. A reader with a specific question should go to the chapter that answers it. A reader without one may find the reading paths in the next section more useful than the table of contents.
How to use this collection
Six hundred and sixty books is an intimidating number, and it is not meant to be worked through. This volume is a reference: a reader should expect to use a selected portion of it, chosen deliberately. What follows explains how the entries are marked and offers four routes in.
Every title sits in one of three levels. The levels describe the background a book assumes, not how good it is. This is the most common misreading of a structure like this one, and it matters: an advanced title is not a better book than a beginner title and choosing above your level is the most reliable way to stop reading.
Two habits are worth adopting before you start.
First: every entry in this volume ends with the limitations of the book it describes. Read that part. It is what will change what you take from the book, and it is exactly what a promotional description would leave out.
Second: read a chapter's introduction before looking at any of its entries. The introduction states what the field agrees on, where it does not, and which widely repeated claims have not survived scrutiny. An entry read without that context can be misleading, because a book may be excellent and still belong to a school of thought the evidence has moved away from.
A note on how much to buy. Very few readers need more than a handful of these books, and the collection is not designed to be purchased. Libraries hold most of the trade titles and can obtain the academic ones through interlibrary loan; the Library of Congress classifications given for advanced titles exist partly to make that request easy. A reader who reads five books carefully will finish better informed than one who buys thirty and reads the first chapter of each.
BEGINNER
Assumes no prior knowledge of the subject and no scientific vocabulary. Written for a general reader, usually by a journalist, clinician or practitioner who explains rather than argues. These are the books to start with in any chapter, and for many readers they are the only books needed. Two hundred and sixty titles.
INTERMEDIATE
Assumes the basics of the subject and some comfort with how evidence works. Usually written by practitioners or researchers for readers who already train, already cook, or already have a diagnosis, and who want the reasoning rather than the instruction. Two hundred and forty titles.
ADVANCED
Assumes professional or academic background. Textbooks, clinical handbooks and research references, several of which are not meant to be read continuously but consulted on a specific question. These are the titles cataloged by Library of Congress classification. A general reader is not expected to buy these, but should know they exist, because it is where the claims in the other two levels can be checked. One hundred and sixty titles.
Within each level, one title per section carries a star. The star marks the best place to start, which is not the same as the best book. It usually means clearest, most current, or most likely to be useful to the largest number of readers. There are two hundred and forty starred titles across the volume, twelve in each chapter.
★ best place to start
Four routes through the collection, depending on why you are here.
If you are starting from nothing. Read Chapter One first, particularly its third section on evidence, then go to Chapter Nine, which covers why exercise matters and how to begin. Those two chapters will do more for a reader with no background than any six others. From there, follow whichever subject you are actually curious about, and read only the starred beginner titles until one of them makes you want more.
If you have a specific goal. Go directly to the chapter that addresses it and read that chapter's introduction before looking at any entry, since the introduction states what the field agrees on and where it does not. For losing weight, Chapters Seven and Eight. For starting to run, Chapter Eleven with Chapter Nine underneath it. For a joint that hurts, Chapter Thirteen. For sleep, Chapter Fourteen. For a diagnosis you are managing, the relevant chapter followed by Chapter Nineteen, because adherence rather than information is usually the binding constraint.
If you are a professional. The advanced tier of each chapter is the reference layer, one hundred and sixty titles across the volume, and the advanced entries state what each work is authoritative for. The section recaps identify which reference settles which question. If you work with clients or patients rather than only with the literature, the intermediate tiers of Chapters Nine, Ten, Thirteen and Nineteen will be more immediately useful than the advanced tiers of anything.
If you will only read five books. That is a reasonable position and the collection is built to accommodate it. The Essential Reading List that follows this section identifies twenty-five titles and explains why each is there. The first five of those are chosen so that a reader who stops after them has the substance of the collection.
Source: the printed volume, pages 13–14 and 19–20. Reproduced from the printed volume.