Previously on this show: how to write the introduction – a four-paragraph template

The first sentence of the first paragraph of any written piece of text is crucially important, as all writers of fiction know (“Call me Ishmael.”). Make it as strong as you can.

First impressions matter. The subset of potential readers who, after getting lured by the abstract of your scientific paper, have decided to have a closer look will first encounter the first sentence of the introduction. For them, this is another decision point: to read on or to stop. The second important sentence is the second one, and the third important sentence is the third one, and so on. The reader can choose to stop reading at any point, after each and every sentence. This means that the first sentence will be the most read sentence of your paper. Your second sentence will be read by fewer readers than the first, and your third sentence will be read by fewer readers still, and so on (if we assume that readers do in fact begin at the beginning instead of jumping in at random points). You will lose readers sentence by sentence whatever you do. This cannot be avoided.

The stronger the sentences, however, the lower the rate of attrition, and the higher the chance that some readers will make it through to the last one. Make the sentences flow and your readers will stick around. Glue them together with transitional words for clarity; place signposts to guide the reader. Create contrast and tension for excitement. Use cliffhanger endings: pose a question. Answer it in the next sentence.

Journalists use the term lede for the first few sentences of a news story—that is indeed how they spell it, instead of “lead”, presumedly for historical reasons that involve mechanical typesetting and lead (the metal, that is). The lede is the lead portion of a news story—it gives the gist of the story, it sets up the story and, most importantly, entices the reader to read the rest. While the lede should give a clear picture of what the story is about, it should not give the whole game away. The lede should raise questions so that the next paragraphs of the story can satisfy the curiosity of the reader by providing answers and details. Journalists even have their standard schemas for ledes. The inverted-pyramid lede attempts to compress the who-what-where-when-why-how of a story into a single sentence or two, and then adds details in decreasing order of importance. The question lede begins with, well, a question, one that you absolutely need to hear the answer to.

Let us have a look at some great openings and powerful first sentences.

As the first example, consider the first sentence of Battiston et al., “The price of complexity in financial networks”, PNAS 113, 10031(2016): “Several years after the beginning of the so-called Great Recession, regulators warn that we still do not have a satisfactory framework to deal with too-big-to-fail institutions and with systemic events of distress in the financial system”. This is a powerful beginning that immediately tells what the general problem addressed by the paper is. It also forces the reader to read on—after all, who wouldn’t want to know where this story is going?

Another example of a great opening, from Centola & Baronchelli, PNAS 112, 1989 (2015): “Social conventions are the foundation for social and economic life. However, it remains a central question in the social, behavioral, and cognitive sciences to understand how these patterns of collective behavior can emerge from seemingly arbitrary initial conditions.” The problem that drives the research is clearly spelled out in the second sentence. Note that in this paper, the exact research question will not appear before the 4th paragraph. The introduction forms a funnel from the broad problem to the more detailed question.

Finally, here is the first paragraph of Altarelli et al., Phys. Rev. Lett. 112, 118701 (2014): “Tracing epidemic outbreaks in order to pin down their origin is a paramount problem in epidemiology. Compared to the pioneering work of John Snow on 1854 London’s cholera hit [1], modern computational epidemiology can rely on accurate clinical data and on powerful computers to run large-scale simulations of stochastic compartment models. However, like most inverse epidemic problems, identifying the origin (or seed) of an epidemic outbreak remains a challenging problem even for simple stochastic epidemic models, such as the susceptible-infected (SI) model and the susceptible-infected-recovered (SIR) model.”

The above paragraph gets from the topic (tracing epidemic outbreaks) to the research question (identifying the origin of an epidemic) with three sentences, and the authors have even managed to include a brief historical detour of the you-know-nothing-John-Snow variety (sorry, I had to). This a great opening. The reader gets a clear idea of what the paper is about, and becomes curious: how did they solve the seed identification problem?

In the next post, we’ll move from the introduction to methods & results.